Checkpoint snapshot image vehicle detection method, computer storage medium and electronic equipment

Abstract

The invention provides a checkpoint snapshot image vehicle detection method, a computer storage medium and electronic equipment. The method comprises the following steps of: S1, acquiring a checkpointsnapshot image, S2, identifying the checkpoint snapshot image, and performing vehicle detection on the identified vehicle by adopting a detection network, 3, filtering vehicle information with an incomplete vehicle body, and S4, outputting a complete vehicle body area. According to the method provided by the embodiment of the invention, the checkpoint snapshot image data are firstly acquired, a TinyDSOD network is then used for vehicle detection, incomplete vehicle bodies are also detected, but the vehicle body areas lack effective information and cannot be subjected to subsequent analysis, so that the incomplete vehicle bodies need to be filtered, finally the complete vehicle body areas are output, and missing detection and false detection of the algorithm on the vehicles can be reduced.

Description

technical field [0001] The present invention relates to the field of vehicle detection, and more specifically, to a vehicle detection method, a computer storage medium, and an electronic device for a bayonet snapshot image. Background technique [0002] Bayonet vehicle detection is the pre-stage of vehicle video surveillance, which detects the vehicle area in the captured image and inputs it to the subsequent vehicle attribute analysis and recognition link. A bayonet vehicle detection algorithm with high recall rate and precision rate is very important to improve the recognition accuracy and efficiency of road vehicle video surveillance system. In recent years, with the improvement of computer computing power and the promotion and application of deep learning theories and methods, a large number of target detection models based on deep neural networks have emerged, such as YOLO, SSD, MobileNetSSD, TinyDSOD, etc. Among them, the accuracy rate of the YOLO algorithm ranks firs...

AI Technical Summary

Problems solved by technology

For example, TinyDSOD, the model is small enough, but the feature extraction ability is weak. When detecting the captured image of the bayonet, it is prone to missed detection and false detection.

Images