Vision-Based Measurement Of Weld Pool Surface Depression In TIG Welding

As an important manufacturing technology, tungsten-inert-gas (TIG) welding is widely used in industry. Sensing and measuring the information of welding quality is the prerequisite for realizing TIG welding process automation and making it intelligent. There is strong correlation between the weld pool surface depression and the penetration extent of the workpiece. Detecting the weld pool geometry and its surface depression from top-side of workpiece with vision sensor is of great theoretical and practical significance, because it can verify the numerical simulation results of welding thermal process, and it is the basis of welding process controlling.LaserStrobe technique is used to eliminate the strong arc light interference. A novel vision sensor is used and a measuring system is developed to sense the weld pool geometry and its surface depression by means of LaserStrobe technique. Parameters such as the distance from LaserStrobe to the workpiece, the aperture, the slope angle of camera, the position and slope angle of grid and frost-glass are adjusted. The images of distorted laser stripe on weld pool are captured. Imaging principle and characteristic are analyzed.An algorithm of software for processing and analyzing weld pool images is developed. The gray distribution characteristics of the images are analyzed. According to the different characteristics of gray distribution of laser stripe and weld pool edge, selective masking smoothing method, median filter and different dynamic search algorithm are developed in order to detect the laser stripe and weld pool edge.Mathematic model of pinhole camera is established, the camera is calibrated, and the affection factors of calibration error are analyzed. The weld pool geometric parameters are obtained based on calibration result.