Study On Video-based Detection, Recognition And Tracking Method Of Vehicle In Its

This thesis mainly discusses the fundamental theories and key technologies of Intelligent Transportation System based on computer vision. The application of computer vision technology to traffic detecting and information collecting is an important subject in ITS, and moving vehicle detecting, recognition and tracking is the basic part. To the above problems, the further study is done in this paper and new methods are brought forward. Moreover, experiments are implemented to demonstrate the validity. The main contents of the study include such aspects as follows:1. A new self-adaptive background approximating and updating algorithm based on optical flow theory is first presented in this paper. It can adapt to the variance of illumination. Moreover, the difference image is obtained by using a color image difference model, and then a self-adaptive thresholding segmentation method for moving objects detection based on Gauss model is developed and implemented. The proposed new background updating method can update the background exactly and quickly, and the detection result of moving objects is good without background image.2. An effective shadow-eliminating algorithm based on contour information and color features is first developed. Firstly, the shadow direction is obtained based on the vehicle contour. Then, the contour for the moving region is picked up and the inflex points are calculated to get the shadow seed points. Finally, the shadow region is eliminated by using color clustering method and morphological processing. Experimental results demonstrate that the proposed method can eliminate the shadow accurately and real-time.3. To the situation that there are always trunk, car, motorcycle and passerby on the road in our country, a hierarchy of image recognition based on multi-feature for moving objects is proposed in this paper. Firstly, support vector machines are used to realize the classification based on the shape moment features. Secondly, fuzzy integral is used to implement the multi-feature fusion classification based on the importance...