Weld three-dimensional reconstruction method based on two-dimensional linear structured light

Abstract

Provided is a weld three-dimensional reconstruction method based on two-dimensional linear structured light. The method includes following steps: (1) structured light image capturing and calibration of a weld: performing scanning along a target weld, shooting the target weld at intervals in a fixed time interval, then obtaining a two-dimensional structured light image of the target weld, performing image calibration on the structured light image, and obtaining the practical length and width of one pixel point in the structured light image; (2) structured light image processing of the weld: pre-processing the two-dimensional structured light image of the weld, then extracting a weld center line of the structured light image by employing a contour average method, and obtaining the practical coordinate of each pixel point of the weld center line; and (3) performing three-dimensional surface fitting by employing an interpolation method based on a B-spline through the practical coordinate obtained in step (2), of each weld center line to reconstruct a complete three-dimensional weld surface. According to the method, the reliability is high, the method is simple, and the detection and identification efficiency is high.

Description

technical field [0001] The invention relates to the field of automatic tracking of welding robots, in particular for realizing three-dimensional reconstruction of welding seams based on visual sensing and providing a basis for welding automation. Background technique [0002] In the study of using visual sensing technology to obtain the topographical characteristics of welds, monocular vision generally can only obtain two-dimensional plane information of the target weld. If the three-dimensional information of the target weld is to be obtained, stereo vision is required , the commonly used stereo vision methods mainly include binocular stereo vision and structured light vision. However, most of the above-mentioned methods are aimed at straight line welds or small curvature welds, and for complex multi-layer multi-pass welds, relatively sufficient research results have not yet been obtained. If binocular stereo vision is used, then two images are required. image matching, wh...

AI Technical Summary

Problems solved by technology

[0003] In order to overcome the disadvantages of low reliability and low detection and recognition efficiency of the existing three-dimensional reconstruction methods of welds, the present invention provides a welding method based on two-dimensional line structured light with high reliability, simple method and high detection and recognition efficiency. Seam 3D Reconstruction Method

Images