Method and system for abnormal behavior detection based on video analysis

Abstract

The invention provides a video analysis-based abnormal behavior detection method and system. Wherein the method comprises the following steps: extract pedestrian foreground image from video frame; carry out grid division to video frame, divide into a plurality of grid areas, and set the grid area where pedestrian foreground image is located as motion area; adopt nearest neighbor method Mark the motion area and associate the motion area of ​​adjacent video frames; calculate the optical flow features in the marked motion area; obtain the weighted direction histogram according to the optical flow feature; calculate the entropy of the weighted direction histogram; The Gaussian model selects the detection threshold, detects whether abnormal behavior occurs according to the detection threshold and the entropy of the weighted direction histogram, and updates the detection threshold. It can automatically detect abnormal behaviors in video scenes in real time, avoid abnormal behaviors from disrupting the order of public places, endanger public safety and personal safety, reduce the workload of personnel, and avoid potential safety hazards caused by missed and false detections.

Description

technical field

[0001] The invention relates to the field of behavior recognition, in particular to a video analysis-based abnormal behavior detection method and system. Background technique

[0002] At present, most video surveillance systems are still in the traditional mode, that is, "only record but not judge", so they can only investigate abnormal situations and obtain evidence through post-event video playback. Intermittent monitoring of activities in the scene, day and night on duty, heavy workload, susceptible to human sensory fatigue, resulting in missed and false detections, loses the significance of the monitoring system for on-site real-time monitoring. On the other hand, with the expansion of the scale of the monitoring system and the massive increase in the number of videos, it is becoming more and more difficult to obtain useful information or intelligence, and the search efficiency is low, which is difficult to meet the needs of the monitoring system. Conte...

Images