基于主动视觉的焊缝成形尺寸测量和缺陷识别研究  

Study on Dimension Measurement and Defect Detection of Weld Based on Active Vision

作　　者：褚慧慧 王宗义[1] 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《热加工工艺》2017年第21期206-209,共4页Hot Working Technology

摘　　要：为了实现焊缝成形质量的在线检测,建立了一个基于主动视觉的焊后焊缝成形质量检测系统,同时对整个系统最重要的部分-图像处理进行了深入探讨。在图像处理过程中,首先进行图像预处理;在特征提取部分,用质心法提取结构光中心线;然后用RANSAC算法来拟合焊接母材;用面积法来提取结构光中心线上的特征点。最后得到焊缝成形的几何参数,还可以利用视觉检测系统得到的三维信息进行焊缝三维形貌重建。试验结果表明,该焊缝成形质量检测系统的图像处理方法可靠性高、鲁棒性强、运算速度快。In order to realize the weld quality on line detection, a weld forming quality inspection system based on active vision was established, and the image processing as the most important part of inspection system was also discussed. In the image processing, the image preprocessing was performed. And in the feature extraction part, the structured light center line was extracted by using centroid method, and the RANSAC algorithm was used to fit the base metal. The area method was used to extract characteristic points on the structured light center line. Finally, the geometric parameters of the weld forming were obtained and the 3D weld surface was reconstructed by using the three dimensional information obtained by visual inspection system. The experimental results show that the image processing method of the weld forming quality inspection system has high reliability, robustness and operation speed.

关 键 词：机器视觉 焊缝成形 结构光 特征点提取 

分 类 号：TG409[金属学及工艺—焊接]