Multi-target tracking method based on Leiyu fusion

Abstract

The invention discloses a multi-target tracking method based on thunder-vision fusion, and the method comprises the steps: obtaining the state information of a plurality of targets of a radar frame through the analysis of a radar signal; an image acquired by the video monitoring equipment is input to the convolution model for target identification, and video state target state information is obtained; transforming the radar signal and the target signal into the same coordinate system by using perspective transformation, establishing data association between a radar target and a video target, fusing the thunder information and storing the thunder information into a tracker; predicting the state of the target in the tracker at the next moment, associating the data information of the prediction target and the observation target, and carrying out track matching; and filtering and updating the associated signals by using a filtering algorithm and a parameter threshold to complete multi-target tracking. According to the method, radar and video information are fused, the problems of target inaccuracy caused by a single information system, inaccurate detection of a static target by the radar, multi-detection of a large target, easy interference of ambient light on single video detection and the like are solved, and more comprehensive and accurate target detection is realized.

Description

technical field [0001] The invention belongs to the field of intelligent traffic detection, in particular to a multi-target tracking method based on multi-source information, in particular to a multi-target tracking method based on Lesion Vision Fusion. Background technique [0002] In recent years, with the increasing status of "smart city" construction, traffic monitoring systems have become more and more widely used. In order to reduce the incidence of traffic accidents, reduce traffic casualties and traffic losses, it is necessary to accurately identify and track the monitoring area. [0003] Traditional target tracking is based on the analysis of signals collected by a single sensor, which has various drawbacks. Millimeter-wave radar detects and tracks targets by transmitting and receiving modulated continuous waves in the detection area, and can obtain target position and speed information more accurately. When the target is stationary or large in size, the signal dis...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the shortcomings of the above-mentioned single detection and tracking method, to provide a multi-target tracking method based on Lesion Vision Fusion, using a detection method based on deep learning to detect video targets, using cluster analysis to achieve radar point detection, and radar fusion and video detection information to achieve accurate tracking of multiple targets

Images