Research On Intelligent Monitoring Technology Based On Computer Image Processing

Intelligent monitoring, based on computer vision technology, has been developed rapidly as the growth of the demand for security product. Some relevant technologies have become the research focus amongst domestic and foreign scholars. The paper lucubrates some of the technical difficulties with the existing research results.The original images, obtained by camera, need to be preprocessed before detecting moving targets. The paper adopts common method, called median filter algorithm, for image de-noising, and combines with the histogram equalization algorithm to enhance the interesting areas, which is convenient for further processing. It is demonstrated that this method is effective through experiments.Moving target detection is one of the priority research areas. The paper introduces several traditional detection methods. Visual background extraction algorithm(ViBe) may lead to ghost shadow easily and can't detect foreground object clearly and accurately based on background subtraction under the dynamic background. This paper puts forward an improved ViBe algorithm. The improved method adopts spatio-temporal background modeling based on the ViBe. It initializes single frame image and consecutive images separately, then combines them through the relationship of time and space dimension that can eliminate the influence of ghost effectively. The threshold's ability to adapt to the dynamic background is enhanced by way of updating the background models separately and introducing the adaptive threshold value method in the process of matching.In the aspect of target tracking,Kalman filter and Camshift tracking algorithm are researched carefully. Due to the poor effect for target shade using Camshift tracking algorithm. The paper judges blocking moment in use variation of moving target area and introduces least square method to predict the movement trajectory during blocking moment, which reduces the chance of tracking lost when blocked. It can improve the effect of the multiple targets tracking when improved Camshift tracking algorithm combines with Kalman filter.Finally, the paper designs system function and summarizes the full text. It discusses the development of intelligent monitoring technology.