Research On Algorithm Of Intelligent Monitoring System

In recent years, the computer plays increasingly important role in everyday life in the context of the information era of big data. The rapid development of computer hardware, high-capacity storage device after another and a substantial increase in processor power provides the basis for the widely used image processing. In image processing, intelligence has been a very clear direction. It is a very challenging research direction that make a computer instead of people to recognize a scene. A complete video surveillance system includes extracting and tracking moving targets, more advanced system should have the features of intelligent image understanding.Looking a image is the most intuitive way of understanding of things, current computer applications primarily in the text or speech, the future trend will be toward intuitive intelligent direction. This mission gives the computer era in the field of image processing vitality in recent years, a variety of machine vision algorithms layer see stacks up, machine vision have become a very popular research field.In the video surveillance system, target detection and tracking is a fundamental issue,many applications can be derived from its research, such as launching missiles, satellites and launch tracking, delineation of enemy military targets, modern intelligent traffic monitoring systems and intelligent monitoring system of social security.Therefore, in the current situation the research of the application of intelligent video surveillance system has very important value, this paper do a thorough research on the basic aspects of video surveillance systems, the main research results obtained are as follows:1.We do a detailed comparison of motion detection algorithms, use different algorithms for comparing test results do the same scene, choose a low sensitivity to light and accurate motion detection algorithm as the basis for moving object detection. The original algorithm has been improved for problems that the exist in light impact and computational complexity, making the detection algorithm has greatly improved in accuracy, but also take into account the real-time algorithm.2.We add a background update process on the base of improved algorithm, while in the background model testing for motion detection in complex background, In the context of the model while testing, follow-up study of the actual situation in the video, so that makes the detection algorithm is more widespread application scenarios, test results more accurate and consistent with human cognition.3. The advantages and disadvantages of common moving object tracking method is intuitive showed through practical application and introduced to the theory. Select the particle filter tracking algorithm as the basis for its targeted improvements. In view of the fact that the particle filter is not good at tracking the target is the background scenery surrounding high similarity.I made certain improvements of the particle filter that the moving target can be tracked continuously, with the lost situation reduced.4. Add the cordon detection algorithm. The moving object detection, object tracking, cordon barricaded system detects a three-part practical use, and the code has been optimized, the entire system has good usability.