Long-time target tracking method based on multi-correlation filtering model

Abstract

The invention relates to a long-time target tracking method based on a multi-correlation filtering model, and belongs to the field of computer vision. The method comprises the following steps: S1, extracting HOG and HOI features of a video image, and training a long-term correlation filter; s2, during a tracking process, judging whether target tracking fails or not by utilizing a maximum responsevalue generated by a long-time correlation filter and a target and a detection threshold value; if the target tracking succeeds, estimating the translation of the target by adopting an optimal displacement correlation filter in an MCCT algorithm and obtaining the position information of the target, and if the target tracking fails, activating an online detector to reposition the target and takingthe detection result of an online classifier SVM as the position information of the target; s3, after determining the translation position of the target, determining the scale of the target in the frame by using a scale correlation filter; and S4, finally, updating the filter model under the condition of meeting the target updating condition. According to the invention, the time overhead is reduced, and the performance is superior.

Description

technical field [0001] The invention belongs to the field of computer vision and relates to a long-term target tracking method based on a multi-correlation filtering model. Background technique [0002] Target tracking is to predict and determine the target position in the next frame according to the known target position of interest, so that the target information can be updated continuously through continuous tracking. [0003] Early research mainly focused on the learning of generative methods, but generative methods did not consider background factors, and tracking would fail when faced with interference from some complex scene factors. Afterwards, researchers consider the background information of the target through the perspective of discriminative classification, and discriminative-based tracking methods have been widely studied for their excellent performance. Among them, the tracking based on correlation filter has the dual advantages of accuracy and speed, so it h...

AI Technical Summary

Problems solved by technology

[0008] In view of this, the object of the present invention is to provide a kind of long-term target tracking method based on multi-correlation filter model, 1) improve the problem that the tracking effect of the appearance model of single feature training is not good, the improved training model has improved the accuracy of tracking ; 2) Improve the problem of updating the filter every frame of the traditional filter, the improved update mechanism avoids the model from being polluted when the target encounters an occlusion; For the problem of short-term tracking, the improved tracking algorithm can perform long-term target tracking, and improve the accuracy and success rate of tracking

Images