Research On Abnormal Behavior Detection Algorithm Of Students In Surveillance Video

Video analysis has become a research hotspot due to its many applications in computer vision.It includes three main operations: moving target detection,object tracking,and recognition.The goal of the research in the project mainly refers to the mobile phone,that is,students' behavior detection using mobile phones during the course of class.The purpose of the detection operation is to find a foreground moving object at the first appearance of the target or in each video frame.In view of the complex background of the specific scenes of the laboratory in the subject,the inconsistent size of the mobile phone,and the uneven illumination,the artic le mainly starts from the following aspects to detect and use the abnormal behavior of the mobile phone:(1)In the video,the information used for detection and analysis is generally a pixel spatial neighborhood and features.In order to better focus the detection of mobile phone targets in a video sequence,image feature extraction is based on image segmentation,edge detection,and morphology processing and the like.Theoretically,this topic proposes the research idea of using image entropy-based image segmentation and sobel edge information extraction features to meet the robustness and accuracy of mobile phone target detection and tracking recognition.(2)Since the mobile phone has the characteristics of dynamic position change,this paper summarizes the common target detection algorithms such as frame difference method,backgrou nd difference method,optical flow method and statistical model.This paper proposes a method based on the cumulative difference method.(3)Based on the above research foundation,the cumulative difference target detection is used to implement tracking modeling of complex scenes.Developed a video motion object feature recognition method that matches these spots with the template in the order of position,shape,and color,and uses the nearest neighbor classifier to identify abnormal behavior as a student using a mobile phone,providing experimental results to prove Tracking accuracy in a complex environment.The experimental simulation analysis in this paper is based on MATLAB to achieve the detection of moving targets.After a series of experimental tes ts and development,the accuracy,real-time and robustness of the moving target detection are proved.