Moving target automatic tracking method and system under similar target and background colors

Abstract

The invention relates to the technical field of target tracking, in particular to a moving target automatic tracking method and system under the condition that a target and a background are similar incolor, and overcomes the defect when a CAMshift algorithm is directly adopted for target tracking. The method comprises the following steps: S1, processing a video stream to obtain a denoised sequence frame image; S2, processing the sequence frame image to obtain a foreground target in the first frame image; removing shadows in the foreground target to obtain a moving target area of the first frame image; s3, reading a next frame of image, taking the next frame of image as a current frame of image, obtaining and processing color-curvature probability distribution maps of moving target areas of the current frame of image and a previous frame of image, and obtaining a candidate area of the current frame of image; s4, if the Bhattacharyya distance between the candidate region of the currentframe of image and the moving target region of the previous frame of image is greater than a distance threshold, taking the candidate region as the moving target region of the current frame of image;and repeating the step S3 and the step S4 to realize tracking of the moving target.

Description

technical field [0001] The invention relates to the technical field of target tracking, in particular to a method and system for automatically tracking a moving target when the target and background colors are similar. Background technique [0002] Target tracking is a research hotspot in recent years. The CAMshift algorithm, which uses color as the target recognition feature, is widely used because of its simple algorithm and less calculation. The algorithm cannot realize automatic tracking of moving targets: [0003] First, the CAMshift algorithm is a semi-automatic tracking algorithm improved from the mean-shift algorithm. In the stage of target determination, it is necessary to manually delineate the region of interest and extract the color information of the region as a tracking feature template. In general, the shape of the moving target is irregular, but when the target area is manually selected, it is usually delineated with a regular shape such as a rectangular are...

AI Technical Summary

Problems solved by technology

In general, the shape of the moving target is irregular, but when the target area is manually selected, it is usually delineated with a regular shape such as a rectangular area, so the area of ​​the delineated rectangular area will be mixed with non-target areas in addition to the real target area , that is, part of the background area, which will interfere with the accuracy of target tracking to a certain extent, and also cause a waste of human resources when performing manual operations;

[0004] Second, the CAMshift algorithm only uses color features for tracking. When the background color is too similar to the target color, the tracking window will contain both the target and the background of similar colors. At this time, the area of ​​the tracking window is much larger than the real target area. In some cases Even lose the target and localize to the background area

In addition, when the target is occluded, the CAMshift algorithm is also prone to tracking failure

Images