Parking state change identification method

Abstract

The invention discloses a parking state change identification method, and belongs to the technical field of parking state identification. The method comprises the following steps: firstly, obtaining aparking space frame in an initial image by using a yolov3 deep learning model, and judging an initial state of a parking space by using the parking space frame; judging whether a moving object existsin each frame of image in the subsequent images or not by utilizing frame difference method motion detection; judging whether the moving frame and the parking space frame are overlapped or not, if so, judging whether the moving object is a vehicle or not by utilizing a yolov3 deep learning model, and if so, obtaining a real contour of the moving object by utilizing the yolov3 deep learning model,and judging the state of the parking space by utilizing the overlap ratio of the real contour and the parking space frame. According to the invention, dependence on a deep learning model is reduced,so that the whole algorithm is lower in power consumption and higher in efficiency. If the moving object is not a vehicle, the frame of image is ignored; if the moving frame and the parking space frame are not overlapped, the state of the parking space is kept unchanged. The dependence on a deep learning model is reduced, so that the whole algorithm is lower in power consumption and higher in efficiency.

Description

technical field [0001] The invention relates to the technical field of parking state recognition, in particular to a parking state change recognition method. Background technique [0002] With the growth of vehicles, there are often many parking spaces on the roadside of urban streets. At present, parking fees are mainly collected manually. A toll collector is needed for a section of the street. Toll collection alone requires a lot of human resources. At present, the automatic recognition of parking on urban streets is not very mature. There are two main types: the first type is to use the yolo deep learning model to track the vehicle appearing in the video. After the coordinates of the parking spaces coincide to a certain extent, it is judged that the vehicle is entering, and the vehicle is judged to be leaving in the same way. Using this method, each frame of the video image needs to be input into the yolo deep learning model for recognition, and the power consumption is p...

AI Technical Summary

Problems solved by technology

[0002] With the growth of vehicles, there are often many parking spaces on the roadside of urban streets. At present, parking fees are mainly paid manually. A toll collector is needed for a section of the street. Toll collection alone requires a lot of human resources.

At present, the automatic recognition of parking on urban streets is not very mature. There are two main types: the first type is to use the yolo deep learning model to track the vehicle appearing in the video. After the coordinates of the parking spaces coincide to a certain extent, it is judged that the vehicle is entering, and the vehicle is judged to be leaving in the same way. Using this method, each frame of the video image needs to be input into the yolo deep learning model for recognition, and the power consumption is particularly high; the second type It relies on some traditional image processing algorithms, which lack robustness and cannot solve many complex situations in real scenes, such as pedestrians walking on the street, light changes, sunlight changes, day and night, etc.

Images