Steel rail abrasion rapid measurement algorithm based on structured light

Abstract

The invention discloses a steel rail abrasion rapid measurement algorithm based on structured light. A steel rail comprises a top straight line segment L1, a middle rail web M1, a bottom rail web M2 and a bottom straight line segment L2, and the method comprises the steps of a) constructing a contour through point cloud, establishing contour coordinate data, and positioning a basic contour line ofthe contour; b) carrying out accurate point cloud fitting on basic contour mapping; c) extracting basic circle centers of the middle rail web M1 and the bottom rail web M2, intersection points of thetop straight line segment L1 and the bottom straight line segment L2 and rail jaw points thereof; d) constructing coarse registration of affine transformation compatible deformation; e) carrying outaccurate registration based on a kd-ICP algorithm; and f) measuring the corresponding detection position after registration. According to the detection algorithm provided by the invention, a recoverymode for deformation contours by noisy non-smooth point cloud data processing, reference invariant region curvature extraction and key feature point extraction is provided, and the precision is improved while the registration speed is ensured.

Description

technical field [0001] The invention relates to the field of rail wear detection, in particular to a fast measurement algorithm for rail wear based on structured light. Background technique [0002] Due to the long-term high-load operation of the train, it brings increasing wear and tear to the rail, which brings great safety risks to the normal operation of the train. Accurate and efficient detection of rail wear is extremely important. At present, due to the rapid development of structured light technology, especially the increase in frame rate, its advantages of non-contact, high-precision, and high-efficiency dynamic detection are widely used in the detection of the rail industry. In recent years, the application cases of structured light on rail wear have increased year by year, but the conventional method is to obtain the center of the middle rail waist and the bottom rail waist, and combine the rail jaw points to form the reference point of the rotation transformation...

AI Technical Summary

Problems solved by technology

[0003] Because the track is in a natural scene, the data acquired by structured light often interferes a lot, resulting in poor data stability; and it will lead to misallocation of attribution point data in the middle and bottom rail waists of the segmentation, resulting in inaccurate fitting of the center of the circle , in addition, because the center angle of the bottom rail waist, especially the middle rail waist is very small, when the center angle is less than 20°, using the conventional least squares to fit the center of the circle will bring a large error; similarly, using the ICP algorithm based on the principle of the closest distance It is sensitive to the initial position of the point set to be paired. When the initial position of the point set to be registered is relatively close to the position of the template point set, the registration effect will be better. Otherwise, the nearest point pair will be changed during each calculation iteration. The ICP algorithm tends to converge to a local optimum rather than a global optimum

Images