An Imaging and Measurement System for Robust Reconstruction of Weld Pool During Arc Welding

Abstract

Robust measurement of the specular weld pool surface can help better understand the complex welding processes and provide feedback for robotic welding. The strong arc light and the specular surface of the weld pool make it difficult for the direct sensing measurement. In this paper, a novel imaging and measurement system is proposed, and it makes use of the strong penetrability of the laser to avoid the interference of arc light and obtain the surface information of the weld pool. By intercepting the reflection of a projected laser pattern twice, the proposed system gives a closed-form solution for each reflected ray and a closed-form solution for the corresponding point on the weld pool surface. The weld pool surface can be thus reconstructed by one-shot structured light projection. To increase the robustness of the measurement system, the least deformation principle is proposed to rectify the coefficient errors of the computed plane equation. Finally, a sequence of 3-D weld pools is reconstructed and compared with state-of-art literature. The comparison verifies that our method is significantly better than all the other methods in measuring the 3-D weld pool shapes.

Single-shot three-dimensional reconstruction based on structured light line pattern

“…Based on the surface characteristics of the object, the non-contact three-dimensional surface imaging techniques could be divided into two categories: techniques to measure the specular surfaces [1][2][3][4][5] and techniques to measure the diffuse surfaces [6][7][8][9][10][11][12][13][14][15][16][17][18] . No matter for the specular surfaces or the diffuse surfaces, the robust measurement with one-shot is very challenging, however, it is highly desired especially when the measured surfaces are dynamic and deformable.…”

“…No matter for the specular surfaces or the diffuse surfaces, the robust measurement with one-shot is very challenging, however, it is highly desired especially when the measured surfaces are dynamic and deformable. As a result, many one-shot methods have been proposed for specular surface measurement [1][2][3][4][5] and diffuse surface measurement [11][12][13][14][15][16][17][18] . Each proposed method has its own merits or limitations and its application is also limited.…”

Automatic Segmentation of Fuzzy Laser Lines with Sub-Pixel Accuracy from the Uneven Background During Robotic Arc Welding

“…However, it lacks accuracy due to the errors caused by finding corresponding points in the two images which has remained a challenging problem in stereo vision. In [4] and [5], the authors make use of the strong penetrability of laser to reduce the arc light effect and utilize the specularity of the weld pool. The dot pattern laser rays are projected onto the specular weld pool and reflected onto the imaging plane.…”

“…The dot pattern laser rays are projected onto the specular weld pool and reflected onto the imaging plane. [4] estimated the weld pool shape based on the slope while [5] computes the weld pool shape with closed form solution. In [6], grating image instead of laser dots was used to estimate the weld pool shape.…”

“…In [7], the authors made use of the thermal emission to estimate the weld pool shape. Among all these state of art methods, only [5] yields a closed form solution and achieves accurate measurement of gas tungsten arc welding (GTAW) weld pool shape. The vision system developed in this paper can be extended to achieve accurate measurement of gas metal arc welding (GMAW) weld pool shape with closed form solution as [5].…”