Context-based target scale self-adaptive tracking method

Abstract

The invention provides a context-based target scale self-adaptive tracking method, and relates to the field of image tracking. A target scale adjustment mechanism is improved by using a target scale adjustment algorithm based on apparent characteristics and context information based on the existing scale direction self-adaptive mean drift algorithm. The method mainly comprises two parts of scale adjustment type determination based on the context information and scale calculation by scheduling an adjustment function and using apparent information and the context information. The context information of the target scale is introduced in the adjustment mechanism, and detailed classification is performed on the adjustment type according to the change of the scale context information on the basis of considering the apparent information so that the adjustment accuracy of the original algorithm for the target scale on the target numerical area and the effective area coverage can be enhanced, and the adjustment accuracy of the original SOAMS algorithm for the target scale can be effectively enhanced.

Description

technical field [0001] The invention relates to the field of image tracking, in particular to a target scale tracking method. Background technique [0002] According to the "2014 Statistical Bulletin on National Economic and Social Development", by the end of 2014, the number of civilian vehicles in China hit a record high, and traffic accidents and uncivilized driving behaviors frequently occurred. Facing increasingly complex management problems, how to track vehicles adaptively by the monitoring system has become the focus of current research on target tracking technology. Compared with the central position of the target, the real-time scale of the target is of great significance for the subsequent recognition and classification of the target. The current tracking algorithm can be divided into the tracking algorithm based on the appearance model and the tracking algorithm based on the motion model according to the difference of the target model used. It is more credible ...

AI Technical Summary

Problems solved by technology

[0008] In order to overcome the deficiencies of the prior art, the present invention aims at the problem that the SOAMS algorithm only uses the Bhattachary coefficient in the scale adjustment mechanism, which causes the problem that the target scale change cannot be accurately tracked during the target zoom-in process, and proposes an improved adjustment mechanism based on context information

Images