Unmanned aerial vehicle shouting method and device based on target identification, unmanned aerial vehicle, and storage medium

By performing scene planning and target recognition on the drone broadcasting system, and utilizing the YOLOv9 model and voice mapping rule base, accurate broadcasting by drones in complex scenarios was achieved. This solved the shortcomings of the existing system in terms of rapid response and multi-module collaboration, and improved information transmission efficiency and control accuracy.

CN122115177APending Publication Date: 2026-05-29GUANGZHOU IMAPCLOUD INTELLIGENT TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Applications(China)
Current Assignee / Owner
GUANGZHOU IMAPCLOUD INTELLIGENT TECH CO LTD
Filing Date
2026-02-09
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing drone-based loudspeaker systems struggle to respond quickly in complex scenarios, their limited voice content fails to meet the diverse needs of different scenarios and target objects, and the lack of efficient multi-module collaboration results in low information transmission efficiency.

Method used

By planning the pre-inspection area and dividing the alarm area, the visual sensor and YOLOv9 target detection model on the drone are used to identify the target, and the violation is determined in combination with the control rules. The appropriate audio is then matched and played from the preset voice mapping rule library.

Benefits of technology

It enables accurate voice communication in complex scenarios, improves information transmission efficiency and on-site control precision, ensures the adaptation of voice content to the scene and target, and reduces manual intervention and delays.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122115177A_ABST
    Figure CN122115177A_ABST
Patent Text Reader

Abstract

Embodiments of the present application provide a target recognition-based UAV shouting method and device, a UAV and a storage medium, the method comprising: scene planning for a pre-inspection area, dividing at least one alarm area, controlling the UAV to collect a video stream of the alarm area, determining coordinate information of a target to be identified in the video stream, determining classification confidence of the target to be identified in the video stream, determining whether a violation occurs based on the coordinate information of the target to be identified, the control rules configured for the alarm area, the control range and the classification confidence, and when it is determined that a violation occurs, matching corresponding shouting audio according to the alarm area and the violation, and controlling the UAV to play the shouting audio in the direction of the target to be identified. The scene image is collected by the visual sensor carried by the UAV, and the target to be identified is identified by a deep learning model; according to the identification result, the adaptive audio is intelligently generated in combination with the control rules corresponding to the alarm area, so that accurate shouting based on the scene is ensured.
Need to check novelty before this filing date? Find Prior Art