Detectors#

The detectors used in the CARS API subsystem are used to find the vehicle and LP in the image. Detectors determine the location of the vehicle and LP, allocate the corresponding BBox.

This section provides information about the vehicles and LP detectors. Detectors are specified in the configuration file /data/vehicleEngine.conf (see Appendix 2).

Detectors can be tested in the CARS_API_tester interface (see CARS Analytics. Administrator manual).

Vehicle detector#

Vehicle detectors are designed to detect vehicle in images. The list of vehicle detectors is given in Table 24.

Table 24. Vehicle detector description

Name	Description
VehicleDetectorV4	Vehicle detector of the latest version that allows you to obtain information about the position of the vehicle in the image. Detector has a number of advantages:
	- Improved performance;
	- Increased speed and accuracy;
	- The presence of an algorithm for redetection of vehicle to improve the accuracy:
	- Birdview images support;
	- Fixed issue of false positive detections for cameras set on houses.
	This version includes additional parameters configurations for vehicle and LP detectors.
VehicleDetectorV2Large	A detector for working with images on which it is necessary to detect a vehicle both near the camera and in the background. That is, in cases where the difference in the sizes of objects on one frame differs. When using another detector in this case, small objects may be lost.
VehicleDetectorV3, VehicleDetectorV2, VehicleDetectorV1	Vehicle detectors of previous generation that allows you to obtain information about the position of the vehicle in the image. The ability to use the detector of a previous version is due to backward compatibility requirements. It is recommended to use the VehicleDetectorV4 detector in all new working circuits

Detection fields received as a result of the detector work are described in Table 25.

Table 25. Description of the vehicle detection fields

Field name	Type	Description	Possible values
detections	string	An array containing the coordinates and size of the detections of each vehicle in the image, as well as an estimate of the accuracy of vehicle detection	List of detected vehicles, each detection has 5 fields: height, score, width, x, y.
execution_time	int	Execution time in milliseconds	-
height	int	Bbox height.	0…1080
score	float	Estimation of TS detection accuracy.	0.0000…1.0000
width	int	Bbox width.	0…1920
x	int	The horizontal coordinate of the upper left corner of the BBox.	0…1920
y	int	The vertical coordinate of the top left corner of the BBox.	0…1080
detector	string	The name of the detector type.	car

Response example:

{
    "detections": [
        {
            "height": 438,
            "score": 0.7983,
            "width": 915,
            "x": 52,
            "y": 150
        }
    ],
    "detector": "car"
}

LP detector#

The list of LP detectors is given in Table 26.

Table 26. LP detector description

Name	Description
PlateDetectorV4	LP detector of the latest version that allows you to obtain information about the position of the LP in the image. Detector has a number of advantages:
	- Improved performance;
	- Increased speed and accuracy;
	- Additional parameters configuration of LP detector.
PlateDetectorV3, PlateDetectorV2, PlateDetectorV1	LP detectors versions 1, 2 and 3 are the detectors of previous generation that allow you to obtain information about the position of the LP in the image. The ability to use the detector of a previous version is due to backward compatibility requirements. The use of the PlateDetectorV4 is recommended for all new systems

Detection fields received as a result of the detector work are described in Table 27.

Table 27. Description of the LP detection fields

Field name	Type	Description	Possible values
detections	string	An object containing a list of all detected LPs on the image.	List of detected LPs, each detection has 5 fields: height, score, width, x, y.
execution_time	int	Execution time in milliseconds	-
height	int	Bbox height.	0…1080
score	float	Estimation of LPs detection accuracy.	0.0000…1.0000
width	int	Bbox width.	0…1920
x	int	The horizontal coordinate of the upper left corner of the BBox.	0…1920
y	int	The vertical coordinate of the top left corner of the BBox.	0…1080
detector	string	The name of the detector type.	grz

Response example:

{
    "detections": [
        {
            "height": 35,
            "score": 0.9343,
            "width": 139,
            "x": 748,
            "y": 444
        }
    ],
    "detector": "grz"
}

Animal detector#

The animal detector is designed for the detection and redetection of animals on multimedia files.

The description of the animal detector is given in Table 28.

Table 28. Animal detector description

Name	Description
AnimalDetectorV1	An animal detector that allows you to get information about the location and position of animals in a sequence of frames

Detection fields received as a result of the detector work are described in Table 29.

Table 29. Description of the animal detector fields

Field name	Type	Description	Possible values
detections	string	An object containing a list of all detected animals on the image.	List of detected LPs, each detection has 5 fields: height, score, type, width, x, y.
execution_time	int	Execution time in milliseconds	-
height	int	Bbox height.	0…1080
score	float	Estimation of LPs detection accuracy.	0.0000…1.0000
type	string	Тип	big_animal, small_animal, bird
width	int	Bbox width.	0…1920
x	int	The horizontal coordinate of the upper left corner of the BBox.	0…1920
y	int	The vertical coordinate of the top left corner of the BBox.	0…1080
detector	string	The name of the detector type.	animal

Response example:

{
    "detections": [
        {
            "height": 708,
            "score": 0.921549916267395,
            "type": "small_animal",
            "width": 1269,
            "x": 1,
            "y": 214
        }
    ],
    "detector": "animal",
    "execution_time": 167
}