A target tracking method based on local features and scale pooling

Abstract

The invention discloses a target tracking method based on local features and scale pools, which belongs to the technical field of grayscale image target tracking and solves the problem of poor accuracy of target tracking algorithms in complex environments such as illumination changes, scale changes and background interference. In the present invention, the first frame of image obtains the target according to the initial information, and the classifier is trained based on the two characteristics of the target, and the target model and the classifier regression coefficient are re-initialized. and feature 2; based on the initialized target model and the classifier regression coefficient, the multi-layer kernel correlation filter response map of the two features is obtained, and then linearly interpolated to a consistent size, weighted and fused to obtain the multi-layer kernel correlation filter response map, and then the target's correlation filter response map is obtained. The predicted position and the predicted scale are to complete one target tracking. If the tracking is not over, the tracking from the second frame image to the third frame image is realized until the loop reaches the last frame image. The present invention is used for target tracking.

Description

technical field [0001] A target tracking method based on local features and scale pools is used for target tracking and belongs to the technical field of grayscale image target tracking. Background technique [0002] Target tracking is of great significance and value in the field of computer vision research. It is widely used in many fields, such as intelligent video surveillance, medical treatment, human-computer interaction and other civil fields. In the military, fast and accurate tracking of enemy moving targets search and tracking, etc. Object tracking is mainly divided into generative models and discriminative models. The generative model completes the matching between the candidate target and the target model by establishing the target mathematical model, and takes the most similar candidate region as the prediction target. The discriminative model trains the classification algorithm with a training set consisting of positive samples belonging to the target and nega...

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned research problems, the purpose of the present invention is to provide a target tracking method based on local features and scale pools to solve the problem of low robustness and poor accuracy of target tracking algorithms in complex environments such as illumination changes, scale changes, and background interference. question

Images