Design And Implement Of A Video Monitoring System With People Counting Function

With the rapid development of video technology, the early infrared, gravity sensing technology of people counting gradually are eliminated. Those technologies are low counting accuracy. Statistical data is often large deviation, which losses of data analysis value. With the development of home monitoring, computer vision technology has been gradually into people’s vision. The people counting statistics use the intelligent image analysis to count the people in some places.Two angles are analyzed and introduced in detail in the paper. One is the development of the video surveillance industry development, and the other is the people counting technology. The paper contents:First, the intelligent monitoring system for market is the core research, including some research of the background and summarize. Intelligent market monitoring system is introduced. In this paper, we consider the intelligent monitoring system in a larger context for thinking of the intelligent monitoring system of shopping malls. According to the current research of video monitoring points, its development tendency is given a simple analysis.Second, three levels of the video image feature extraction algorithms are classified, which respectively are the pixel image based on feature extraction, edge-based image feature extraction and region-based image feature extraction. Each type of image feature extraction method is analyzed in detail, which laid the foundation for the knowledge of image processing system, paving the way for the following topics.Third, four classes of image feature extraction algorithms are introduced in detail, such as the color feature extraction, texture feature extraction, shape feature extraction and spatial feature. Each kind of algorithm is listed the basic principle, and the basic characteristics of some key algorithms, as well as the application of different algorithms. We do some research, collected different needs from different users. We carried out a detailed record and analysis. The final system should have the characteristics, such as user login, video playback, and the real-time people statistical characteristics. Fourth, according to the different system function, we carried out a detailed design and code implemented. Through the experiment, some flaws and shortcomings are finding, we proposed an improvement method, such as the addition of trajectory analysis strategy. The algorithm has the higher statistical accuracy, and the algorithm has better real-time performance than before.