Research Of Image Processing For Seam Track Based On Laser Optical Sensor With Structured Light

With the development of image processing technology and the improvement of computing speed, image processing is widely applied in industry. Vision sensor based on laser structured light uses triangulation principle which laser structured light plane projects onto the workpiece surface at a certain angle and form feature laser stripe. A CCD camera with filter obtain laser stripe image in weld area. Vision sensor is fixed in the front of seam tracking system. The welding image obtained by CCD camera was processed by welding image process and feature extraction algorithm to acquire weld location information. As the core of seam tracking technology, welding image process is extremely important in the seam tracking. Find a effective image processing algorithms has become the target of weld image processing technology and an important research content in automatic seam tracking field.Welding image process algorithm in this paper is based on dynamic ROI (region of interest treatment). Firstly, left ROI, right ROI and a central ROI which contains a complete laser light are acquired. Processing left ROI and right ROI , feature lines equations of left light band and right light band can be extracted. At the same time, the relevant parameters of groove ROI can be obtained. Processing groove ROI, we can obtain feature lines equations of tow oblique light bands. Three feature points'coordinate in the image domain can be solved through four simultaneous feature lines equations. In order to ensure accuracy of the feature line and the feature point, a judgment criterion to determine lines and points are proposed. When the criteria judges error or low accuracy of the feature line or the feature point, according to the previous six consecutive welding images'corresponding line or point using least square method to calculate the lines or points.Under CO2 gas shielded arc welding, seam tracking experiment is executed to test algorithm. Experiment shows image processing algorithms meet the real-time seam tracking, anti-interference performance and precision.