Carrida Camera Program User Documentation

Revision: 1.6
Date: 2019-08-27
Contact: support@carrida-technologies.com
Copyright: 2017-2019 Carrida Technologies GmbH, Ettlingen, Germany
Author: Carrida Support

Home page

Table of Contents

1   Introduction

Carrida is a software library for Automated License Plate Recognition (ALPR) in images. The Carrida Camera integrates an imaging and processing device, together with the Carrida 4 engine into a compact license plate reading sensor.



Typical applications of the Carrida Camera include:

  • Parking lot management
  • Vehicle ticketing/management
  • Automatic barrier opening
  • Vehicle counting
  • Law enforcement applications
  • Traffic planning (count vehicles, map traffic to geographic regions)
  • Traffic flow monitoring


Carrida offers the following features and benefits:

  • Easy to configure
  • Support for State/Country recognition
  • Supports license plates with two rows
  • Supports license plates with stacked characters
  • Robust against illumination, contrast, noise, and rotation


For further details and information see also the related documents:



A list of abbreviations used throughout this document:

ARM64
Stands for processors with the v8 AARCH 64-bit instruction set.
ARM
Stands for processors with the v7 ARM 32-bit instruction set.
ALPR
Automatic License Plate Reader.
dB
Decibel, in the Carrida documents used to describe a relative sensitivity (gain) value in cameras.
fps
Frames per second.
FOV
Field of view.
HW
Hardware.
LP
License Plate.
mA
Milli-Ampere, one 1/1000:sup:th of an ampere.
ms
Millisecond, one 1/1000:sup:th of a second.
OSD
On-screen display
PC
Personal Computer.
PK
Product Key (for software licensing).
px
Pixel. The smallest unit of resolution in an image. For example, VGA images are made up of 640x480 pixels.
ROI
Region of Interest, typically a marked area in an image.
SDK
Software Development Kit.
SW
Software.
x86
Stands for Processor Architectures with the Intel instruction set. 32-bit hardware.
x86_64
Stands for Processor Architectures with the Intel instruction set. 64-bit hardware.
VC
Vision Components GmbH.
Image symbols used in this document
Symbol Meaning
Note Sign Points to especially important aspects of the manual or the software.
Hint Sign The lightbulb highlights the hints and the ideas that may be helpful.
Warning Sign This warning sign alerts of possible pitfalls to avoid. Please pay careful attention to sections marked with this sign.


The Carrida camera user interface uses the following symbols and icons:


Symbol Description
1q

Add button

add single entries, new actions or new servers to the already existing ones.
2q

Add user button

add new User profiles.
3q

Apply changes button

apply changes.
4q

Export as csv:

export the text result table as csv.
5q

Export as excel

export the text result table as an excel table.
6q

Export images

export images for filtered results of the text result table.
7q

Upload button

upload (import) certain files (ex.: result images, classifiers, single list entries or entire lists as .txt files).
8q

Download button

download (export) certain files (ex.: entire lists as .txt files).
9q

Cancel button

cancel editing and to stash the changes.
10q

Edit button

edit some settings in the configuration menu.
11q

Delete button

delete all entries in a list.
12q

Filter button

filter the results from the Text results menu by date, time, plate and/or state confidence.
13q

Full-screen button

stream the video from Live view in full-screen mode.
14q

Info button

display additional information
15q

Left button

move left
16q

Right button

move right
17q

Remove button

remove a single entry (ex.: when manually adding list entries, a single defined period of time can be removed this way).
18q

Reload/ Sync button

it is used in:

  • Text results to reload the page and show the newest results.
  • to synchronize the Brightness ROI with the ALPR ROI

2   Quick start with the Carrida Camera

Once the camera is connected to your local network, it can be accessed by using a web browser. Enter the provided IP address in the address bar of your browser (i.e. Chrome, Firefox, Opera, etc.) and the Carrida login page will open. A username and a password are required to sign in.

The Carrida User Interface offers four different user profiles: administrative, basic user, setup and viewer. The main difference between these profiles is the level of access to configuration and system settings. The factory Login credentials are as follows:


Administrator: Basic User: Setup: Viewer: SuperUser:

username: admin

password: secret

username: user

password: simple

username: setup

password: setup

username: viewer

password: pass

username: su

password: car0000rida as default, but this should be changed on the first start of the server.

Access to all settings. Has the authority to add, delete and edit other users. Access to the most commonly used settings. Access to view live image streaming and reading results. Has the authority to change the ROI and classifier settings. Access to view live image streaming and reading results. No authority to change any parameters. Access to all settings, deleting admin users and access to UI Design setting.

tt

To switch between User modes, click on the User button in the upper right corner of the interface.

Administrators have access to all camera, system and configuration settings.

We recommend not to change those configuration settings which are visible in the administrator mode, unless absolutely necessary.



Note

Note Sign

All graphical images in the remaining parts of the manual will display the interface under admin mode.

2.1   SuperUser

Starting from the 4.3.2 release a further user type is introduced to Carrida. This user has the username su and has the highest priority, which allows the deletion of admin users, as well as design changes in the UI of the camera.

The password for this user is constructed using the following naming pattern car<SN of the camera>rida.


Note

Note Sign

As factory default the SuperUser password is set to car0000rida. It is highly recommended to change the default password for this user immediately.


When logging in with this username for the first time, a dialog window that allows immediate changes to be made will appear, as illustrated in the figure below. For later password changes, the standard method should be used as described in the section Users.


./images/2_su_password_dialog.png

2.2   Configure areas for reading and brightness control

ROI, short for Region of Interest, refers to the area in which some processing is done. There are two different types of ROIs used by the Carrida Camera:

  • the ALPR detection ROI to define where plates are to be detected
  • the ALPR brightness control ROI to define in which area the camera needs to adjust the brightness of the video stream.


Note

Note Sign

  • The ALPR detection ROI is drawn as a yellow polygon.
  • The ALPR brightness control ROI is drawn as a green polygon.


How the ROIs are displayed and how to configure them is explained in the following table.

Live view Camera settings Server settings
t1 t2 t3

The OSD on the Live result page continuously displays a live image stream.

The yellow polygon represents the ALPR ROI.

The polygon can be adjusted in Settings, in the Engine page of the camera.

Both ROI for brightness control and ROI for ALPR detection can be defined in the Settings page.

The green polygon represents the Brightness control ROI. defined in the Settings page.

The Brightness control settings are only applied to the area inside this polygon. This allows you to optimize the control over areas where it is required and ignore unwanted effects.

It is recommended to limit the brightness control ROI to the image area where the vehicles might move through the screen.

The ROI found in the Engine section represents the ALPR detection ROI.

It is displayed as a yellow polygon. License plates will only be detected within this area. It is recommended to make this ROI as small as possible.

View in Live results tab. ROI changes in Settings -> Camera ROI changes in Settings -> Engine


Hint

Hint Sign

Resize the polygons by clicking on any of the highlighted corners and drag it to the desired position. When finished resizing, click on the Apply button.



Hint

Hint Sign

The ALPR detection ROI should be larger than and enclose the brightness control ROI, in order to achieve better performance.



2.3   Display results

Two pages in the camera interface are used to display reading results either as a live video stream with a graphical overlay or in the form of a table:

  • the Live result page displays live images with reading results and updates the latest results in a compact table view.
  • the Text result page displays the results in a tabular form. Clicking on a list entry opens the associated image, which can also be downloaded. It is also possible to export the entire list as a CSV or an XLS document.

The result list can be limited to a certain amount of entries and/or have a time limitation. When one of those conditions is met, the oldest entries will be removed from the list.

The camera can optionally display the following information over a live image, configurable for each element:

  • state of vehicle registration,
  • state recognition confidence,
  • vehicle number plate,
  • number plate recognition confidence,
  • character recognition confidence.
./images/Image08.png

Processed image with an overlaid reading result.

In the example image above, AT stands for the state in which the vehicle is registered, and it is recognized with 68% confidence. W _ _ 4 3 3 is the vehicle number plate, recognized with 100% confidence. The confidence of recognition is displayed above each character; in this case, it is 100% for each character.

3   The Carrida 4 Camera User Interface

The Carrida 4 camera user interface gives you access to all Carrida features which are supported on your camera. The software is based on modern web technologies so that you can use a standard web browser to access the user interface from any standard browser, which is installed on Windows, Linux, Apple IOS or a Mac.

By using the Carrida discovery tool (see Carrida Camera Discovery) you can find all cameras which are active on your local network and with the click of one button access any of them in the browser.

The rest of this document describes the Carrida 4 user interface in detail.

4   Logging into the Carrida camera

Before the camera can be accessed over the web browser, you have to log in into the device, very similar to other web services.

The Carrida user interface offers several levels of access control with different sets of permissions. Only the administrator is allowed to modify all settings and parameters of the camera.

The username and password you can use for logging into the camera are typically provided by the administrator of the camera(s), but a set of default users is stored in the camera (see Quick Start).



Note

Note Sign We strongly recommend changing the default passwords for the user accounts before or during the installation of a camera.



You can access any camera in your network by typing its IP address in the browser address bar. This brings you to the login page of the camera.


./images/1_login.png

Login page of Carrida Camera Configuration


Submitting a valid username and password opens the homepage as illustrated in the image below.

