Video-Based Motion Human Abnormal Behavior Recognition Algorithms

Abstract

The existing video-based algorithms for the recognition of abnormal behaviors of moving human beings have problems of low timeliness and recognition rate, so the video-based algorithm for the recognition of abnormal behaviors of moving human beings is proposed. The median filtering method is adopted for the equipment and the environment in the video produced by the influence of noise to filter out, on this basis, the background difference method is used to detect human movement accordingly, based on test results, through regional correlation method to track human movement, will get the image binarization processing, to extract of human movement behavior characteristics, based on human movement behavior characteristics, through the fuzzy algorithm to criterion of human movement behavior, achieved the identification of the abnormal behavior of human movement. Through experiments, it is found that the time efficiency of the proposed abnormal behavior recognition algorithm is reduced by 18.23% and the recognition rate is increased by 31.7%, which fully indicates that the proposed abnormal behavior recognition algorithm has better recognition effect.