计算机视觉焊缝识别技术研究  

Study on the weld detection technology based on computer vision

作　　者：丁度坤[1] 高向东[1] 赵传敏[1] 

机构地区：[1]广东工业大学机电工程学院,广东广州510090

出　　处：《焊接技术》2006年第6期19-21,共3页Welding Technology

基　　金：国家自然科学基金资助项目(60375012);广东省自然科学基金资助项目(020176)

摘　　要：对焊缝识别技术进行了深入研究,包括对焊缝图像的中值滤波以降低噪声干扰、对图像的最佳阈值化以获取焊缝信息以及对图像的数学形态学处理以消除焊缝周围孤立点的干扰。在此基础上,提取到较为清晰的焊缝图像,进而采用最小二乘法建立焊缝的轨迹方程,为焊接控制提供依据,提高焊缝跟踪精度及焊接质量。A deeply research on the weld detection technology has been carried in this paper, including the filtering process of the weld images in order to reduce the disturbance of noise, best threshold cutting of the images to get the information of the weld and the process of the weld images based on the morphology technology to clear up the disturbance of the insulated points. A clearer weld image may be extracted on this base. Moreover, the tracking equation of the weld using the least square algorithm has been built, which can supply a basis to the later weld controlling. And it can improve the precision of weld tracking and welding quality.

关 键 词：计算机视觉 焊缝跟踪 图像处理 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置]