./images/1_home.png

Homepage of Carrida Camera Configuration


4.1   The User Menu

Located in the upper right corner of the homepage is the User menu, which allows you to easily switch between user profiles. For example, you can switch between a regular user and an administrator here.

./images/0_logoutbutton.png

4.2   Camera System Information

The camera information section of the home page displays a compact overview of the current software version and classifier currently used by the camera.

Classifiers contain pre-trained data sets, which are used to enable the recognition of characters and symbols for a specific country. They contain information about plate specific details like font styles, plate dimensions and valid text patterns. Detailed information about which classifier is currently in use is listed in the camera information section in Section Classifiers.

The bottom of the Home page contains the navigation links to different sections of the camera interface for quick access. Hover the mouse pointer over the icons to obtain a short description of the section functionality.

5   Live Result

The main purpose of the Live Result page is to verify the camera operation and configuration settings in real-time. To accomplish this, a live image stream from the camera is displayed together with additional information, such as reading results, confidences, etc.

The details of the overlaid information can be configured in Section Engine.

In addition to the video stream of the camera, the latest nine reading results are listed in the table on the right side of the page in order to provide a quick overview.

It is possible to display the video stream in full-screen mode by clicking on the green button in the lower, right corner of the live image. You can leave the fullscreen mode by pressing the ESC (escape) key.

The yellow polygon overlaying the image represents the region of interest (ROI) within which the camera reads the license plates. The camera will only process license plates that are detected inside of this area. You can add and define the positions of up to 5 regions of arbitrary shape in the image. The polygons can be added and/or adjusted on the Settings page, under the Engine tab (see Section Engine).

./images/2_liveresult.png

6   Text Result

On the Text Result page, all reading results of the camera are displayed in a table view. The table consists of

  • a Timestamp with the date and the time when the plate was detected,
  • License plate information,
  • the Confidence for the given license plate recognition,
  • the detected Country of the vehicle license plate,
  • the Confidence for country recognition.

Note

Note Sign The result table is generated when the Text Result page is opened, it is not automatically updated. To update the table while watching it, click on the blue Reload button on the right side below the table.

The number of results which are stored on the camera is configurable through the History List Length (max 10000 entries), and the History Hour (max 10000 hours) input fields at the bottom of the page.

Note

Note Sign History List Length and History hour are, by default, set to 0.

When either the defined time limit or the maximum number of stored results has been reached, the oldest results are discarded and overwritten.

./images/3_textresult.png

ALPR Text results.

6.1   How to browse the text result table

Navigation through a table with too many entries to fit in one page is possible by using the buttons below it:

  • click on the Previous or Next buttons to switch pages
  • directly jump to a page of the text results table by clicking the page number box
  • type a page number into the GoTo Page field and click the right arrow

6.2   Searching the table

The result table is fully searchable, it can be queried for any given piece of information like plate number, date or time.

The simplest method is to type the text you are looking for (e.g. the license plate string or a part of it) into the search field in the right upper corner of the page.

Note

Note Sign

All numbers in the result table are considered to be text for the purpose of the search function.

You can search for multiple strings by separating them with a blank character in the search field. In this case, all of the strings must be contained in a reading result.

Example: if you type 'BR 80' the text result field must contain the two text fragments BR and 80.


./images/3_textresult_search.png

Search field for ALPR text results.


The search results from above can be further refined by restricting them to a

  • Date range
  • Time range
  • Plate confidence
  • State confidence

The filter option dialog opens by clicking on the Filter button above the Search field. If you provide only one of the parameters in the parameter field (e.g. only the starting date or the lower confidence value), the other parameter is set to a value which maximizes the search range (e.g. 100 for the upper confidence).


./images/3_textresult_filter.png

Example:The combination of the search field and filter options allows you to look for license plates which contain the string 'BAC3', which have been detected between the 3rd and the 5th of April 2019, between 8 am and 8 pm, with a minimum confidence of 80 as shown below:
./images/3_textresults_filteredsearch.png

6.3   Exporting the table

The content of the filtered text result table can be exported from the camera in several different formats to your PC. By clicking on one of the three blue/white buttons on the top left side above the text result list, you can choose to export as:

  • CSV file
  • XLS file readable by Excel
  • ZIP archive which contains all images. The naming format of the images is <timestamp_platestring_confidence_state_confidence.jpg>

The export buttons are shown below:

./images/3_textresult_export.png

6.4   Images linked to text results

The right side of the Text Result page is reserved for images related to table entries. To display an image related to a reading result, click on the desired entry in the table.

Individual images can be downloaded by clicking on the Download button above the image.

Right below the image, actions and their result states which are connected to this image are listed. This provides useful information about which concrete action was defined for this reading result, and if the action has been successfully performed.

./images/3_textresult_download.png

7   ACCESS Configuration

The ACCESS configuration page of the Carrida camera provides the tools to define what the camera has to do when a license plate has been detected. For example, an action could send the image and license plate string to an ftp server when its string has been matched with an entry from the Blacklist.

It is useful to understand the following terms used by Carrida:


Events
An event is a state change within the camera. This can be the reading of a new license plate (LP), not detecting a LP within an externally triggered reading operation, a LP moving in a wrong direction, or a LP with low confidence, which is possibly of bad quality.
Events may also be tied to a LP string, which is defined in one of the three lists which are supported by the camera.
Events can trigger actions - this is the basic mechanism used by the camera to do something when a LP has been (or has NOT been) detected.

Lists
provide the means to define which license plates can trigger actions. The three categories of lists Whitelist, Blacklist and Ignorelist can each have their own associated actions triggered.

Actions
tell the camera what to do when a specific event happens. Actions can either be triggered by general events, e.g. the detection of a new license plate, by specific events related to lists or they can be related to external trigger lines.
Actions allow the user to make full use of the I/O capabilities of the camera, including the network interface, I/O lines and the internal flash storage.
As mentioned above, specific Whitelist match/mismatch, Blacklist match/mismatch and Ignorelist match/mismatch events may be used to trigger actions. In this case, the action will be executed when a plate from the list is detected (match) or when a plate is detected but is not found in the list (mismatch).

The ACCESS configuration page provides two tabs (Actions and Lists), which are used in conjunction to set up the camera as needed. The figure below shows the Access page with the two menus for actions and lists:

./images/5_actions.png

Access control configuration page.


7.1   ACCESS lists

The three access lists available on the camera allow you to specify three sets of license plate types, each of which can perform its own actions.

License plate entries may either be defined as full strings, defining only one specific license plate or as patterns (regular expressions). In the latter case, you can define to match a whole set of strings with one entry in the list, e.g. match all license plates which start with the character 'B'. These patterns are called Regular expressions and are explained in detail in Section Regular Expressions below.

The three types of Access Lists of the Carrida camera are:

  • Whitelist,
  • Blacklist
  • Ignorelist.

Every time the camera reads a new license plate, it compares its string with all of the entries stored in each of the lists. If an entry matches, a previously defined action, if specified, is triggered and performed. Likewise, if the newly found license plate does not match and an action is defined for that case, that action will be performed.


There are four types of events which may trigger actions:
  • Events are considered to be Read events when the camera reads any license plate.
  • NoRead events occur when NO plates are detected when the camera is configured to be triggered by an IO line. For example, if Start and Trigger timeout is defined and no valid plates are detected, the camera will perform a NoRead action if such one is defined. Triggers are explained in detail in Section Trigger. NoRead events can also be produced periodically in the Free_run Trigger Mode. For any interval of length defined by NoRead Event Interval (frames) without any plate readings, any defined no-read action will be executed.
  • Wrong Direction read events occur when a license plate is detected with a motion direction which is not within the allowed direction range defined for the area within which this plate has been detected. The areas and associated direction ranges can be defined in the ALPR ROI editing page, see Section ALPR Region of Interest.
  • Low confidence read (LCR) events - a Low confidence read event is created when the plate is read, but the confidence level of the reading is lower than set in the parameter Limit of Low Confidence. This type of event can be used to collect low-quality readings for further analysis.

./images/4_listcat.png

A list entry contains at least the plate number, but it can optionally also contain a date range and a time range. Only matches within the defined date and/or time range will be accepted as valid.

There are two ways to create or update the lists: they can either be created directly in the dialog window in the camera interface or they can be imported as text files from the PC. The latter method is more efficient if a large number of entries has to be created or if an external database of license plate entries (e.g. employees of a company) already exists and needs to be stored on the camera. Both methods for access list creation are described in the following sections.



7.2   Manual creation of a list entry

A new list entry can be directly created using a dialog in the camera's interface. Click on the green '+' button in the lower right corner of the Lists page to open the Add List Entry dialog.


The first entry in the dialog box is the plate number. When Always enabled is turned ON, a match will not be restricted to a specific time or date range. When Always enabled is turned OFF, time and date limits apply. These can be defined through the date and time fields below.

The time window can be defined through the start and end date, start and end time, and on individual days of the week (marked with So, Mo, Tu, We, Th, Fr, Sa). By default, the time span is set to 24/7, Monday to Friday, for a month from the current date.

In the example on the right side, Monday-Friday is checked, meaning that if an LP is detected on any working day between the set dates, a certain action will be triggered.

After finishing the list, click on Add (the green plus button) and Save.

ls

Note

Note Sign Each plate can have multiple date/time windows attached (e.g. * Monday 8am-12am* and Wednesday 2pm-5pm). Use the Add button to attach a new time/date definition to the plate. After all definitions have been made, press the Save button to store the entry for this license plate in the access list.

