Multi-target tracking method based on multi-model fusion and data association

Abstract

The invention discloses a multi-target tracking method based on multi-model fusion and data association; the tracking method comprises the following steps: firstly using an interframe difference method to detect a motion target contour and center of mass coordinates; fusing a pyramid optical flow method with Kalman filtering so as to predict the center of mass coordinates of the motion target in the next moment; using Euclidean distances between the center of mass coordinate predicted value and the center of mass coordinate detection value at next moment to form a benefit matrix, and using a Hungary algorithm to obtain the optimal matching through data association; finally removing certain portion unable to satisfy requirements in a tracker, and building a tracking unit for non-assigned detections, thus realizing multi-target tracking. The tracking method can be less affected by light changes and background noise interferences, thus solving the tracking failures caused by target blocking or mutual interferences between targets, providing multi-target tracking accuracy, and providing well instantaneity and robustness.

Description

technical field [0001] The invention belongs to the application fields of image processing, video detection and artificial intelligence cross technology, and in particular relates to a multi-target tracking method based on multi-model fusion and data association. Background technique [0002] Multi-target tracking is a research hotspot and difficulty in the field of computer vision, and it has important application value in intelligent traffic control, intelligent video surveillance and other fields. Due to the complexity of the real environment, problems such as background noise and target occlusion need to be solved urgently. The current tracking algorithms are mainly used: model-based tracking, target contour-based tracking, region-based tracking and feature-based tracking. [0003] (1) Model-based target tracking: Firstly, it is necessary to obtain the prior information of the tracking target to model the structure and motion state of the target. Although it can achieve...

AI Technical Summary

Problems solved by technology

[0003] (1) Model-based target tracking: Firstly, it is necessary to obtain the prior information of the tracking target to model the structure and motion state of the target. Although it can achieve better results, if it cannot obtain enough information about the target, it will fail. The tracking effect deteriorates, and at the same time, it cannot meet the real-time requirements

[0004] (2) Target tracking based on target outline: Because of the robust invariance of the outline information, the outline of the object is used to represent the moving target, and it is continuously updated, which has strong anti-light transformation characteristics, but in the case of a relatively complex background There will be a tracking failure

[0005] (3) The method based on area tracking: establish a target template and track the target according to the target. When the moving target is blocked, it will not be able to continue to track accurately, and the target will not automatically recover when it reappears, causing the tracking algorithm to fail.

This algorithm can maintain good tracking when the target is occluded, but if the feature points of the target change, such as scaling, rotation, etc., the tracking effect will be affected

Images