An unmanned aerial vehicle recognition method and system based on cross-modal perception fusion

CN122310201APending Publication Date: 2026-06-30YANGTZE DELTA REGION INST (QUZHOU) UNIV OF ELECTRONIC SCI & TECH OF CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
YANGTZE DELTA REGION INST (QUZHOU) UNIV OF ELECTRONIC SCI & TECH OF CHINA
Filing Date
2026-03-13
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing drone identification technologies mainly rely on single sensing methods, making it difficult to accurately distinguish between legitimate and illegal drones in complex environments. Furthermore, multimodal information is difficult to effectively integrate, resulting in insufficient identification accuracy and robustness.

Method used

By coordinating the processing of radio frequency sensing and visual sensing information at the temporal, spatial, and feature representation levels, cross-modal feature alignment and adaptive fusion are achieved, including temporal alignment, spatial correlation, and feature mapping, and weighted fusion is performed by combining radio frequency signal-to-noise ratio and visual confidence.

Benefits of technology

It improves the accuracy and robustness of UAV identification, adapts to complex environments, breaks through the bottleneck of shallow multimodal fusion, promotes the development of integrated sensing systems, and forms a complete technical system that can be engineered.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122310201A_ABST
    Figure CN122310201A_ABST
Patent Text Reader

Abstract

This invention belongs to, but is not limited to, the fields of wireless communication and intelligent sensing technology, and discloses a method and system for drone identification based on cross-modal sensing fusion. The method acquires radio frequency signals and visual information, extracts physical parameter features and semantic features of the drone respectively, and aligns and correlates them in the time and spatial domains, mapping the two types of features to a unified representation space. Semantic consistency is achieved through cross-modal feature alignment, and adaptive fusion is performed in conjunction with sensing reliability to complete drone category identification and physical parameter output. This invention maintains high recognition accuracy and robustness even in low signal-to-noise ratio, visual occlusion, and complex environments, effectively distinguishing between legitimate drones, illegal drones, and non-target objects. It is applicable to scenarios such as low-altitude security monitoring, airspace supervision, and intelligent drone management.
Need to check novelty before this filing date? Find Prior Art