Full-scene image detection for mixed traffic flow
Sun Liguang, Qixin Shi, Jiji Cao · Research into Practice: 22nd ARRB ConferenceARRB Group Limited · 2006
The purpose of this paper is to present the research on background model and shadow detection of video and image detection. Video/Image Detection is modern and effective for the detection of traffic flow because it can be easily installed and integrates both surveillance and detection. The Mixed Traffic flow exists in so many developing countries that emphasize the importance of relevant information collection methods. The Extended Running Average algorithm for obtaining and updating representation of the background scene is proposed and compared with a background model based on Kernel Density Estimation and Three-parameter Background Model in efficiency. A shadow model is proposed, considering the significant impact of the moving object’s shadow to detection efficiency. After analyzing samples which were captured from real traffic scenes, in several color spaces such as RGB and HSV, the utility of proposed method is demonstrated through experiments on several scenes. Through the analysis of the characteristics of mixed traffic flow and available methods on Video/Image Detection, a full-scene image detection framework based on Background subtraction is proposed.