基于视觉处理的焊缝质量检测与控制系统  

A novel welding quality detecting and controlling system based on machine vision technique

作　　者：王振民[1] 张丽玲[2] 薛家祥[1] 

机构地区：[1]华南理工大学机械工程学院,广东广州510640 [2]南昌工程学院,江西南昌330099

出　　处：《焊接技术》2007年第3期60-62,共3页Welding Technology

基　　金：广东省自然科学基金资助项目(04020100)

摘　　要：采用视觉处理方法是实现焊接质量检测与控制的重要途径之一,而视觉处理算法是其可靠、准确应用的关键环节。采用CCD传感器直接摄像法获取焊接熔池的图像,利用数学形态学的图像处理方法,建立统一的非线性处理算法,对熔池图像进行预处理和特征提取,获取熔池边缘信息并确定焊缝中心和熔宽、熔高等信息,将这些信息与模糊逻辑相融合,实现高性能的跟踪和焊接质量PID模糊闭环控制。试验结果表明,基于视觉处理的焊接质量检测与控制系统具有良好的稳定性和准确性。The machine vision technique is one of the important methods for the detecting and controlling of the welding quality,while the vision processing algorithm is the key factor that affect the application of the machine vision technique.In the paper,the CCD sensor is adopted to obtain the image of the welding molten pool directly;and a universal non-linear image-processing algorithm was established based on the mathematic morphologic theory.The image of the molten pool is pre-processed and the characteristic information such as the center of the weld line,the width and length of the molten pool edge are obtained.According to the character information of the molten pool,the fuzzy inference and PID controlling theory are combined to form a closed-loop control system to control the penetration quality and weld line seam tracking.The experimental results indicate that the welding quality detecting and controlling system based on machine vision technique has perfect stability,self-adaptability and controlling precision.

关 键 词：视觉处理 数学形态学 熔池信息 焊接质量检测与控制 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] TG4[自动化与计算机技术—控制科学与工程]