Research On The Algorithm Of Moving Object Detecting

Intelligent video surveillance system plays a pivotal role in real life as a frontier topic inComputer Vision domain. It has broad application prospects in the fields of defense, power,transportation, industry, etc. Moving target detection technology is the key technology tointelligent video surveillance system, the researching on it has both important theory significanceand application value. However, in the real environment, there is not a common moving targetdetection scheme due to a variety of uncertain factors. This dissertation focuses on the methodsof moving target detection in static scene, the main contents of it are summarized as follows:(1) The research status and background of the moving target detection are introduced briefly,and describes some of the commonly used image processing technology. To reduce the affect ofnoise and edge blur in the detection process, using the differential operator and mask operations.(2) To eliminate the impact of background disturbance, an effective and practical programwas presented to get a stable background. The idea of the program is to control the pixels of thetarget area and updates the pixels of the background area, combined with the inter-framedifference and iteration; it can overcome the drawbacks of the statistical average method,through dynamic update rate and threshold settings.(3) In order to eliminate the noise and holes of the binary objects image, Morphologicalalgorithms and Median filter are applied to it.(4) In the actual target detection, the background with frame difference thought effectivelyintegrates to solve the key technologies, such as target segmentation, background update，ghosting and real-time. Experimental results show that the moving target detection algorithm canget stable pure background, accurate segmentation of the foreground object.