Fastener integrity detection method for unmanned aerial vehicle track fine inspection

Through the visually prior guided overall layout characterization method of fastener and the heuristic knowledge-guided detection network architecture, the missed detection problem caused by occlusion and shadow in railway fastener detection is solved, and the integrity and efficiency of fastener detection are achieved.

CN119693625AActive Publication Date: 2025-03-25BEIJING UNIV OF TECH
6 Cites 0 Cited by

Patent Information

Application Number
CN202411862736.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-17
Publication Date
2025-03-25
Estimated Expiration
2044-12-17

AI Technical Summary

Technical Problem

When detecting fasteners in railway track lines, the traditional general target detection algorithm cannot consider the inherent geometric relationship between fasteners, resulting in false negative cases and missed inspection problems in natural occlusion or severe shadows, and cannot guarantee the integrity of fastener detection.

Method used

A standard characterization method for overall layout of rail fasteners with visual prior guidance SVP-RFR and end-to-end heuristic knowledge-guided fasteners with overall detection network architecture HRFDet is proposed. Fastener detection from rough to fine by designing adaptive anchor frames and custom loss functions.

Benefits of technology

It effectively reduces the problem of missing fastener inspection caused by shading and shadow, ensures the integrity of fastener inspection, and provides inspection coverage and integrity indicators, supporting the optimization of drone inspection routes.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a fastener integrity detection method for unmanned aerial vehicle track fine inspection. The health condition of the steel rail fastener directly affects the stability of the position of the steel rail, the problem of missing detection of the fastener caused by shielding or shadow inevitably exists, and the specific position and statistical data of missing detection cannot be given. The method is established on a designed special standard embedding space, has good geometric interpretability, and can be closest to all the steel rail fasteners in the image, so that all the fasteners are uniformly distributed on the two sides of the steel rail along the direction of the steel rail. By using the fastener layout characterization method, an end-to-end heuristic knowledge guided fastener overall detection network architecture is provided, integrity detection from roughness to fineness of the steel rail fastener in an unmanned aerial vehicle aerial image in a track area is realized while the overall and integrity of fastener detection are ensured, and the detection accuracy of the fastener layout characterization method is improved. And the problem of fastener leak detection caused by heavy shadow or shielding of trees, catenary rods or steel rails is reduced to the greatest extent.
Need to check novelty before this filing date? Find Prior Art