Hint

Hint Sign The plate string may contain regular expressions, see Section Regular Expressions .



7.3   Import of list files into the camera

If a larger number of license plates needs to be imported into one of the access lists of the camera, the import of a text file with definitions may be more efficient than manually entering all the data. The external text file has to be formatted as described in the following paragraphs. First create the text file according to the formatting rules described in the next paragraph On the ACCESS page under the List tab, click on Import and select the newly created document from your PC.

During the import of the text file, all entries are checked for validity. All valid entries will be displayed on the screen after the text file has been loaded. If needed, entries can be edited directly in the table.

When the list is finished, click on Upload to finally save the list to the camera.

2s



7.3.1   Formatting rules for external files

When importing external files, the following naming rules apply:
  • Every list entry must be placed in a single, separate line without blank spaces.
  • The entry order must be license plate; date window(s); or license plate; date window(s); time window(s)
  • Time window definitions are optional, but if given, they must be defined for every date window.
  • Plate number, date and time must be separated with semicolons (;).
  • When listing multiple dates/times, separate them with a comma (,).
  • Date format: YYYYMMDD-YYYYMMDD, with no space between the numbers.
  • Time format: HH:MM-HH:MM. Time is always defined in combination with the date and a symbol (0/1) for the preferred day of the week.
  • Days of the week are represented with zeros and ones (0-not active, 1-active) written inside the paranthesis starting with Sunday, ending with Saturday. For example, "0111110" non-active days are Sunday and Saturday, represented with zeros (0).
  • The number of dates set must be equal to the number of time restrictions ( W223344;20170101-20170101,20170102-20170102,20170103-20170103;03:00-04:00(0111110), 03:00-04:00(0111110),03:00-04:00(0111110) ). If there are more dates than time restrictions (and vice versa), the definition of this entry will be regarded as invalid.


Example: W223344;20170101-20170125;03:00-04:00(0111110) an action will be taken if the vehicle with plate number W223344 is detected from 01.01.2017 (03:00) until 25.01.2017 (04:00) between 03:00 and 04:00 on any day of the week, except Saturday and Sunday (0111110).


Practical examples and descriptions for the rules:

  Example Description
Only plate number W223344 Actions will be taken when a vehicle with this number plate is detected.
Plate number and date   Date format: YYYYMMDD-YYYYMMDD, no blanks.
W223344;20170925-20170926 Actions will be taken if the plate is detected on a date inside set limitations ( 20170926<=[current date]<=20170925).
W223344;20170925-20170926,...,20170529-20170629 Multiple dates can be set, separated by a comma defined in one line, without blanks.
Plate number, date and time   Time format: HH:MM-HH:MM(0111110); 1 and 0 represent days of the week, starting from Sunday(0) till the last day Saturday(0), 0 meaning not active, 1 meaning active.
W223344;20170925-20170925;03:00-04:00(0111110) Time must be set in combination with the date.
W23344;20170925-20170925,201701012-20171112,...;03:00-04:00(0111110),05:00-06:00(0111110 ),...
When there are multiple dates, if time is set for one date it must be defined for all the dates or the limitation will be found invalid.
In this example, two dates with their respective times are set where date(1) 20170925-20170925 is combined with time(1) 03:00-04:00, and date(2) 20170112-20171112 is combined with time(2) 05:00-06:00 and so on.

To trigger actions when any plate is detected, use the regular expression !* as list entry:

The action is triggered for all detected license plates pc

7.3.2   Regular Expressions

Regular expressions are used to conveniently define patterns for license plates or groups of license plate strings.

Carrida uses a powerful set of regular expressions, of which the most important details are described in the following paragraphs.


Note

Note Sign Regular expressions can be used in the dialog for list entry of the camera or in the list text files which are imported. In both cases, a license plate number which is made up of a regular expression must be preceded with a "!" character, e.g. !B1*



Hint

Hint Sign Details to all features of the Carrida regular expression syntax can be found here: https://www.boost.org/doc/libs/1_61_0/libs/regex/doc/html/boost_regex/syntax/perl_syntax.html



In Carrida regular expressions, all characters match themselves (example: A is always A, B matches B, etc.) except for some special characters:

.
[]
{}
*
+
?

The '.' character matches every single character. E.g. the regular expression

G.23456

will match

G123456, or
GU23456, etc.

(In the above expression, the "." is matched by 1 or U respectively)



A character range is defined by a list of characters enclosed in []. For example [A-D] will match any single character in the range 'A' to 'D'. This character range may also be defined as [ABCD].

An atom is defined as being a character or character range.

A single atom can be repeated with the * , +, ?, and {} operators.



The * operator will match the preceding atom zero or more times. For example, the expression

A*B

will match any of the following:

B
AB
AAAAAAAAAB


or the expression

A[BC]*D

will match

AD
ABCD
ACCCBBD

(between A and D, the characters B and C as defined by the atom [BC] may be repeated any number of times.


Note

Note Sign A regular expression beginning with a * is invalid because the * operator does not precede any other atom. Use the combination .* instead!


The + operator will match the preceding atom one or more times. For example, the expression

A+B

will match any of the following:

AB
AAAAAAAAB


But will not match:

B


The ? operator will match the preceding atom zero or one time. For example, the expression CA?B will match any of the following:

CB
CAB

But will not match:

CAAB


An atom can also be repeated with a bounded repeat, where the number of allowed repeats is defined as part of the regular expression:

A{n} Matches 'A' repeated exactly n times.

A{n,} Matches 'A' repeated n or more times.

A{n, m} Matches 'A' repeated between n and m times inclusive.



For example:

A{2,3}

Will match either of:

AA
AAA

But neither of:

A
AAAA


Following are some useful examples for regular expressions and use cases:



To match any license plate string of at least length 1:

.+


Match license plate strings which consist of digits only:

[0-9]+


Match all taxis, under the assumption that taxis have a license plate string that ends with TX:

.+TX


Match all license plates that start with a letter and end with a digit:

[A-Z]+.*[0-9]+


Match all license plates from the Carrida company, assuming that their license plates consist of a CAR string and a following 3 number digit (like CAR001):

CAR[0-9]{3}


Note

Note Sign When making a list by using regular expressions in either the camera dialog or the text file used for import, the definition for the regular expression must always start with a "!" character.


7.4   Actions

Once the camera detects a license plate, it checks whether some action has to be taken based on this reading. The camera checks also regularly for other state changes (for example a signal on a digital IO line) which can create an internal event. As mentioned earlier, there are several types of events which can trigger actions:

  • Read events occur when the camera detects any license plate.
  • NoRead events occur when NO plates are detected in a triggered reading mode. The use of external triggers is explained in detail in Section Trigger.
  • Wrong Direction read events occur when a new license plate is detected, but it moves in a direction which is not within the allowed direction range defined for the area in which this plate is located.
  • LCR events - a Low Confidence Read event is created when the plate is read, but the confidence level is lower than set in the parameter Settings-Access control-Limit of Low Conf.
  • Whitelist match, Blacklist match and Ignorelist match - the license plate is found in one of those lists.
  • Whitelist mismatch, Blacklist mismatch and Ignorelist mismatch - the license plate is NOT found in one of those lists.

7.5   How to setup Actions

The definition of actions is done from the Actions menu of the Access page. Here you can choose which action should be performed based on the event type.

./images/5_actions.png

Access actions categories.

The following possible actions can be created (details follow below):

  • Ftp Upload: Create a new file from the reading result and transfer it over ftp.
  • Ftp Database: Update a database file on an ftp server with the new reading.
  • Store: Save the current image in a specified location on the camera.
  • Digital out: Set a digital output line of the camera for a defined time period.
  • Serial: Format and send a message over the serial port of the camera.
  • TCP: Format and send a TCP message.
  • HTTP: Format and perform a request on a specified HTTP server.
  • LEDBlink: Activate the LEDs on the camera for a specified time interval.

7.6   Adding Actions

To add an action, first choose the event category for which the action is to be added, then press the ADD button to open a dialog window for configuration. A specific action can then be selected and configured in the dialog window, as shown below.

./images/5_actions_add.png

7.7   The use of Placeholders

All actions which create dynamic results (like TCP messages) can use placeholders for actual values of an event (like the current license plate string or the current date/time). This allows you to create dynamic action messages.

The concept of Placeholders is used to specify the text elements which are needed in order to generate your dynamic text. Placeholders have the form %placeholder_element%, this text in your message will be substituted by the actual element it represents (like the license plate number) when the action is performed.


Example:

A placeholder for the license plate is %LP%, for the date is %DAY% %MONTH% %YEAR%, level of confidence for plate recognition is %CONF%.

The camera detected a vehicle with number plate W223344 on the 03.08.2017, with confidence 98%.

The dynamic text

Detected %LP%(%CONF%) on %DAY%/%MONTH%/%YEAR%.

would result in the actual created text

Detected W223344(98) on 03/08/2017.


A complete list of placeholders can be found in the Section Placeholders for naming patterns.

The following sections describe in detail all possible actions that can be triggered by an event.


7.7.1   Ftp Upload

f

Ftp upload is the most commonly used action. It is used to transfer image files of detected plates to a remote ftp server.

Enter the Host ftp server IP address in the form of (s)ftp://xxx.xxx.x.xxx/[...].

