High-speed railway rail profile detection method and device

The high-speed railway rail profile detection method is determined by combining the spatiotemporal context algorithm and DBSCAN clustering with the least squares method, which solves the real-time and robustness problems of high-speed railway rail profile detection and achieves high-precision light bar center extraction.

CN114119957BActive Publication Date: 2025-09-09CHINA ACADEMY OF RAILWAY SCI CORP LTD +2
1 Cites 0 Cited by

Patent Information

Application Number
CN202111137538.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-09-27
Publication Date
2025-09-09
Estimated Expiration
2041-09-27

AI Technical Summary

Technical Problem

Existing high-speed railway rail profile detection methods cannot meet the real-time detection requirements at a speed of 350 km/h, and are sensitive to rail grinding and changes in external ambient light, resulting in inaccurate extraction of light bar centers, affecting detection accuracy and robustness.

Method used

The spatiotemporal context algorithm is used to determine the pixel points in the rail area of ​​interest, the DBSCAN clustering algorithm is used to remove interference point clusters, the maximum gray value point sequence and the least squares method are used to determine the center of the light stripe, and the standard rail template is used for detection.

Benefits of technology

The accuracy and robustness of detection are improved, ensuring real-time performance and anti-interference capabilities in high-speed railway environments. The average deviation of the sub-pixel light strip center results is 0.0581 pixels, and the mean square error is 0.0460 pixels.

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

Abstract

The present invention discloses a method and device for detecting the profile of a high-speed railway rail, wherein the method comprises: obtaining a high-speed railway rail image; determining the pixel points of the rail region of interest using a spatiotemporal context algorithm based on the high-speed railway rail image; scanning the pixel points of the rail region of interest using a DBSCAN clustering algorithm to determine the interference point clusters of the rail region of interest; extracting the maximum grayscale value point sequence from the rail region of interest after removing the interference point clusters to obtain the initial value of the light bar center; determining a first center point sequence based on the initial value of the light bar center and a standard rail template; determining a second center point sequence based on the first center point sequence using a least squares method; and performing high-speed railway rail profile detection based on the first center point sequence and the second center point sequence. The present invention can perform high-speed railway rail profile detection, improve detection accuracy, and ensure real-time and robustness.
Need to check novelty before this filing date? Find Prior Art