Y.B. Han, P. Lei, A. Donofrio

Pages: 137-146

Abstract
In order to overcome low recognition accuracy and long tracking time in traditional road multi-vehicle recognition and tracking algorithms, a road multi-vehicle recognition and tracking algorithm based on computer vision is proposed. Firstly, the road multi-vehicle vision image is collected through the internal and external parameter model of the computer vision sensor; Then the visual image data is fused by multi-thread control, and the fused image is preprocessed by graying, filtering, denoising and enhancement; Finally, the method of row gray value statistics is used to dynamically delimit the area of interest, and identify multiple vehicle targets on the road in this area, and form a tracking door for the identified targets, and associate multiple targets and tracking objects falling into the same tracking door to obtain the multi-target tracking results. The simulation results show that the proposed method has high accuracy and short tracking time.
Keywords: computer vision; road multi-vehicle; multi-thread control; areas of interest; IEPF algorithm