Algorithm Study And Realization Of Vehicle Detection Based On Video Image Technology

With the development of computer hardware and computer vision technology, Intelligent Traffic System has made great progress. Society's advancement and economic development also request ITS become consummate step by step. As an important part of ITS, the automobile incident detection system get more attention from many academicians.Vehicle detection is very important in a good automatic transport incident detection system. There are many traditional detection methods, but these methods are not very flexible. The detection method based on video image has its unique advantages, such as huge detection regions, small engineering time, easy equipment, low expense and rich information etc, which will take over the traditional methods. As the development of computer hardware, vehicle detection and tracking based on video image is becoming very popular step by step.The thesis's research object is a freeway video image which got from a single CCD camera .It's aim is to detect the moving vehicle in the video image. This thesis research and analyze the most important problem related to vehicle detection, such as background extraction and update, shadow detection and vehicle segmentation etc. Based on these methods, this paper also realizes a video vehicle detection system. The main contents can be listed as follows:1. On the basis of traditional background subtract method, a new car detection method which uses wavelet transform and mathematical morphological is proposed. At first adopted the statistical method -quick serial equalization to extract the background model. Because the effect of rapid beam change to background is very obvious, the thesis introduces a corresponding simple and rapid background update method. Do background subtraction using the built background image, then integrate morphological method and wavelet transform to detect vehicle accurately.2. This paper proposes a moving car segmentation method based on improved OTSU method.3. Design and realize a video car detection system. This system is developed using Visual C++ 6.0 and the image development package OpenCV 5.0. This system mainly realizes background extraction, image filtering, image threshold, Morphologic -al transform, vehicle detecting and segmenting, shadow detection etc.