KCF tracking target loss detection method and system based on foreground detection

Abstract

The invention relates to the technical field of visual target tracking, and the embodiment discloses a method and system for detecting the loss of a KCF tracking target based on foreground detection. The KCF tracking target loss detection method based on foreground detection proposed in this application uses the foreground detection algorithm to extract the foreground image, uses the KCF tracking algorithm to extract the target detection frame, obtains the foreground area corresponding to the target detection frame, and calculates the target contour area in the foreground area. The proportion value of the total area of ​​the area, and the method of setting the proportion value threshold to determine whether the KCF tracking target is lost, solves the problem that the KCF algorithm itself cannot detect the loss of the target, and the KCF tracker will continue to track the background information as the target. Without affecting the real-time performance of KCF target tracking, accurate judgment is achieved when the tracking target is blocked by a large area or leaves the lens.

Description

technical field [0001] The invention relates to the technical field of visual target tracking, in particular to a KCF tracking target loss detection method and system based on foreground detection. Background technique [0002] Visual object tracking is one of the most challenging topics in the field of computer vision, which is widely used in video surveillance, human-computer interaction, smart cars, medical images and other fields. Currently, the most used trackers are based on correlation filtering algorithms. The main features of correlation filtering methods are low algorithm complexity, fast operation speed, and real-time requirements, including KCF trackers and their variants. At present, the missing detection methods for KCF tracking target mainly include KCF target loss detection method based on outlier detection and tracking target loss detection based on perceptual hash algorithm. [0003] When the target is affected by factors such as scale change, severe occlu...

AI Technical Summary

Problems solved by technology

[0003] When the target is affected by factors such as scale change, severe occlusion, similar target interference and out of view, the KCF algorithm will still lose the target

Once the target is lost, the tracker itself cannot detect it and does not have the ability to detect target loss, which is an inherent defect of the KCF tracking algorithm

During the trajectory tracking process, the loss of the tracking target will cause the trajectory to jump or shift

Images