DSP-based Correlation Tracking Method with Low Power Consumption

Abstract

The invention discloses a low-power-consumption DSP-based relevant tracking method. The target tracking method adopts the manner of combining an NProd (normalized product correlation)-based template matching algorithm with sheltering determination and motion estimation, a search center is set through Kalman motion estimation; a subtemplate correlation matrix based on NProd function is utilized as a matching metric, and a target is searched in a wave gate; and if the relevancy error between subtemplates meets the sheltering condition, template updating is suspended, and a motion estimation value is directly adopted as the best matching position till the target leaves the sheltering area. According to the tracking method, the target in a complex background, similar to the background color distribution and appearing large-area shielding suddenly can be tracked effectively; and the advantages of small size, low-power-consumption performance and strong applicability and the like can be achieved.

Description

[0001] This application is a divisional application with application number: 201310309831.2, application date: 2013.07.22, and name: "DSP-based correlation tracking method". technical field [0002] The invention relates to an image processing technology, in particular to a DSP-based target correlation tracking method and device thereof. Background technique [0003] Target tracking based on image processing has many practical applications in the fields of human-computer interaction, intelligent video surveillance and military applications. But the difficulty lies in how to make the tracking process more robust, because the shape change and occlusion of the target will affect the reliability of the tracking. [0004] The correlation tracking method based on template matching is a method in target tracking. The template image containing the target is obtained through automatic detection or manual setting, and then a certain criterion is used to find the best match in the sear...

AI Technical Summary

Problems solved by technology

The correlation tracking method has the advantages of high precision, stable tracking and insensitivity to the background color distribution, but when the target is deformed or occluded, the tracking accuracy will be reduced or even the tracking will fail.

By updating the template, the influence of target deformation on tracking can be suppressed to a certain extent, but the small error on each frame will gradually accumulate with the relative motion process, which may cause the target tracking point to deviate farther and farther from the original aiming center , that is, the template drifts

In addition, special attention should be paid to when the target enters the occlusion, if the template update is not stopped immediately, the occluder will enter the template, which will seriously affect the follow-up tracking

[0005] Existing technical systems are generally difficult to deal with sudden interference such as target occlusion. In order to more effectively improve the robustness of correlation tracking, research can be carried out from two aspects at the same time: first, construct a more robust matching metric function; second , study the appropriate template update method

Images