Yang Jianguo, Yin Xuquan, Fang Li, Li Jian, Wang Zhao'an
(School of Electrical Engineering, Xi'an Jiaotong University, Xi'an 710049, China)
Abstract: To detect and track moving vehicles and achieve the movement trace by image
processing in traffic video sequences, a new algorithm based on the fixed template is
proposed, where self-adaptive contour matching method is combined with false detection
judgment and contour division. The algorithm accurately detects the vehicle contours and
avoids the false detection due to two vehicles approached with each other. The
self-adaptive contour matching method is once introduced into the moving vehicle detection
to ensure a precise tracking results. Compared the computer detecting results with the
manual ones, it is concluded that under the certain conditions this algorithm effectively
decreases the false detections, and the accuracy reaches to 95.01%, and fewer missed
detections can be made up by the interpolation.
Keywords: video; template; vehicle detection