Moving object tracking method and device and computer equipment

Abstract

The invention relates to the technical field of computers, in particular to a moving target tracking method, device and equipment. The method comprises the steps of obtaining a first FHOG feature of atarget, and constructing position response output and size response output of Gaussian distribution function distribution; According to the FHOG characteristic, the position response output and the size response output, training to obtain a corresponding filter; Obtaining a video second frame selected area, obtaining feature information of the video second frame selected area, and calculating thefeature information and a position change filter according to a corresponding model to obtain a target motion position; Selecting a plurality of image areas with different sizes by taking the position center of the second frame of target as a center to obtain a plurality of sub FHOG features; And respectively calculating the plurality of sub FHOG features and a size change filter according to corresponding models to obtain a target size. According to the method, the adaptability to the scale change of the target is improved, and the target can be accurately tracked; And the algorithm reducesthe calculation amount and improves the real-time performance of the algorithm.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a moving target tracking method, device and computer equipment. Background technique [0002] Video object tracking can be used in vehicle monitoring, intelligent security, etc., and it has become an important research content of computer vision. [0003] Target tracking can be mainly divided into 5 parts, namely motion model, feature extraction, appearance model, target positioning and model root. Among them, the motion model can predict the area where the target may appear in the current frame based on the position of the target in the previous frame. Now most of them use The method of particle filter or correlation filter is used to model the target motion. The method based on correlation filter tracking uses the special properties of time domain transformation of circulant matrix to convert the budget into the frequency domain for calculation, which can speed up the trainin...

AI Technical Summary

Problems solved by technology

[0004] However, the current algorithm cannot adapt well to the transformation of the scale of the target itself and the transformation of the movement position, so the accuracy of target tracking cannot be guaranteed.

Images