Fight Detection Method in Intelligent Video Surveillance

Abstract

The invention discloses a fighting detection method in intelligent video monitoring. The method employs a dynamic energy characteristic to quantify the possibility of fighting in a video image. The dynamic energy characteristic comprises kinetic energy, potential energy and a confusion degree, wherein the kinetic energy needs motion speeds and object quality, the potential energy needs the object quality and distances between objects, the confusion degree needs a total motion direction of the objects and the motion direction of each point in the objects, and the object quality, the motion speeds and the directions are not easily obtained. According to the scheme, the object quality is simplified to proportions which the objects occupy in the video image, for the motion speeds and the directions of the objects, the motion direction and the speed of each point in the objects are obtained by use of an optical flow method on the basis of preceding extraction of a foreground through a background in a difference mode, and finally, through combination with a corresponding control process, fighting detection in the video image is effectively realized, the algorithm complexity is low, the operation speed is fast, the misjudgement rate is rate, and the stability is good.

Description

technical field [0001] The invention relates to intelligent video monitoring technology, more specifically, a fighting detection method in intelligent video monitoring, which belongs to video image processing technology. Background technique [0002] With the development of science and technology, intelligent video surveillance has been widely used in people's production and life. As an intelligent community security system, intelligent video surveillance has gradually developed from simple video evidence collection to intelligent recognition of video images. In video image intelligent recognition, it usually includes face recognition and abnormal human behavior detection. [0003] Existing abnormal human behavior detection algorithms are usually aimed at the detection of abnormal movements in the case of a single person. This method is suitable for the situation where there is only one person in the monitoring area, but in practice, there may be violent conflicts among mul...

AI Technical Summary

Problems solved by technology

[0009] Aiming at the deficiencies of the prior art, the present invention provides a fighting detection method in intelligent video surveillance, which integrates the degree of chaos of the moving mass into the dynamic energy feature, and solves the problem that the existing algorithm only uses kinetic energy and potential energy as the dynamic energy feature The defect of high misjudgment rate caused by factors, at the same time, this method uses the buffer to store the dynamic energy feature values ​​of several consecutive frames of video images to compare with the threshold to improve the accuracy of the algorithm

Images