Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Note
titleAttention!

To ensure the correct CARMENAR-Auto module operation, do the following:

  1. Make sure that "KB3033929" security update and Hotfix are installed.
  2. If Carmen GX drivers are installed, then they should must be uninstalled from Windows. 
  3. Make sure that there is not a single folder with any drivers left at C:\Windows\System32\DRVSTORE\ (if there are folders, they should must be removed).
  4. Restart the computer.
  5. Download the  Carmen Freeflow 7.3.1 distribution kit from the  AxxonSoft  manufacturer's website and install it. If an older version was installed earlier, then it must be uninstalled before installing a newer version.
  6. Download the distribution kit with the license plate recognition algorithm from the AxxonSoft algorithm from the manufacturer's website and install it via the Engine Manager utility. The procedure for installing the recognizer using the Engine Manager utility is described in the engine_install_windows.txt file located in the folder with the corresponding recognizer.

Setting up the CARMENthe AR-Auto program module is performed in the following way:

  1. Go to the CARMENthe AR-Auto module object settings panel, which is created on the basis of the LPR channel object.
  2. Enter the value of minimal minimum validity of recognition in percent in the Minimum validity of LP type recognition field (1). Validity is defined determined by the degree of similarity between results of recognition and the LP template and it allows distinguishing the LP from other marks unrelated inscription on the vehicle.

    Info
    titleNote.

    The validity is displayed in the List of parameters of the Active monitor dialog box.


  3. In the LP delimiter field (2), enter symbols which will separate LP symbols from region symbols. Maximum number of delimiter symbols 2. If a recognized license plate is to must be written without any delimiters, leave the LP delimiter field emptyblank
  4. In the Recognition frame rate, fps field (3), enter the number of frames per second processed by the recognizer. If you enter 0 or leave the field emptyblank, the recognizer will process the maximum possible number of frames.
  5. Click the Apply button to save the changes (4).

Setting up the CARMENthe AR-Auto program module is completed.

For the best quality of recognition, consider the following characteristic while the CARMENwhen configuring the AR-Auto configuring module:

  1. The less lower the resolution, the higher the quality of recognition. It ie is recommended to use the following resolution: 320x240 - 800x600px. It is highly recommended not to use resolution more than 1.2 Mpx.
  2. It is recommended to specify the search area. It results in decreasing of factual resolution, which go is fed to the entry on input of the recognizer. The less smaller the area, the less lower the resolution, and consequently, the higher the quality of recognizingrecognition.
  3. It is recommended to use recognition by trigger command (motion detection, the Parking mode) instead of permanent recognition. If a lot of recognizers are to many recognizers must be configured on the one server, use the Parking mode. Detailed For the detailed information about Parking mode is presented in the , see Scripts used in the Auto PSIM software package section.
  4. The less fewer recognizers are created on the one computer, the higher speed of video processing the quality of recognition, because the speed of video processing is shares shared between all created recognizers. Low speed of processing results in plate recognition delay and bad quality of recognition.

    Here's the approximate data of speed of video processing:
    For the Common characters template:

    ResolutionFPS
    320x24019
    480x36018
    640x48013
    1024x7688
    1280x9608

    For other templates:

    ResolutionFPS
    320x24012
    480x3609
    640x4807
    1024x7685
    1280x9604
    Info
    titleNote.

    These The data can differ decreasingly of frequency of move down in frequency of processed frames at different lighting conditions, on different processors, etc.