Multi-target unmanned aerial vehicle tracking method under multi-view field condition of unmanned aerial vehicle cluster

By numbering and matching the visual features of multiple drones in a drone swarm, and utilizing the visual information of multiple drones, the problem of tracking multiple targets in a multi-field-of-view scenario is solved, achieving higher accuracy and real-time performance.

CN116501084BActive Publication Date: 2026-03-24BEIJING INST OF TECH
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-04-12
Publication Date
2026-03-24

AI Technical Summary

Technical Problem

Existing UAV target tracking methods are mainly designed for single-target UAVs under a single field of view, and cannot effectively achieve multi-target UAV tracking under multiple fields of view, especially continuous tracking of multiple targets in UAV swarms.

Method used

This paper proposes a multi-target UAV tracking method under multi-field-of-view conditions in UAV swarms. The UAV swarm is numbered, and the master and slave UAVs respectively perform video shooting and target detection. Visual features and target position matching are used, and Kalman filtering and Hungarian algorithm are combined to perform target matching and identity number correction, making full use of the visual information of multiple UAVs to supplement the method.

Benefits of technology

It improves the accuracy and tracking speed of multi-target tracking, enables better real-time tracking, and enhances the performance of multi-target tracking algorithms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116501084B_ABST
    Figure CN116501084B_ABST
Patent Text Reader

Abstract

The application discloses a multi-target unmanned plane tracking method under multi-view field conditions of an unmanned plane cluster, and comprises the following steps: numbering unmanned planes in the unmanned plane cluster, wherein one unmanned plane is a master machine, and the rest unmanned planes are set as slave machines; each unmanned plane shoots a target cluster through a video, and carries out target detection on the shot video clip; and the master machine and the slave machines track the target respectively. The multi-target unmanned plane tracking method under multi-view field conditions of the unmanned plane cluster disclosed by the application fully utilizes visual information obtained by multiple unmanned planes, and uses visual information of other unmanned planes to supplement tracking performance of a current tracking unmanned plane, so that the precision and tracking rate performance of multi-target tracking are improved.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Method for tracking motion state of aerial target by unmanned aerial vehicle cluster

    CN114627151A