Research Of Some Key Algorithms In Video Moving Target Detection

Intelligent video surveillance system is one of new arising advanced topics in Computer Vision fields. Compared to traditional video surveillance system, it has advantages of higher surveillance quality, less labor-intensive, better stability, less security risks, and has significantly improved in real-time and accuracy. It has cheerful prospect, the researching of this system also has important scientific significance.The major contents this dissertation researches is research of moving target detection with single static camera. On the basis of analysis and study of the relevant existing algorithms, select two of them to improve, and combining related theory propose a new method detect shadow, the major contents this dissertation can be summarized as the following three aspects:1,In order to reducing the affect came from noises, after researched and analysed existing filtering algorithms, this dissertation improve a filtering algorithms which can reduce the pixel value difference between original image and the image have been filtered.2,By comparing the existing Moving Target Detection algorithms, this dissertation choose Background Subtraction Method to detect moving target. Aims to the disadvantage of Mixture Gaussian Model, proposed a method detect moving target combine with the improved Average Background, and update Mixture Gaussian Model parameters according to the background condition changes which gets from the improved Average Background. At end, verified the superiority of this dissertation's algorithm by experiments.3,After analysed the existing shadow detection algorithms, this dissertation combine with knowledge of optics and illumination model proposed a shadow detection algorithm based on RGB color space, and verify the effectiveness and feasibility of the algorithm by experiment.