Multi-scale vehicle tracking method and device based on road surface texture context information

Abstract

The invention relates to a multi-scale vehicle tracking method based on road surface texture context information, and the method comprises the steps: S1, obtaining the central position of a target vehicle under the condition of a linear space road texture, and S2, obtaining the central position of the target vehicle in a dual space under the condition of a nonlinear space road texture; and S3, obtaining the central position of the target vehicle, and combining the central position with the optimal scale of the image of the current frame to obtain a more accurate target vehicle position. The invention further discloses a multi-scale vehicle tracking device based on the road surface texture context information. According to the road surface texture area at the bottom of the target vehicle, in the moving process of the target vehicle, the relative position of the target vehicle and the road surface cannot be greatly changed, the road surface texture is stable; according to the road surface texture information, namely, the target vehicle is accurately positioned according to the relative relation between the target vehicle and the road surface area, and the target frame is prevented from drifting.

Description

technical field [0001] The invention relates to the field of computer vision, and more particularly to a multi-scale vehicle tracking method and device based on road surface texture context information. Background technique [0002] Machine learning (Machine Learning, ML) is a multi-field interdisciplinary subject, involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines, and target tracking is an important application in the field of machine learning. [0003] The invention patent with the publication number "CN108776974B" (the application date is 2018.05.24) discloses "a real-time target tracking method suitable for public transportation scenarios. The initial position P(i) on the current i-th frame; Step 2, use P(i) to train the correlation filter tracker; Step 3, obtain the image of the target in the i+1th frame; Step 4, use the correlation filter and the Correlation calculations are performed on ...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is to provide a multi-scale vehicle tracking method and device based on road surface texture context information to solve the problem of low accuracy of the target vehicle position in the above-mentioned background technology

Images