A multi-target tracking method based on radar and vision fusion

CN114299417BActive Publication Date: 2025-04-04LIANYUNGANG JARI ELECTRONICS CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202111499048.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-09
Publication Date
2025-04-04
Estimated Expiration
2041-12-09

AI Technical Summary

Technical Problem

The traditional single detection and tracking method has problems such as inaccuracy, heavy impact on weather and light, and limited information in the process of target detection and tracking, making it difficult to achieve accurate tracking of multiple goals.

Method used

The multi-objective tracking method based on lightning vision fusion is adopted to detect video targets through deep learning, and the radar point detection information is analyzed using cluster analysis, and the radar and video detection information are fused to achieve accurate tracking of multiple targets.

Benefits of technology

This method can provide accurate data in large detection areas, ignore the impact of climatic conditions, provide on-site image information, and achieve accurate tracking of multiple targets, enhancing the robustness and comprehensiveness of multi-objective identification tracking.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114299417B_ABST
    Figure CN114299417B_ABST
Patent Text Reader

Abstract

The present invention discloses a multi-target tracking method based on radar-visual fusion, which obtains multiple target state information of radar frames by analyzing radar signals; inputs images obtained by video monitoring equipment into a convolution model for target recognition to obtain video-state target state information; transforms radar signals and target signals to the same coordinate system using perspective transformation, establishes data association between radar targets and video targets, fuses radar-visual information and stores it in a tracker; predicts the state of the target in the tracker at the next moment, associates the predicted target with the observed target data information and performs track matching; uses a filtering algorithm and a parameter threshold to filter and update the associated signal to complete multi-target tracking. The method of the present invention integrates radar and video information, overcomes the problems of target inaccuracy brought by a single information system, radar inaccuracy in static target measurement and multiple measurement of large targets, and single video detection is susceptible to interference from ambient light, and realizes more comprehensive and accurate target detection.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Image millimeter-wave radar fusion target detection method

    CN111652097A

  • Vehicle detection and tracking method based on radar signal and visual fusion

    CN112991391A