Username and Password of the ftp server must be written in the form of [username]:[password]. Example: for the username "user123" with the password "pass123" the correct form is: user123:pass123.

Use placeholders to include specific dynamic information (date, time, plate...) in the file name.

Images can be uploaded as:

  • Empty file- the filename contains all information, most efficient
  • Full image- needs more bandwidth
  • License plate image only- bandwidth efficient

Empty file sends files with no images. This saves memory and increases the upload speed.

Image quality refers to the compression rate, it can be set from 1 to 100%, 1 being the maximal compression level which will produce images with the lowest quality, and 100 being the minimal compression level with the highest image quality. We recommend setting this parameter to 65-70%.

Enable the action by sliding the Enable button to the right.


7.7.2   Ftp Database

4

This action creates and automatically updates a text file on the specified ftp server.

Enter the Host ftp servers IP address in the form of (s)ftp://192.168.3.127/[...].

Username and Password of the ftp server should be written in the form of [username]:[password]. Example: for the username "user123" with the password "pass123" the correct form is user123:pass123.

Filename stands for the name of the database to be updated (created) with a new result string.

Use placeholders to include specific dynamic information (date, time, plate...) in the result string.

Enable the action by sliding the Enable button to the right.


7.7.3   Store

Warning

Warning Sign

This action is recommended for expert users only!

5

This action stores image data on the camera. Use this action type to collect images for training, and only if you are an expert user! Currently, the filename is fixed as a combination of current date and time, and a user-defined string ('postfix'), which may be appended. Only the storage path on the camera's file system may be chosen. The filename is therefore constructed as

path/date-time_postfix.png

The postfix field may be left empty.

Note that we recommend using the file path /root/data as a storage location on the camera.

The Free space parameter defines the minimal free disc space required to be available on the camera in order to allow the action to perform. If there is less space available on the camera than defined here, the files will not be saved.

Enable the action by sliding the Enable button to the right.


7.7.4   Digital Out

6

This activates a specified digital output line on the camera for a specific time period (Output duration).

Select which Output pin will be activated by the action.

Output duration defines the duration of the output signal in ms.

The Invert signal option can be disabled/enabled in order to invert the polarity of the IO line.

Enable the action by sliding the Enable button to the right.


7.7.5   Serial

8

Construct a message using strings and placeholders and send it through the serial port of the camera with the specified Baud rate.

Enable the action by sliding the Enable button to the right.


7.7.6   TCP

9

Write a message for the action using TCP with text and placeholders (Section Placeholders for naming patterns).

Enter the IP address of the receiver of the TCP address and the Port used to send the TCP message.

Enable the action by sliding the Enable button to the right.


7.7.7   HTTP

ö

Run an HTTP action with the specified Request Type.

Define the URL and enter the credentials, username and password for the URL.

In patterns define the name of the file using the guidelines in Section Placeholders for naming patterns

Enable the action by sliding the Enable button to the right.


8   Settings

The Carrida camera provides the necessary configuration parameters to optimize its operation for different application scenarios. Modification of the settings for image acquisition, detection sensitivity, etc. ensures that the camera is adjusted properly to avoid common problems like motion blur or low contrast.

The engine settings allow you to optimize the OCR performance.

8.1   Overview

The Settings page divides the available camera parameters into several logical menus, which can be opened and closed by clicking on the triangle next to the menu name.

As shown in the image below, the settings are divided into 7 tabs (Camera, LED Flash, Trigger, Engine, Mode of operation, Access Control and Display Information). The tabs are opened by clicking on the right-pointing arrow. In the example image, all menus except Display Information are closed.

./images/8_settingsdropdown.png

Access actions categories.


Hint

Hint Sign

There are three types of control parameters used for configuration:

  1. Drop-down menus to choose from a predefined list of options.
  2. Input fields where users can enter values within a certain range.
  3. Enable/disable options. They can be changed by clicking on either the green checkmark (enabled) or a red cross mark (disabled).

All changes are confirmed by clicking on the Apply button; a green circle with a white, right-pointing arrow.


The following sections describe each of the menus in detail.

8.2   Camera

The Camera settings page is used to optimize the imaging performance of the camera. It allows the user to modify brightness control parameters, and define gain settings and exposure times. How to adjust these parameters properly and what they mean is described in the following paragraphs.

Gain and Shutter options directly impact image acquisition. Gain, expressed in decibels, represents the camera's sensitivity to light, and Shutter speed measures the exposure time in microseconds.


What does the green polygon in the live image represent?

The area inside the green polygon is called the brightness control region. It defines the area within which the camera measures and corrects the image brightness by controlling exposure time and gain.

It is recommended to limit this brightness control area to the actual image region where license plates are expected to be visible. This maximizes the effect of the control parameters and allows the camera to achieve the best image quality in all possible scenarios.

How to adjust the brightness control polygon and for more information about it, see Brightness control Region section.


What is the shutter resp. exposure time?

A camera shutter is a mechanism that lets the light in through the lens onto the sensor of the camera. It has two states- open and closed, and the time it stays open is measured in microseconds (μs) in the Carrida camera.

Fast shutter speed means the shutter will open and close quickly, limiting the amount of light that can hit the sensor. A slow shutter speed resp. long exposure time means that more light can pass through to the sensor, producing brighter images.

The maximum allowed shutter time gives you control over motion blur, should vehicles move too fast. A shorter shutter time reduces the chance for motion blur on moving vehicles.


The following four parameters define the limits for gain and exposure time as used by the camera's automated brightness control algorithm. When these limits are set to extreme values, the camera may be able to capture optimally exposed images, but possibly at the cost of reduced image quality.

For example, if the gain is allowed to become too high, images may become too noisy for detection, or if the exposure time is becoming too high, motion blur may be visible.


8.2.1   Minimum Gain (dB)

Gain controls the amplification of the image signal from the camera's sensor, and thus affects its overall light sensitivity. A higher gain setting boosts the image brightness at the cost of introducing noise. A lower gain setting results therefore in general better image quality.

The Minimum Gain parameter defines how low the camera can go with the gain control when it tries to adjust the image brightness. For best image quality it is recommended to set this parameter to 0.


8.2.2   Maximum Gain (dB)

The Maximum Gain parameter defines how high the camera can go with the gain control when it tries to adjust the image brightness. The higher the camera can turn up the gain, the more noise can be introduced into images, thus making them harder to read.

We recommend a maximum gain value of 10 for the Carrida Camera.


8.2.3   Minimum Shutter

The Minimum Shutter parameter defines how short the shutter time can become when the camera tries to adjust the image brightness. The Minimum Shutter time must always be lower or equal the Maximum allowed shutter time.

The lowest possible value for Minimum shutter time is 10 μs, but we recommend to set a value of 50 μs, especially when the camera is located indoors.

Note

Note Sign If the Minimum Shutter time is set too high, the camera may not be able to adjust the image in very bright illumination conditions. E.g. if the Sun is directly behind the camera and its light is being reflected from license plates.

8.2.4   Maximum Shutter

The Maximum Shutter time defines how long the shutter can stay open when the camera tries to adjust the image brightness. The Maximum Shutter time must always be higher or equal to the minimum allowed shutter time.

If the Maximum Shutter time is allowed to become too high (e.g. allowing up to 20000 microseconds in an urban environment), vehicles which move too fast are likely to become blurred.

If the Maximum Shutter time needs to be reduced to avoid motion blur, and the resulting image brightness becomes too low, try the following measures to increase it:

  • set a longer LED flash time (see Section LED Flash for a description).
  • try to increase the power of the built-in LED flash by configuring more LED current
  • increase the maximum allowed gain

Hint

Hint Sign
As a general recommendation, the following shutter times are advised:
  • shutter time of 50-3000 μs on highways, for high-speed vehicles
  • shutter time of 50-5000 μs in urban areas, for medium-speed vehicles
  • shutter time of 100-10000 μs in parking situations, for slow vehicles
How to estimate the maximum allowed exposure time:

The maximum possible exposure time can be calculated from the viewing geometry and image resolution. The basic principle is that a vehicle at maximum possible speed should not move more than 1 px in the image during the exposure time period.

It is possible to either compute this value exactly using measurements of the visible street section, viewing angles, etc. or it can be approximated.

By measuring the time (T) it takes a vehicle from top to bottom or left to right of an image - whatever time is shorter - and the image resolution in this direction (i.e. the number of pixels) is known, the maximum allowed exposure time in ms can be computed as follows:

exposureTime(ms) = 1000*(T/resolution)


8.2.5   Low Resolution

The Carrida camera sensor can operate in two resolution settings

  1. a low-resolution mode of 640x480 pixels and
  2. a high-resolution of 1280x960 pixels

The lower resolution mode allows the camera to achieve a higher processing rate so that the faster moving vehicles are captured with higher accuracy. The disadvantage of the low-resolution mode is that the camera can read license plates only up to a maximum distance of about 5 meters.

The high-resolution mode enables the camera to read license plates up to a distance of about 9 meters, but at the cost of a 30% slower processing rate. The lower processing speed is typically uncritical in almost all situations, only when the camera is applied to freeflow like scenarios (with fast moving vehicles, this could become an issue that should be taken care of).


8.2.6   Image Rotation (deg)

In situations where the camera has an oblique view onto the street, it might be preferable to rotate the camera image after capture so that the license plates appear horizontally. This is not a strict requirement for good readings, but it can improve recognition quality. Enter a rotation angle here in order to optimize the image alignment. The effect of the rotation parameter is directly visible after you set a new value.


