Method for directional cross-border detection and mixing line detection in video

Abstract

The invention discloses a method for directional cross-border detection and mixing line detection in a video. The method includes the following steps of S1, inputting a monitoring video, S2, initializing the video, wherein the mixing line and an interesting region are set in the video, and the minimum filtering movement target area is selected according to the resolution ratio of the input video to be detected, S3, detecting the foreground, wherein movement targets are detected and tracked, and the movement target needing to be detected is filtered out, S4, carrying out mixing line detection or cross-border detection, wherein the cross-border detection comprises the two phenomena of intrusion detection and fleeing detection, S5, carrying out the S3 and the S4 repeatedly for every later new movement target. By means of the method for directional cross-border detection and mixing line detection in the video, the movement targets are filtered, targets which do not need to be detected such as birds and small animals are excluded, detection accuracy is improved, detections of leaving and the movement direction of the movement target can be achieved on the basis of mixing line detection and intrusion detection, detection is fully functional, and the method better meets actual application requirements.

Description

technical field [0001] The invention belongs to the field of intelligent monitoring video combined with computer vision and pattern recognition, and in particular relates to a directional cross-border and line-mixing detection method in the video. Background technique [0002] Intelligent surveillance video is to use computer vision technology to process digital surveillance video to achieve the purpose of automatic real-time detection of abnormalities and alarm. It can replace the working method of manual fixed-point regular inspection or manual long-term monitoring of surveillance video, which can avoid fatigue caused by missing inspections or manual long-term monitoring of surveillance video. The computer automatically detects abnormalities in real time, and is useful in the security of public and private places. Huge application prospects. [0003] At present, the intelligent monitoring video technology is becoming more and more mature, and some real-time monitoring vid...

AI Technical Summary

Problems solved by technology

[0006] Therefore, the existing methods mainly have the following three problems: 1) Simply set the line, and there is no direction setting, but many scenes need to detect retrograde targets, allowing one-way passage, such as urban one-way streets, etc. ;2) The setting of the area of ​​interest is only the detection of intrusion, but in many scenes, it is necessary to detect whether there is a target leaving, and there is no abnormality for the entering target. This area is only in and out, such as schools in class; 3) Movement The target is not filtered. Cats, dogs or flying birds in public places sometimes do not need to be detected, and such abnormalities need to be excluded

Images