Method for detecting moving object based on frame difference method and moving point clustering

Abstract

The invention relates to a method for detecting a moving object in the field of video processing, which comprises the steps of S4, obtaining moving point sets by frame difference, wherein the moving points are discontinuous, and the gray and white parts are broken and not connected; S5, executing intra-frame moving point clustering, and connecting the gray and white moving point sets to form a complete contour of a person; and S7, enhancing the moving contour of the person through inter-frame moving point clustering. According to the method, the moving points are obtained by adopting the framedifference method, and anti-interference analysis and combination processing are performed on the moving points based on intra-frame and inter-frame clustering algorithms. The method is better than the Gaussian background modeling algorithm and the ViBe background modeling algorithm in adaptability, and can adapt to the problems such as a blocking effect, a flash screen and lamplight sudden changes of the video. In addition, the method is superior to the frame difference method through the combination processing for the moving points, is high in sensitivity, can detect slight movements of theobject and obtains a continuous moving region of the object.

Description

technical field [0001] The invention relates to a moving target detection method in the field of video processing, through which a continuous video frame sequence is analyzed to detect moving objects in the video. Background technique [0002] Currently, the frame difference method, Gaussian background modeling and VIBE background modeling methods are the most commonly used methods in the field of moving object detection. However, the object detected by the frame difference method will appear hollow, and only the outline of the moving object can be detected, and there will be breaks and discontinuities. Gaussian background modeling and VIBE background modeling methods can detect the entity of moving objects and there will be no empty scenes, but these two methods cannot adapt to sudden changes in the environment, and have poor adaptability to situations such as block effects, jumps, and low contrast in the video . [0003] Various problems encountered in the detection of m...

AI Technical Summary

Problems solved by technology

However, the object detected by the frame difference method will appear hollow, and only the outline of the moving object can be detected, and there will be breaks and discontinuities

Gaussian background modeling and VIBE background modeling methods can detect the entity of moving objects and there will be no empty scenes, but these two methods cannot adapt to sudden changes in the environment, and have poor adaptability to situations such as block effects, jumps, and low contrast in the video

[0003] Various problems encountered in the detection of moving objects in the Internet security of existing shopping malls. The specific problems include: 1) The video acquisition end, the equipment is cheap, the image effect is poor, and after the video is transmitted through the network, there will be low contrast, many noises, and flashes. screen, video block effect and other problems; 2) the application scene is complex, the installation position of the camera is uncertain, and the close-focus application, the color diversity, texture, and edge details in the background interfere with the foreground analysis; 3) in addition, the pedestrian movement pattern facing the store is complex, On the one hand, the movement is slow and there is a phenomenon of stagnation (browsing products, queuing to pay fees, etc.)

Images