8.2.7   Brightness control Region

The Brightness control region is the area enclosed in the green polygon shown in the live image when the Camera menu is selected in the Settings page:

./images/7_settings_camera.png

The Brightness control area defines the image region used by the camera to continuously analyze and adjust the brightness of the next image to grab and process. To this end, the camera tries to adjust shutter time and gain setting in a way that gives the image the best possible average brightness even in difficult situations.


Resizing the Brightness control polygon
Click, drag and release the corner of the polygon. Click on Apply next to the Brightness control region option under the image to save changes.
example
In this example the camera is mounted in front of the gate. The polygon captures the path that the vehicles might take so that the brightness control area is set to the smallest possible area.

A good image should have neither too many overexposed pixels nor too many dark pixels.

Overexposed pixels are particularly bad because all information is lost, pixels which are too dark are less problematic because they still contain useful information.

All pixels with brightness levels below 100 will be considered too dark, all pixels with the levels over 220 are considered to be overexposed.

For many application scenarios, especially in entrance ways of parking garages or during night-time, the just mentioned goal of a perfectly exposed image is very hard to achieve, if not impossible. The differences in illumination encountered in those situations are just too high to be faithfully captured by a sensor. The Carrida camera brightness control algorithm therefore puts the emphasis on not overexposing images. It allows you to tolerate only a certain number of overexposed pixels and also a minimum number of good pixels that is to be reached.

In any case, lowering the image brightness to avoid overexposure has priority in the brightness adjustment algorithm.

Hint

Hint Sign We recommend setting the brightness control area to the smallest possible area so that the camera can optimize the image brightness for the relevant image region. A good idea is to synchronize the BC area with the ALPR ROI, this can be easily done by pressing the blue Sync button in the right lower corner of the live image.

./images/Image15_sync.png

8.2.8   Use Brightness control

This setting turns the automatic brightness control on or off. Our recommendation is to leave this parameter enabled at all times, as it will enhance image quality and produce better ALPR results.

In the case that the Brightness control is disabled, the values for gain and shutter will be set to specified maximum values.

Note

Note Sign

Pixel counts respective percentage measures are indicated relative to the Brightness ROI polygon size.


8.2.9   BC max bright pixels (%)

During automatic exposure control, the camera counts all the 'good' pixels within the brightness control area. Good pixels are neither too dark nor overexposed as explained in the introduction above.

In cases where there are too many overexposed pixels, as indicated by this parameter in percent relative to the polygon size, the camera will change its exposure settings to make the image darker.


Hint

Hint Sign What to do if the license plates are constantly overexposed?

If the license plates are constantly too bright, the percentage value of maximum saturated pixels needs to be reduced, so that the camera tries to decrease the count of those pixels on its own.

Example: reduce the value from 5% to 3% and observe the effect on the live image stream.

Another possible cause for constantly overexposed images could be that the Minimum Shutter time is set too high, so that the camera is not allowed to decrease the brightness of the image any more. See Section Minimum Shutter for more information on how to reduce the Minimum Shutter time.


8.2.10   BC min correct pixels (%)

In order to assure good image quality the camera checks not only for overexposed images, but also for a minimum number of well exposed pixels.

This parameter defines the minimum number of good pixels as defined above, and tries to make corrections if necessary. In this situation the camera will only try to enhance the image brightness if the limits of gain and shutter value allow for. BC min correct pixels is of secondary importance for brightness control and should only be used if absolutely necessary. It may be deprecated in future versions of the Carrida software.

Hint

Hint Sign What to do if the license plates are constantly too dark?

If the plates are constantly too dark, the value of minimum correct pixels needs to be increased, so that the camera is forced to increase the image brightness.

Example: modify the value from 3 to 5% and observe the effect in the live image.

Another possible cause for constantly underexposed images would be setting either the Maximum Shutter time or the Maximum Gain too low, so that the camera is not allowed to brighten up the image more. See Section Maximum Shutter time on how to increase the Maximum Shutter time or Section Section Maximum Gain on how to increase the Maximum Gain.

As third option try to increase the LED flash power, see Section LED Flash or more information.


8.3   LED Flash

The Carrida camera has a built-in IR flashlight which can be controlled through the LED flash settings. It is recommended to have the flash turned on at all times to optimize the contrast of license plates. In addition to the LED, the Carrida camera is also equipped with a daylight blocking filter which is tuned to the LED wavelength. This feature reduces the influence of sunlight as well as reflections from the headlights of vehicles.


The relative strength of the LED light is controlled via two parameters:
  • The LED flash duration which defines how long the LED flashes last when triggered by the camera. The flash duration may be as short as 10 microseconds and at most as long as the camera shutter time, but never longer than that.
  • The LED current which defines the total power of the flash. The flash current is set in mA and may range from 150-1500 mA. Flash current setting of less than 150 mA will turn the flash off.
./images/8_ledflash.png

LED Flash settings on the Configuration screen.


8.3.1   Enable Flash

Enabling the flash will turn the IR LED flash on.

We recommend leaving the flash enabled at all times.


8.3.2   Flash Duration (µs)

The Flash Duration parameter defines the duration for which the flash is actually emitting light when triggered by the camera. The flash is synchronized with the beginning of image acquisition by the camera. Its duration may be shorter than the current exposure time, but it will never be longer. E.g. if the current exposure time by the camera is 500 µs and the flash time is set to 800 µs, it will nevertheless be cut off at 500 µs.

As the flash duration is set to shorter times, its ability to freeze vehicles in motion gets better. For example, shorter flash durations are desirable when dealing with vehicles that are moving fast. In a typical city environment with up to 70 km/h (45 mph), a flash duration of 1000 µs would be a good choice.



8.3.3   Flash Current (mA)

Independent of the flash duration, the current which flows through the LEDs is another means to control their brightness. The more current is sourced through the LEDs, the brighter they become.

The LED controller circuit covers a current range from 150 to 1500 mA, a value of less than 150 mA turns the LEDs off.


It is recommended to set the current in a range from 500 mA to 1000 mA, so that

  1. it is bright enough even for LP at larger distances
  2. the flash is strong enough to illuminate license plates even in the presence of vehicle lights. This becomes especially important when the Carrida camera is used indoors.

Note

Note Sign

If the duration of the LED flash is set to 150 mA or less, the flash will be turned off by the controller circuit.


8.4   Trigger

The Carrida camera offers several options to control its operation through the use of digital I/O lines. These I/O signals allow the user to limit the time that the camera actually spends reading in combination with some basic timing parameters. E.g. the camera can be set up to start reading for 5 seconds after the trigger signal from a ground loop has been received.

By also creating an associated action, the camera may execute a command if a LP has been read or even if no LP has been detected at all (see Section Actions for further explanation of actions and events).

There are five different trigger modes available as shown in the following table:

Overview of Carrida cam trigger modes
Free_run read without considering any trigger signals
Single_Line read as long as the trigger line is active
Start_Stop start and stop reading when the defined I/O lines are activated
Start_Time start reading for a defined time upon receiving a trigger
Start_Sequence start reading for a defined number of images after the trigger signal

Note

Note Sign In the following section, the term session will be used to describe the time between the start signal by a trigger and the end of a triggered period, which may either be signaled through another I/O line or by a timer.

A session is the only time period in which the camera reads license plates. During an open session, it operates as defined in the reading modes, e.g. like in the non-triggered parking mode, etc.


The trigger setup is controlled in the Settings page, in the Trigger menu as shown in the image below. The following paragraphs explain the possible trigger modes, their setup and operation in detail.

./images/9_trigger.png

Trigger configuration page.

Hint

Hint Sign The electrical requirements for digital IO and triggering are defined in the VC pro Z Series Operating Manual.


8.4.1   Trigger mode

Different trigger modes are described in the following table:

  • The Setup section lists the parameters which are related to the mode and are used for its configuration
  • The Start section describes the signals which will start a session
  • The Stop section describes the condition which will end a session

After choosing the mode from the drop-down menu, press the Apply button to activate the trigger mode and set the parameters as needed.



fr Setup: No other parameters need to be defined for this mode.
Start: The session starts simultaneously with the camera. No external signal is required to start the camera.
Free run should be the default mode when an external trigger is unavailable or not required.
sl
Setup: Define Start trigger line, Start trigger polarity.
If Start trigger polarity high is enabled, the session will start when the input signal is transitioning from low to high. If Start trigger polarity high is disabled, the session will start when the input signal is transitioning from high to low.
Start: The session will start when the signal from Start trigger line is received.
Stop: The session is ended when the signal on the input line changes its state again.
ss Setup: Define Start trigger line, Start trigger polarity. In addition, define Stop trigger line and Stop trigger polarity. If Start trigger polarity high is enabled, the session will start when the input signal is transitioning from low to high. If Start trigger polarity high is disabled, the session will start when the input signal is transitioning from high to low.
Start: The session will start when the signal from Start trigger line is received.
Stop: The session stops when the signal from Stop trigger line is received.
In case a new trigger signal is received while the old session is still open, the camera closes the old session and opens a new session immediately.
st Setup: Define Start trigger line, Start trigger polarity. and Trigger timeout (ms). If Start trigger polarity high is enabled, the session will start when the input signal is transitioning from low to high. If Start trigger polarity high is disabled, the session will start when the input signal is transitioning from high to low.
Start: The session will start when the signal is received.
Stop: The session lasts for a defined time period as set in Trigger timeout. After the time passes the session is closed.
In case a new trigger signal is received while the old session is still open, the old session will be extended to a new timeout period.

