Python Module for data transferring

CARS Stream has a functionality for transferring information about the detection of vehicle and LP to external systems (CARS API, CARS Analytics) via the python module «callback_manager.py».

The data is processed by the «process_callback» method. This method receives a set of discovery results in key-value format.

Translation of the response in JSON format:

{
                    'trackId' (int): 12,
                    'vehicleData' (dict): {
                        'licensePlate' (str): 'H919CH40',
                        'licensePlateScore' (double): 0.999257,
                        'vehicleType' (str): 'B',
                        'vehicleTypeScore' (double): 0.999257,
                        'vehicleEmergencyType' (str): 'not_emergency',
                        'vehicleEmergencyTypeScore' (double): 0.98,
                        'carBrand' (str): 'Kia',
                        'carBrandModel' (str): 'Cerato',
                        'carBrandModelScore' (double): 0.999257,
                        'country' (str): 'RUS',
                        'countryScore' (double): 0.997342,
                        'carImg' (numpy.ndarray, optional): [...],
                        'licensePlateImg' (numpy.ndarray, optional): [...],
                    },
                    'detections' (list(dict)): [{
                        'frameId' (int): 12,
                        'x' (int): 1282,
                        'y' (int): 1390,
                        'width' (int): 200,
                        'height' (int): 120,
                    }, ...],
                    'operation' (str): 'recogn_update',
}

«trackID» and «vehicleData» blocks include general information about the vehicle and LP. The description of the «trackId» and «vehicleData» fields is given in Table 12.

Table 12. «trackId» and «vehicleData» fields

Field	Description	Type
trackId	Vehicle track id	Int
licensePlate	Car license plate’s data	String
licensePlateScore	Score confidence of responsed license plate	Double
vehicleType	Type of the vehicle	String
vehicleTypeScore	Score confidence of determining the vehicle type	Double
vehicleEmergencyType	Vehicle emergency type	String
vehicleEmergencyTypeScore	Score confidence of responsed vehicle type	Double
carBrand	Recognized car brand	String
carBrandModel	Recognized car model	String
carBrandModelScore	Score confidence of responsed car model	Double
country	Car license plate country	String
countryScore	Score confidence of responsed license plate country	Double
carImg	Car img represented as numpy array	numpy.ndarray
licensePlateImg	Car license plate img represented as numpy array	numpy.ndarray

«detections» block includes information about the position of the vehicle and LP on the frame. The description of the «detections» fields is given in Table 13.

Table 13. «Detections» fields

Field	Description	Type
frameId	Bbox vehicle ID	Int
x	X coordinate bbox	Int
y	Y coordinate bbox	Int
width	Bbox width	Int
height	Bbox height	Int

|«operations» field includes the type of operation performed. The description of the «operations» possible values is given in Table 14.

Table 14. Description of the «operations» field

Field	Description	Type
recogn_update	The best shot of the current track	String
track_end	Track end event	String