Research Of Behavior Detection Method In Video Surveillance For Primary And Secondary School

The video surveillance system has been widely used in primary and secondary schoolsin which by use of takes the video processing and other techniques, such as advancedcomputer network, photoelectric sensor, artificial intelligence and automatic control,therefore, it has Therefore, how to make full use of these massive video information inadvance in order to found the unexpected behavior and the case in campus has become animportant research content. The several key technologies in video pedestrian detectionapplication for small and medium-sized school has discussed and researched in this paper,the specific contents are as follows:(1) Aimed to the problem in which HOG feature cannot distinguish pedestrians andbackground, on the study of the LBP feature extraction method, a new pedestrian detectionalgorithm by using LBP to measure and extract the features of image texture informationof local adjacent area is put forward in this paper based on SVM and HOG-LBPcharacteristics, which can make up the shortage of HOG features, improves the accuracyrate of classification.(2) In view of the existing single feature pedestrian detection algorithm in thecharacterization of pedestrian ability insufficiency, feature extraction, recognition and lownoise is more based on based on HOG feature and LBP feature, on HOG and LBP werecharacteristics of the combination of analysis, the pedestrian HOG-LBP combined featuredetection algorithm is put forward based on SVM. The completion of the pedestriandetection test in the INRIA sample database and the campus monitoring samples, theexperimental results show that based on the mixed characteristics of pedestrian detectionalgorithm is relatively single feature detection has higher detection accuracy and speed.(3)An intelligent surveillance video is analysis based on the OpenCV softwareplatform and the system designed and realized. The system implementation process is discussed in this paper, such as the code key module, at last, its feasibility and validity isproved in practical application.