Target tracking method based on soft characteristic theory

Abstract

The invention relates to a target tracking method based on a soft characteristic theory and belongs to the technical field of computer vision. The method comprises the following steps: initializing a target area; extracting soft characteristic information of a target; carrying out predecessor forecast on the soft characteristic information of the target; repairing the soft characteristic information of the target; carrying out target tracking. The target tracking method has the benefits that a moving target is tracked by the adoption of soft characteristics, and the anti-interference performance on target shape change and scale expansion and contraction is good, so that the method has higher accuracy and stability and well robust performance on the tracking of the seriously-deformed moving target; when the predecessor impact strength is adopted for forecasting a target predecessor, the problem of loss caused by the reason that the target is blocked can be better solved; as the target detection scope is reduced, the soft characteristic data amount is smaller, a plurality of attitude models of the target need not be stored, the calculation complexity and the space complexity are lower, and the tracking real time is higher.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to a target tracking method based on soft feature theory. Background technique [0002] Moving target tracking is an important and traditional research problem in the field of computer vision at home and abroad. It has been widely studied in recent decades. According to past research history and current research status, although target tracking has been developed for a long time Time research has also made great research progress, but the target tracking in complex backgrounds is still not well resolved. [0003] The research of target tracking technology mainly has two directions, one is the tracking method based on the analysis of motion information, and the other is the target tracking method based on model analysis. There are many methods of using motion information analysis to study target tracking, and many algorithms and standards have been formed. The ta...

AI Technical Summary

Problems solved by technology

There are many methods of using motion information analysis to study target tracking, and many algorithms and standards have been formed. The target tracking method based on motion analysis is simple in principle and convenient to implement, but it is difficult to distinguish the foreground and background information of the occluded target. Adhesion and loss of heel

Another model-based target tracking method can overcome the interference of local occlusion information, and is currently the mainstream method in the field of computer vision. When there is a large occlusion, deformation and scale movement, effective tracking cannot be carried out for a long time, and drift and loss of tracking are prone to occur

[0004] At present, although the target tracking technology has made significant progress, it still mainly adopts the classic technology in image / video processing, pays more attention to the target information in the current scene, and lacks the analysis and research on the target's predecessor information.

Images