With the development of computer technology, Computer Vision research has made rapid progress. At present, technologies of Motion Detection, Recognition and Tracking in Computer Vision have been using in National defense, Aviation, Navigation, Medical, Security, etc. Motion Detection in the environment with severe noise interference, purpose of which is to detect and locate the object that occurs in monitoring area, is an important research direction in Computer Vision.This article proposes a new motion detection algorithm based on2-D spatiotemporal states histogram. The new algorithm combined the idea of image change detection based on2-D histogram and the one of spatio-temporal entropy image segmentation. Innovatively, new algorithm has completed the conversion of video sequence to histogram image sequence. We name the processes of assess as TDF (Time Domain Filter) and SDF (Space Domain Filter).Missions of TDF and SDF are to increase values of the pixels with well state Continuity in time and space respectively. By them, most of foreground pixels value can be distributed in higher levels, intensively. Innovatively, this algorithm quantifies the original data in a time way and a space way. After this, put both channels of output data from TDF and SDF into2-D histogram. In the2-D histogram, a new curve division method helps to separate the foreground state points and the background ones, more accurately. Experimental results show that the motion detection algorithm based on2-D spatiotemporal states histogram can almost Meet the requirements of real-time, detect out moving object in the environment with severe noise. |