Target tracking method and device

Abstract

The invention provides a target tracking method and a device, which relate to the technical field of target tracking processing. The method comprises the steps of: based on the background informationof video images, establishing a sparse representation model based on target subspace and background subspace is established; by processing the target image region, extracting the gray-scale feature, HOG feature and LBP feature of the target, and selecting the distinguishing feature, combining multiple features to represent the target sparsely, and considering the relationship between the candidatesamples, and selecting the best candidate samples to achieve the accurate tracking of the target. The sparse representation is applied to target tracking, which not only takes into account the globalinformation of samples, but also takes into account the local information and spatial information of targets, so as to increase the accuracy of target tracking. At the same time, according to the sparse representation model of the target structure, the candidate targets are classified to improve the robustness and adaptability of the target tracking method in complex environments such as occlusion, background interference, illumination change, fast motion and so on.

Description

technical field [0001] The present invention relates to the technical field of target tracking processing, in particular to a target tracking method and device. Background technique [0002] The process of visual target tracking is to process the video or image sequence, perform feature extraction, target positioning, tracking detection and classification recognition on the target of interest, and finally obtain the position or motion characteristics of the target. This information can be used to perform more advanced vision tasks, such as object behavior recognition, scene understanding, motion pattern classification, object behavior understanding, etc. Visual object tracking technology is one of the hotspots in machine vision research. This technology is widely used in the fields of intelligent video surveillance, medical diagnosis, driving assistant, behavior recognition and human-computer interaction. Although researchers have proposed various object tracking methods, ...

AI Technical Summary

Problems solved by technology

This type of method can adapt to tracking problems such as light changes, attitude changes, and partial occlusions, but it is prone to tracking drift.

Images