| As science and technology is increasingly update, compared to traditional video surveillance, more intelligent remote video monitoring system has attracted more and more attention. There is a monitoring system which visual analysis algorithms is regarded as the core of computer and combine with the technology of web, it was suggested that a practical and feasible treatment options, focusing on the following main elements:The core of the web-based video monitoring system is a moving target detection and tracking, describes the moving target detection and tracking algorithms to analyze the advantages and disadvantages of each method and the experimental results were compared. By considering the remote video monitoring system for real-time requirements, are given in terms of improved target detection based on the combination of symmetric difference and background difference method algorithm improves the hole problem symmetric difference method to extract the original goal, to achieve the goals of the movement Real-time effectively extract; given in terms of a real-time target tracking target tracking algorithm based on compressed sensing, the experiments show that the algorithm is consistent with real-time, robust, able to effectively track the target.According to the difference of the color histogram between frames to find mutations and gradient frame in a video sequence, so that the video is divided into many segments, calculated time properties and movement properties of each segment. Fuzzy k-means algorithm for clustering grouped an average amount of exercise and judge exercise exceptions and anomalies.Web-based video monitoring system is built into Web procedures, combined with JFreeChart make visual analysis based on visualization. Users can view the behavior of the monitoring area, the monitoring system is more convenient. Given algorithm for remote visual image analysis is valid, has a high practical value the web-based video monitoring system experimental results show. |
