作　　者： ; ; ; ;

机构地区： 华南理工大学机械与汽车工程学院

出　　处： 《焊接技术》 2009年第1期37-40,共4页

摘　　要： 根据机器人水下焊缝跟踪的要求,设计了一套双目立体视觉传感器。采用卤钨灯辅助光源加复合滤光系统,较好地去除了噪声,拍摄到较为清晰的焊缝图像。针对水下焊缝图像模糊、对比度低的特点,采用模糊增强图像处理技术,对Pal处理算法进行了改进,改进算法中采用OTSU法选择最佳阈值,提出了新的隶属度函数。试验表明,该算法提高了图像处理的效率,提取的边缘比较精细,为空间焊缝跟踪中焊缝识别及立体匹配奠定了基础。 According to the requirements of robot underwater seam tracking,a binocular stereo vision sensor was developed.An auxiliary light source of halogen Tungsten Lamp and suitable filter was used.Noise was removed and good seam image could be captured by this sensing system.Due to the characteristics of fuzzy and low contrast about underwater welding seam images,a fuzzy enhancement image processing technology was presented.And Pal processing algorithms was improved.OTSU algorithm method was used to select the best threshold and a new membership function was proposed.The experiment indicated that the algorithm was an efficient and valuable edge detection algorithm,which was helpful for seam identification and stereo matching in space seam tracking.

关 键 词： 水下焊缝跟踪 双目视觉 模糊增强 边缘检测

领　　域： [金属学及工艺] [自动化与计算机技术] [自动化与计算机技术]