Study on Groove Edge Extraction During Pulsed Argon Arc Welding Based on Improved Canny Algorithm and Hough Transform
The image collected by the passive light sensor during pulsed argon arc welding has the problems of fast light transformation,few target pixels,and it is difficult to extract the groove.A Canny algorithm based on improved threshold selection and double-threshold detection was proposed.The image was filtered by median filtering and bilateral filtering instead of Gaussian filtering in the traditional Canny algorithm,the number ratio of background pixels to foreground pixels was introduced to improve Otsu's maximum variance,and the threshold of the gradient value near the groove was calculated.The complete groove edge information was obtained by the optimized double-threshold detection,and the edge information was subjected to Hough transform to extract the clear bevel edge.The experimental results show that,compared with the traditional Canny algorithm,the improved Canny algorithm improves the success rate of groove extraction,and obtains more complete groove edge information.
NETL
NSTL
万方数据
