A Novel Inclinometer-Integrated Structured-Light Weld Vision Sensor and Its Calibration Method and Weld Seam Recognition

Abstract

In scenarios such as crawling welding, achieving precise robot positioning is challenging, rendering the acquisition of weld seam spatial states through kinematics unfeasible. However, it is crucial to recognize weld seam inclinations that enable robots to plan appropriate welding process parameters. Currently, there is a notable absence of a modular sensor solution tailored for this purpose. This article presents the design and calibration method of an innovative low-cost inclinometer integrated structured-light vision sensor that can modularly gauge the weld seam inclinations in the world frame and its position and pose in the camera system. First, a method of calibrating the inclinometer pose based on data optimization is proposed utilizing external turntables and two-dimensional data from the inclinometer. Second, a description method of the weld seam coordinate system is designed, including two-dimensional seam-ground parameters to represent the weld seam inclination in space. Then, the method for determining the seam position and pose in the camera system as well as the seam-ground parameters are elucidated. Finally, experiments demonstrate that a calibrated sensor prototype exhibits satisfactory recognition results and satisfies welding requirements.