NOTE: in this case, the camera does NOT generate NoRead events for the interrupted session.
sseq
Setup: Define Start trigger line, Start trigger polarity. and Sequence length in frames.
If Start trigger polarity high is enabled, the session will start when the input signal is transitioning from low to high. If Start trigger polarity high is disabled, the session will start when the input signal is transitioning from high to low.
Start: The session will start when the signal from Start trigger line is received.
Stop: The session lasts for the defined sequence length.


8.4.2   Noread Event Interval (frames)

This parameter specifies the interval of images without any plate readings before a NoRead event is triggered. This parameter is only used in freerun mode.


8.4.3   Start Trigger Line

This parameter defines a start trigger line which begins the session.

When the input line on the camera is activated/triggered, it will send a signal to the camera to start the session.

An overview of active input and output pins can be found in System settings.


8.4.4   Start Trigger Polarity High

This parameter defines the active polarity of the input line in order to start a session.

If Start trigger polarity high is enabled, the session will start when the input signal is transitioning from low to high.

If Start trigger polarity high is disabled, the session will start when the input signal is transitioning from high to low.


8.4.5   Stop Trigger Line

This parameter defines a stop trigger line which ends a session.

When this pin is activated/triggered it will send a signal to the camera to end the current session.

An overview of active input and output pins can be found in System settings.


8.4.6   Stop Trigger Polarity High

This parameter defines the active polarity of the input line in order to stop a session.

If Stop trigger polarity high is enabled, the session will stop when the input signal is transitioning from low to high.

If Stop trigger polarity high is disabled, the session will stop when the input signal is transitioning from high to low.


8.4.7   Trigger Timeout (ms)

This parameter defines time duration (in milliseconds) of one session. The session starts when Start trigger line signal is received and ends after the defined time passes.


8.4.8   Sequence length (frames)

This parameter defines the number of frames which will be processed in one session. The session starts when Start trigger line signal is received and it ends after a defined number of frames has been acquired and processed.


8.5   Engine

The Engine page controls all parameters necessary to set up the actual reading of license plates. The provided default values will usually result in good general reading results, but they can be modified to tune the camera operation to specific situations.

./images/10_settings_engine.png

8.5.1   ALPR Region Of Interest

The ALPR Region of Interest (ROI) is the area enclosed in the yellow rectangle shown on the live stream. The ALPR ROI defines the image region used by the camera to process the images. The ALPR engine will only search for plates within the ROI borders.

An ROI polygon can have an arbitrary number of contour points, but it must contain at least 3 points. The shape of the ROI polygon can be modified using the following actions:

  • reshape the ROI by selecting one of the polygon points and dragging it to the desired location.
  • add more points by double clicking on an existing point, this will create a new point. You can then drag the point to a new image location.
  • remove a polygon point by right clicking on it.

Additionally, more than one region of interest can be defined by writing the number of desired regions in the Area Count field ALPR settings window.

./images/10_settings_engine_alpr.png

The name, direction and range parameters define which motion direction of a detected license plate is accepted within the polygon area. A sector of allowed motion has a center direction in degrees and a sector with or range in degrees around the center direction.

In the example below a direction from top to bottom of the image with a range of 90 degrees is defined. This means that only license plates which move from top to bottom +-45 degrees are being accepted.

./images/direction_range.png

Note

Note Sign

It is only possible to compute a motion direction if at least two detections of a license plate have been made. In the case that the minimum number of detections is set to 1, the situation may occur that only one instance of a license plate has been detected as valid. In this situation the license plate is accepted regardless of the direction and range specification of the ROI.

If the direction evaluation of a license plate is mandatory in your application, the minimum number of detections has to be set to, at least, 2.


8.5.2   Reading mode

The Carrida engine can operate in several performance modes, which differentiate by the recognition/processing speed and reading accuracy. The faster modes offer a higher reading speed on the camera at the cost of lower reading accuracy and vice versa. The Carrida camera offers the following reading modes, ordered from fastest to slowest:

  • Fastest
  • Fast
  • Standard
  • High
  • Best

The modes are explained in detail in the table below.


  Mode Description
mo Fastest Fastest possible processing speed with lower recognition quality than in other modes. Recommended for open traffic where recognition speed has priority.
Fast Fast processing speed with slightly better recognition quality than in the fastest mode. Also intended for open traffic situations.
Standard A good compromise between speed and reading performance, recommended to be used in most situations.
High High mode is about half the speed compared to fast mode (e.g. from 50 ms to 100 ms), the reading accuracy of the high mode is typically 2-3% better than the fast mode.
Best The best possible reading quality. Can be up to two times slower than in the fast mode. At the same time, reading confidence will be about 3-4% better.

8.5.3   Min letter height (px)

Defines the minimal height of characters, in pixels, which are to be recognized. The camera will only accept characters larger in height than set with this parameter. This value should not be less than 8 pixels.

Use this parameter if license plates above a minimum size should be read- e.g. by enforcing a value of 16 px, license plates which are too far away from the camera will be ignored.


8.5.4   Max letter height (px)

Defines the maximum height of characters, in pixels, which are to be recognized. The camera will only accept characters smaller in height than set with this parameter.


8.5.5   Max plate angle (deg)

Should be used if there is a chance that the vehicles will appear rotated relative to the camera. This parameter allows adjusting the range of accepted angles.

The best results are achieved when Max plate angle is between 5° and 20°, we recommend a maximum value of 40°.

Try to avoid rotated LPs by installing the camera so that LPs appear level.


8.5.6   Minimal plate confidence (%)

Minimum plate confidence refers to the minimum confidence level at which the camera will accept the plates as valid. All readings with confidence smaller than this value will be ignored.


8.5.7   Maximum Plates

Refers to the maximum number of plates that can be detected in a single image. Setting it to 0 means that an unlimited number of plates can be detected.


8.5.8   Minimum contrast

This parameter defines the minimal contrast to differentiate foreground (characters) from the background (the license plate itself). Reducing this parameter increases both the reading performance and the processing time. Modifying the minimum detected contrast may help read LPs in difficult lighting situations, at the cost of slower reading time.

It is recommended not to modify this parameter and use the default value provided by the camera unless absolutely necessary.


8.5.9   Scale Height/Width

The values are provided in percent of the original image size (100 = original size).

Scaling means resizing an image in percentage relative to its original size, with separated scales for width and height.

Scaling may help in situations where you have a view from the top or side at a large angle. The LP will be distorted, and to adjust for that, it is possible to re-scale the opposite image dimension to a value < 100. This would compensate for the original perspective effects.


8.5.10   Search inverted

Search inverted is an option that controls the way the Carrida engine will search for the LP in terms of contrast of the characters (black-on-white or white-on-black contrast) as follows:


Search inverted Description
Disabled The camera will only search for plates with a bright background and dark characters This is the fastest mode and it is recommended if only this type of LP can occur.
Search The camera will only look for white-on-dark license plates when no dark-on-white license plates have been detected. This option is a reasonable compromise between speed and detection rate. It is still recommended to turn it on only when necessary.
Always The camera will always look for white-on-black plates. This option is slow and should only be used if absolutely necessary.

./images/Image27.png

Example of inverted and regular plates.


8.6   Mode of Operation

Depending on the application of the Carrida camera, different methods for image acquisition and internal data processing might be needed. These configuration settings affect the way the camera handles image sequences, duplicate readings or small reading errors. For example, the camera should only read and transmit unique license plates in access control applications, so that the server in the background is not inundated with reading results.

In order to provide good starting points for parameter sets in respect to camera configuration, Carrida offers several pre-configured setups which are useful for specific application scenarios. The configuration settings are called camera modes or modes in the following sections.


./images/11_modeofoperation.png

Modes listed below are different in the way the minimum number of correct readings, buffer size, etc. are set up, but the camera is not restricted to these settings. By changing individual parameters, you can adjust the camera operation specific to your needs.

The following table describes the default setups and their typical use case, the sections following after that describe each control parameter in detail.



  Mode Usage Output
m4 Single The camera reads every single image as quickly as possible. Each input image is considered on its own and unrelated to images before and after it. Every single detected license plate.
m3 Parking

This mode is designed for access and parking control uses. Each vehicle number plate is required to be detected at least the number of times defined in the Minimum Detections parameter.

If Minimum Detections value has been reached, the LP will be available as a detection, and it will be ignored until it is no longer visible.

Unique license plates, redundant readings are ignored.
m2 Offline

In this mode, the results are available with a delay of about 25 frames. The reading quality is the best in this mode as the engine has more frames to analyze before producing results. The main purpose of this mode is to detect LP with the best possible quality without the need for real-time operation.

Note: This mode is not recommended for real-time applications.

Unique license plates, best possible reading quality.
m1 Freeflow

This mode is designed for freeflow operation. Considering the Minimum Detections parameter, the camera tries to read as quickly as possible. Results may become available with a delay of up to 15 frames.

The difference to Parking Mode is the larger Buffer Size which increases reading quality at the cost of a higher delay until a LP is transmitted.

Unique license plates. Longer delay until LP are read, but close to real-time.


