Monitoring and warning of lawn trampling behavior in surveillance video

In this paper, we present a monitoring and warning method for the lawn trampling behavior in the surveillance video. Through analyzing the surveillance video image in detail, a background model is established at first. Then the lawn area in the video is determined and the background difference method is used to detect and extract the motion target. Then, according to the positional relationship and the overlap area of the lawn area and the moving target, whether the pedestrian has trampled the lawn is judged. When lawn trampling behavior has happened, the monitoring information will be displayed on the screen of the monitoring video in time and a warning sound will be issued. Experiments show that the method can extract the moving target framework clearly and can detect the human lawn trampling behavior and alert them in time. The method also has good real-time performance.