The Research Of Financial Outlets Intelligent Video Surveillance Technology

People’s way of life has been changed with the development of science and technology, saving banks and the ATM brings us convenient also brings safe hidden trouble at the same time. The working mode of security personnel before the screen has been changed by Intelligent Monitoring. It will be warning to remind the security personnel handle it when the possible security problems appears.Intelligent video applications to the technology of image processing and pattern recognition which has become an important research field in computer vision, and have made the tremendous contribution in control and prevention field.Trailing behavior, Robbery behavior and Bank card theft are the three abnormal behavior financial networks using intelligent monitoring system in this paper. To detect the moving targets is the first step and basic way to identifies the abnormal behavior, so the three commonly moving object detection algorithm are introduced in this paper firstly. To compare the advantages and the disadvantages combined with the map of the experimental effect of three methods, and explains its appropriate application places. In the subsequent work according to the different needs to take a different approach.To began recognize the abnormal behavior after detected the moving object. By combining with the characteristics of the trailing behavior, the identification method combining the characteristics of relative mass center with motion trajectories is proposed. The moving objects in the background are extracted by using the frame difference method, and then whether there is trailing behavior is determined. The method of histogram weighted by the amplitude be using to identify the Rob behavior. The direction feature histogram weighted by the values of the optical flow vector magnitude normalize. Direction histogram which is weighted would obvious the motion region information clearly, and introducing the concept of regional entropy to judge. The Bank Card Theft behavior recognition has two aspects, the judgment method of The YCbCr color space based on skin color area of face region is used in the masked behavior, if the skin color area is less than a certain threshold, then think the masked behavior set up. The face matching method based on the LBP operator is used in the problem of matching the former face and the latter face in withdrawal operation, coding facial image based on the LBP operator. The encoded the LBP map as the recognition basis of matching, with the comparison of the LBP histogram similarity between every two samples to determine the face of two video frames to judge whether they are similar or the same person, if the matching fails is suspected of bank card theft occurred. Once the behavior is judged on the establishment just alarm, prompt staff make further processing. After verification by experiment, the application and all the methods are practical, algorithm is simple, suitable for real time monitoring system in this paper.