8.6.1   Minimum Detections

The Minimum Detections parameter controls how many readings of the same license plate have to be made until a definite result is accepted by the camera.

A new reading is accepted as valid if:

  • The number of character differences to previously read strings is not greater than Maximum character Mismatch, see Section Maximum character mismatch.
  • If it is detected inside the ALPR region of interest

Note

Hint Sign Minimum Detections delays the transmission of images and readings because the camera is set up to wait until this number of correct readings has been made.


8.6.2   Maximum characters mismatch

Maximum plate mismatch refers to the maximum number of characters by which two consecutive readings are allowed to differ and still be considered identical.

The useful range for this parameter is 1-3.

Example:
If Maximum plate mismatch is set to 1, then two readings B12345 and B1245 are still considered to be identical, but not B12345 and B12456.

8.7   Access control

Access control is always enabled by default, which means that the consideration of the blacklist, whitelist and ignorelist and their related user-defined actions is always on.

./images/12_accesscontrol.png

8.7.1   Low Confidence Threshold (%)

This parameter is used to set a limit for which the reading confidence will be considered low, but still valid. The lower confidence limit must be equal to/greater than the Min Plate Confidence.

The typical use of this parameter in conjunction with the Low Confidence Read action is to detect and store low quality (= low confidence) readings for further analysis or retraining.


8.7.2   Whitespace Matching

This parameter defines whether whitespaces (' ') should be considered during list matching. This applies to the following situations:

  • matching of strings in the whitelist
  • matching of strings in the ignorelist
  • matching of strings in the blacklist

8.8   Display Information

The set of parameters in the Display Information menu allows you to configure the display of information in Live View with the Heads Up Display (HUD). The display properties are modified with the several parameters of the Display information menu.

  • HUD Size
  • HUD Position
  • Text Info
  • Draw Info
  • Draw Time
  • Draw Date
  • Display Shutter
  • Display Gain
  • Display Camera Framerate
  • Display reading Framerate
  • Draw plate
  • Draw state
  • Draw character confidence

Enabling/disabling any of these options will take immediate effect on the image display in the web interface of the camera as well as on all stored images.

The following sections describe the display options in detail.


./images/13_displayinformation.png

8.8.1   HUD Size

This parameter affects the size of the Display characters. Select between small, medium or large.


8.8.2   HUD Position

The HUD can be positioned on the top/bottom left side, or top/bottom right side of the image.


8.8.3   Text Info

Text information is an optional arbitrary string that can be displayed as a part of the HUD. Typically this field describes the location or use of the camera ('exit North').

The display of this text field is enabled with the Display info flag, see Section Display Info


8.8.4   Display Info

Display Info, when enabled (checked), displays the Text information defined in the text parameter above.


8.8.5   Display Time

When enabled, the current time will be overlaid on the result images and over the live stream images.


8.8.6   Display Date

When enabled, date information will be overlaid on the result images and over the live stream images.


8.8.7   Display Shutter

By enabling this parameter, the average shutter speed resp. exposure time (in microseconds) will be displayed.


8.8.8   Display Gain

If enabled, information about the camera analog gain (in dB) will be displayed.


8.8.9   Display Camera Framerate

Framerate refers to the number of individual frames (images) per second, which are currently retrieved from the camera sensor. Not all of those images are being processed. Depending on the processing load of the camera, some images may be skipped, resulting in the actual Reading Framerate - see next Section below.


8.8.10   Display Reading Framerate

Enabling this parameter allows information about the camera's processing framerate to be displayed in the HUD. This rate corresponds to the actual number of readings the camera is currently performing.

This number depends on the image size (low- or high-resolution), the ALPR ROI size (smaller is better) and other settings.


8.8.11   Draw Plate

When enabled, the camera will overlay the reading results over detected LPs (see Example).


8.8.12   Draw State

If Draw State is enabled, the name of the detected state, as well as the reading confidence, will be shown.

This parameter will only be considered in combination with the Display Plate flag. If Display Plate is not enabled, the state information will also not be displayed.


Example:

Enable Display plate and Display state.

A vehicle with a license plate W554433 is detected.

As a reading result, the SW overlays "AT 68 W554433 100" onto the result image.

"AT 68%" AT stands for Austria, the state, and 68% is the percentage of recognition confidence for the state. "100" stands for 100% confidence that the plate read is "W554433"(see Example).


8.8.13   Draw character confidence

By enabling Draw character confidence, the confidence of recognition for each character will be overlaid above the vehicle number plates.


Shown in the table below are examples of overlaying plate, state and character confidence.


Example:

Image (A) displays an LP image with the state and plate visualization enabled.

In image (B) plate, state and character confidence visualization is enabled.

Image (C) is generated with only plate visualization enabled.

vz

8.8.14   JPEG Compression Level

JPEG compression significantly reduces the size of the images - this results in faster data transfers (e.g. ftp uploads) and reduces the bandwidth consumption of the camera when it has to send live images to the web browser on your PC.

The level of compression can be set on the scale from 1 to 100, where a lower number means a higher compression rate with a resulting lower image quality.

Our recommendation is to set the JPEG compression level to 65, as this is a good compromise between image quality and required network bandwidth. If the camera is connected to a slow network, the JPEG compression should be reduced to a value of 50 or less.


Note

Note Sign It is a typical sign of a slow network connection if live images from the camera are not fully displayed in your browser. In this situation, the lower part of images is either sporadically or constantly missing during the live stream because the camera doesn't have enough time to transmit the whole image with the available bandwidth.


9   System Configuration

The System Configuration page allows you to set general system-related properties of the camera. The following configuration settings can be modified:

  • Network - set the IP address of the camera, network mask, gateway IP and DNS IPs.
  • Time - turn the network time protocol on or set the local timezone.
  • Users - create, delete or change user profiles.
  • Classifier - upload a classifier and view the supported state list.
  • System - view trigger input/output activity and the camera SW version.
./images/14_sysconf.png

9.1   System Actions

The System actions button is located on the upper left side of the screen. As illustrated in the image below, clicking on it reveals a drop-down menu offering a group of settings and parameters used to perform maintenance tasks on the camera:


./images/14_sysconf_systemactions.png

9.1.1   Restart ALPR

This option restarts the camera firmware including the webserver.


9.1.2   Reboot system

This option completely reboots the camera. This is equivalent to a cold start of the camera.


9.1.3   Download Logs

All camera logs can be downloaded by clicking on the Download logs button. The log files will be collected and compressed into a .ZIP file.

After unzipping the log files on your PC, you can view them with a regular text editor.


9.1.4   Save Configuration

Complete camera configuration settings can be exported from the camera and stored on your PC with the Save configuration button.


The following information is saved from the camera:
  • all Classifier files
  • date and time settings
  • network settings
  • parameter settings and region of interest settings, reading modes, etc.
  • available users and their permissions
  • all access lists
  • information about Carrida software versions on this camera
  • the serial number of the camera

Saving the configuration is a three-step process, consisting of collecting the files on the camera, creating an archive and storing it on your PC, as illustrated in the image below.

./images/20_configuration_save.png

9.1.5   Import Configuration

The state of a camera can be restored by importing a previously saved configuration file by clicking on Import configuration. This effectively resets the camera to the state of the configuration file.

It is possible to import configuration settings from another camera - this may be useful when you want to quickly copy and distribute camera setups. If you import the configuration from a camera with a different serial number you will be notified with a warning dialog.

For best compatibility, we recommend that you exchange only configuration files of the same firmware versions.


Note

Note Sign

Import has some limitations regarding the camera firmware versions:
  • version 4.3.1 cannot import configurations other than 4.3.1
  • version 4.3.2 can import configurations of version 4.3.0 and 4.3.1.

Warning

Warning Sign

A warning dialog will appear if the serial number of the imported configuration and the actual camera are different. You will be asked to choose whether to proceed with the import or cancel it.


You can choose which system settings to import by enabling or disabling features from the import dialog.

After the camera is done importing the settings, a confirmation window appears to inform you about settings which have been restored and to notify about a reboot which is necessary at this point.

./images/22_configuration_import.png

Importing server and network configuration


9.1.6   Change Camera Password

The root password of the camera can be changed with this setting.


9.1.7   Factory Reset

This option is used to reverse the camera configuration back to factory settings. All custom configuration data will be lost.


9.2   Network

This page is used to modify the following Network settings of the camera:

  • DHCP - on/off
  • IP Address - set the IP address of the camera
  • Network mask - set the network netmask
  • Gateway IP - set the gateway IP address
  • DNS IP - set the IP address
  • DNS IP 2 - second IP address, if needed

Hint

Hint Sign

The Dynamic Host Configuration Protocol (DHCP) is a network management protocol used on TCP/IP networks where the DHCP server dynamically assigns an IP address and other network configuration parameters to each device on the network. A DHCP server enables computers to request IP addresses and networking parameters automatically from the Internet service provider, reducing the need for a network administrator or a user to manually assign IP addresses to all network devices. In the absence of a DHCP server, a computer or any other device on the network needs to be manually assigned an IP address.

Any of the above parameters can be changed by clicking on the Edit button, as shown in the image below:

./images/15_sysconf_network.png

System configuration for network settings.


Note

Note Sign

If DHCP is turned OFF, manual network settings for all fields have to be provided in order to set up networking on the camera.

