3D Otsu Moving Vehicle Detection Method Based on Steepest Ascent
Shangwan Chen, Zhengjie Wan · 2013
In the outdoor traffic situation the large number of noise can be eliminated by using traditional three-dimensional (3D) Otsu method, but the work is time-consuming and can not satisfy the real-time demand. To solve this problem, a method for 3D Otsu moving vehicle detection based on steepest ascent is proposed which combines with Mixture Gaussian Backround Modeling (MGBM). Experimental results show that it can detect moving cars accurately for real-time video processing in the case of complex backgrounds and jittering in the video.