A method of underwater acoustic passive tracking of multiple targets based on YOLO algorithm

Through the YOLO algorithm and split beamforming technology, the identification and association problem of passive tracking of multi-objective water acoustic targets in the existing technology is solved, and the ability to identify multiple water acoustic targets in the same frame input data is realized, which improves the accuracy of detection degree and data association accuracy.

CN114706084BActive Publication Date: 2025-01-07HANGZHOU DIANZI UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202111661212.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-30
Publication Date
2025-01-07
Estimated Expiration
2041-12-30

AI Technical Summary

Technical Problem

The prior art is difficult to identify multiple water acoustic targets in the same frame input data, especially when the target signals overlap, and it is impossible to effectively associate multiple target data.

Method used

The YOLO algorithm is used to detect the multi-beam LOFAR spectrum, and the frequency range and azimuth center of different targets are identified through convolutional neural networks, and the precise azimuth angle is obtained through split beam formation, combined with Kalman filtering to estimate states to achieve accuracy of multiple targets. track.

Benefits of technology

Without extracting the target signal alone, the frequency and orientation range of multiple targets can be identified, which improves the accuracy of multi-object detection and the accuracy of data correlation, and can handle overlapping situations in frequency and orientation, compared with the current situation of current target detection. More efficient with technology.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114706084B_ABST
    Figure CN114706084B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for passive tracking of underwater acoustic targets based on the YOLO algorithm, and the steps include: S1, obtaining a multi-beam LOFAR spectrogram; S2, detecting the multi-beam LOFAR spectrogram through a YOLO neural network model to obtain the frequency range of different targets and the azimuth center of the target; S3, determining the data association of the azimuth beam pointing angle of the target and the target; S4, further obtaining the precise azimuth of the azimuth beam pointing angle of the target through split beam formation; S5, filtering and estimating the precise azimuth of the target, and then estimating the state of the target at time K; S6, repeating S1-S6 to update tracking information. The above technical solution is adopted to identify multiple targets in the same frame of input data, and the frequency range and azimuth range of the target are output, so that the azimuths of different targets can be identified; when detecting multiple targets, the YOLO algorithm gives the prediction probability of different targets, which can be used for data association, and its accuracy is higher than that of the existing data association method.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Method of measuring speed of ship by the adoption of seabed echo theory under circumstance of ship pitching

    CN103197320A

  • Azimuth trace acquisition method and device for flexible towed linear array sonar noise target

    CN112285720A