9.3   Time

The Time menu allows you to set the current time on the cameras as well as the local timezone.

The camera can use the NTP protocol to synchronize its internal clock with an NTP time server. The camera brings along a general list of NTP time servers, but you can customize the server list according to your needs. A click on the Edit button brings you to the edit dialog with the list of active NTP servers:

./images/16_sysconf_time.png

In order to manually set the camera's time, NTP has to be turned OFF. Then use the Set Time field to manually enter date and time to program the internal clock of the camera.

9.4   Users

The Carrida camera allows you to manage a number of user profiles, which are differentiated by the level of access to configuration and system settings. By default, each user has the following sign-in credentials:


  Username Password Description
Administrator admin secret Immediate access to all settings.
Basic User user simple Access to the most commonly used settings.
Setup setup setup The same access level as "viewer", plus changing the ROI and the classifiers.
Viewer viewer pass Permission to view reading results and live image streaming.
SuperUser su car0000rida
This is a default password and it should be changed on the first start of the server.
Access to all settings, deleting admin users and access to UI Design setting.

The addition of new users, deletion of old ones or the password changes can be performed in the user menu.


Note

Note Sign

The Users dialog is only accessible in advanced modes (Administrator and SuperUser).


After clicking on the Create new button, a dialog box will appear, as shown in the image below, allowing you to create a new user profile on the camera.


./images/17_sysconf_users.png

Note

Note Sign

The camera password is used for the internal system software, and it is not to be confused with the user passwords mentioned in this section.

9.5   Classifier

Classifiers are files that contain state- and country-specific information about license plates. This information can, for example, be the size and the font of letters and numbers, the alphabet used, etc.

Classifiers are created specifically for different regions of the world (Europe, USA, Thailand...) by Carrida.

The Carrida camera stores classifiers as a part of its firmware, and this list of classifier files may be updated and expanded by the user. The Classifier Menu shows you the currently stored classifier list on the camera in the Available Classifiers section (see also the next image below).


Note

Note Sign

Only one classifier may be active on the camera at a time, but a classifier can contain information about an arbitrary number of countries (e.g. all countries of the EU).


The Classifier menu of the Configuration page allows you to upload new classifiers, delete old ones or activate any classifier which is stored on the camera.

The dialog to modify classifier parameters can be called by clicking on the blue pencil icon on the right side of the Classifier menu:


./images/18_sysconf_currentclassifier.png

9.5.1   Switching Classifiers

As mentioned above, the camera can store a large number of classifiers as a part of its firmware and those classifiers can be switched anytime.

./images/19_sysconf_switchclassifier.png

Switching classifiers


A click on the drop-down menu (illustrated in the image above) will display a complete list of the classifiers which are currently available in the firmware. By selecting a new entry from the list and pressing the green apply button, the selected classifier will be activated.


9.5.2   Uploading Classifiers

As mentioned above, the camera has the capacity to store a large number of classifiers as part of its firmware. The list of available classifiers can be extended by uploading new classifier files to the camera. In this way, you can use the camera in new regions of the world or update and improve already existing reading capabilities.

New classifiers can be uploaded by clicking on the green upload button, which brings up a new dialog window into which a new classifier can be dropped and uploaded, as illustrated below.

Alternatively, a click on the green button on the top right bottom of this dialog opens a new dialog that prompts you for a classifier file from your PC.

./images/20_sysconf_uploadclassifier.png

Upload classifiers


While uploading a new classifier, you can opt for deleting all previous classifiers by turning on the option Delete other classifiers.

A pop-up dialog which appears after the upload operation will ask whether you want to set this new classifier as new default.

9.6   System

The system menu displays information about firmware software versions as well as all available digital input and output lines and their respective states.


./images/21_sysconf_system.png

9.6.1   Libraries

This table lists all versions and build dates of the Carrida software libraries on the camera.


9.6.2   Inputs/Outputs

Outputs and Inputs section shows information about the status of the digital IO lines of the Carrida camera.

The status of the available digital input lines is displayed in the Inputs section.

For testing purposes, the outputs can be activated in the user interface by clicking on the respective circle.

10   Placeholders for naming patterns

This section describes all the special keywords that can be used as placeholders in messages used within actions. See Section Actions for a description of actions.

During the creation of a message as a part of an action, placeholders will be replaced with their respective dynamic values.

A thorough explanation of how to use the naming patterns can be found in the Naming patterns section.

Listed in the table below are all the keywords and their descriptions (highlighted keywords are new in the 4.3.2 Release):


Placeholder Description
   
"%LP%" This keyword will be replaced by the license plate value string.
"%LP_WS%" This keyword will be replaced by the license plate value string containing found white spaces.
"%STATE%" This keyword will be replaced by the state string.
"%CONF%" This keyword will be replaced by the confidence value string.
"%CONF_STATE%" This keyword will be replaced by the state confidence value string.
"%NUM_ROWS%" This keyword will be replaced by the count of rows of the license plate as a string.
"%REGION%" This keyword will be replaced by the region string (if available). Example: "Bayern"
"%REGION_CITY%" This keyword will be replaced by the city string (if available). Example: "Munich"
"%REGION_SHORT%" This keyword will be replaced by the region abbreviation. Example: "M"
"%ENDING%" This keyword will be replaced by the LP suffix (if available). Example: "Taxi"
"%ENDING_SHORT%" This keyword will be replaced by the LP suffix abbreviation (if available). Example: "TX"
"%COL_PLATE%" This keyword will be replaced by the color of the plate as a string.
"%COL_CHARS%" This keyword will be replaced by the color of characters as a string.
"%POS_X%" This keyword will be replaced by x position in the image of the left upper corner of the license plate as a string.
"%POS_Y%" This keyword will be replaced by y position in the image of the left upper corner of the license plate as a string.
"%POS_W%" This keyword will be replaced by the width of the license plate in the image as a string.
"%POS_H%" This keyword will be replaced by the height of the license plate in the image as a string.
"%ANGLE%" This keyword will be replaced by the direction angle in degrees (0..360) of the found plate.
"%YEAR%" This keyword will be replaced with the year string. (When a license plate was detected)
"%MONTH%" This keyword will be replaced by the month string. (When a license plate was detected)
"%DAY%" This keyword will be replaced by the day string. (When a license plate was detected)
"%HOUR%" This keyword will be replaced by the hour string. (When a license plate was detected).
"%MINUTE%" This keyword will be replaced by the minute string. (When a license plate was detected).
"%SECOND%" This keyword will be replaced by the seconds string. (When a license plate was detected).
"%FRACSEC%" This keyword will be replaced by a fraction of seconds string. (When license plate was detected).
"%CR%" This keyword will be replaced by the carriage return sign.
"%LF%" This keyword will be replaced by the line feed sign.
"%FULL_IMG%" This keyword will be replaced by base64 encoded full image string.
"%LP_IMG%" This keyword will be replaced by base64 encoded cropped LP image string.
"%AREA_INDEXES%" This keyword will be replaced by a string constructed with the indexes of areas where the plate is detected, separated by a comma.
"%AREA_NAMES%" This keyword will be replaced by a string constructed with the names of areas where the plate is detected, separated by a comma.



Example:

The camera has detected a German plate "TEST123" on 01.01.2017 at 10:15:20. You can form a naming pattern as follows:

" License plate value: %LP% from %STATE% " the form will be displayed as : " License plate value: TEST123 from DE "
" Filename_%HOUR%_%MINUTE%_%SECOND%.jpg " the form will be displayed as : " Filename_10_15_20.jpg"

11   General Changes in the User Interface

From version 4.3.2 on, the Carrida web interface allows you to modify some aspects of the user interface design. You can now change the colors of UI elements and you can replace the icons with your own icon sets.

These changes are only allowed for the SuperUser from the design editing page:

./images/22_design.png

The page consists of two tabs with the following content:

11.1   Style Change

In order to provide the possibility to personalize the Carrida UI, the website design is split into multiple groups. The color for each group can be edited independently. The following colors can be edited:

  • Primary and secondary colors: The color of the page details such as buttons, table headers, navigation, etc.
  • Font color: The color of the font.
  • Body color: The color of the body background.
  • Sidebar color: The color of the navigation sidebar.
  • Card color: The color of the background of the UI cards.
  • Panel color: Color of the panels contained in the access and settings pages.

Other than the colors, the SuperUser can also change the border radius of the UI elements.

The Style Tab contains a table with the current values for all properties. These values can be edited in the dialog window, which can be opened by clicking on the Edit button. The Edit dialog window contains color pickers for color values and a number input for the border radius. The effects of new values can be seen instantly in the UI after changing them.

Click on the Apply button to permanently save the changes. The camera will store the chosen values and show a message dialog with the following info: "You have to perform a hard reload of the page or clear the cache in order to see changes in the current browser".

The Reset to default values button can be used to revert the UI back to factory settings.

Use the Cancel button to exit the UI design page without saving.

./images/22_design_style.png

11.2   Icon change

Each icon used in the Carrida Web UI can be replaced with a custom version. All currently used icons are displayed in the Icons tab. By clicking on any of the icons, a dialog window pops up with the options for icon replacement. You can choose to upload a new image or to reset it to the default version.

./images/22_design_figures.png

After changing the icon, a message dialog window with the notification about the status of the changes will pop up.

In cases where the change was successful, a hard reload of the camera UI webpage or a cache clear should be performed to see the changes in your browser.