The Research On Algorithms Of The Moving Vehicle Detection And Tracking Based On Video Image

Traffic information detection technique based on video image is widely concerned in recent years, because of its wide detection range, rich collection information and easy installation and maintenance. Moving vehicle detection and tracking are the key techniques in traffic information collection and detection system.In the thesis, traffic scene images are research objects, and the key techniques of background extraction and updating, shadow removal, vehicle detection and vehicle tracking are studied and analyzed.Based on the studying of conventional background extraction algorithms, an algorithm of combination of histogram and multi frame average for background extraction.The algorithm firstly find several maxima from the histogram, and using the average method to determine the believe degree of the several maxima, and then select the highest believe degree as background. an algorithm for background updating and an algorithm of alternate threshold selection for background difference image bi-valuation are given. Shadow of vehicle is mistaken for target for the influence of illumination and lighting equipment, an algorithm combined grabs and line-scan is introduced to remove shadow. After using connected component labeling to process binary image, moving target is detected and tracked by vehicle tracking algorithm based on Kalman filter. And it provides a foundation for traffic incident detection and analysis.The algorithms proposed above are experimented in the situation of heavy vehicle flow, mutual occlusion and adhesion of vehicle. The experimental results show that it can deal with vehicle occlusion problem better and realize the vehicle detection and tracking in real time.