Method for automatically detecting and tracking weld seam

Abstract

The invention belongs to the technical field of automatic welding and discloses a method for automatically detecting and tracking a weld seam. The method includes the following steps that workpieces (1) to be welded are spliced in advance, the gap between the workpieces (1) is formed, the width of the gap is 1.5 mm-3 mm, a bottom light source (3) is arranged underneath the workpieces (1), and an image collecting unit is used for extracting pre-calibrated trails of light transmitting the gap in an actual weld space coordinate system; then the workpieces (1) are tightly spliced to form the weld seam (2) of which the width is smaller than 1 mm; the image collecting unit is used for collecting images of the surfaces of the multiple workpieces (1), a search area containing the pre-calibrated trails is determined in each image, the weld seam (2) is searched for, and a welding gun (4) is guided to perform welding. The weld seam (2) can be found accurately through the bottom light source, the search range of the weld seam is greatly narrowed through determining one search area in each image, in this way, workloads of searching for the weld seam are reduced, working efficiency is improved, and welding by mistake is avoided.

Description

technical field [0001] A method for automatically detecting and tracking weld seams belongs to the technical field of automatic welding. Background technique [0002] In automatic welding, it is necessary to find the weld seam first, and then guide the welding torch to walk along the weld seam trajectory. At present, the weld seam on the surface of the workpiece is automatically identified through the image acquisition unit. At present, after the image of the workpiece surface is extracted by the camera, a very large area must be processed and processed. Finding the weld seam requires a lot of work and low detection efficiency. Moreover, the inventor found during the test that the surface of the workpiece is not smooth, and there are often scratches and indentations on the surface of the workpiece. The image acquisition unit can easily detect the scratches or indentations. Mistaking it as a weld seam increases the work of image processing, reduces the efficiency of automatic...

AI Technical Summary

Problems solved by technology

[0002] In automatic welding, it is necessary to find the weld seam first, and then guide the welding torch to walk along the weld seam trajectory. At present, the weld seam on the surface of the workpiece is automatically identified through the image acquisition unit. At present, after the image of the workpiece surface is extracted by the camera, a very large area must be processed and processed. Finding the weld seam requires a lot of work and low detection efficiency. Moreover, the inventor found during the test that the surface of the workpiece is not smooth, and there are often scratches and indentations on the surface of the workpiece. The image acquisition unit can easily detect the scratches or indentations. Mistaking it as a weld seam increases the work of image processing, reduces the efficiency of automatic welding, and even directly welds deep scratches or indentations by mistake, resulting in the scrapping of the entire steel plate, and even damage to the production equipment in severe cases

Images