改进Canny算子在焊缝边缘检测中的研究  

Research on Improved Canny Operator in Weld Edge Detection

作　　者：吴房胜[1] 李如平[1] 施冬冬 WU Fangsheng;LI Ruping;SHI Dongdong(College of Information Engineering,Anhui Business and Technology College,Hefei 231131,China)

机构地区：[1]安徽工商职业学院信息工程学院,安徽合肥231131

出　　处：《新乡学院学报》2020年第9期46-49,共4页Journal of Xinxiang University

基　　金：安徽省高校自然科学研究重点项目(KJ2019A1166,KJ2019B001);安徽省高校学科拔尖人才学术资助重点项目(gxbjZD2016094);安徽省高校优秀青年骨干人才国内访学研修项目(gxgnfx2020133)。

摘　　要：因为管道焊接优劣主要取决于管道焊缝是否有缺陷,所以对焊缝缺陷轮廓的精确识别直接影响到管道的质量。分析了图像边缘处理的Canny算子,并对Canny算子进行了改进,提出用基于数学形态学的滤波算法替代传统的高斯滤波方法来平滑焊缝缺陷轮廓。采用3×3邻域有限差分算法,添加两个方向的梯度幅值计算以提高轮廓边缘检测精度。提出最大熵自适应阈值算法,用自适应阈值代替人为设定的阈值。实验结果表明:算法对焊缝缺陷轮廓边缘提取的封闭性比传统算法好,且抗噪能力强。Because the quality of pipeline welding mainly depends on whether the pipeline weld is defective, the precise identification of the weld defect contour directly affects the quality of the pipeline. The Canny operator of image edge processing was analyzed and the Canny operator was improved. A filtering algorithm based on mathematical morphology was proposed to replace the traditional Gaussian filtering method to smooth the contour of weld defects. Using the 3×3 neighborhood finite difference algorithm, the gradient amplitude calculation of two directions was added to improve the detection accuracy of the contour edge. The maximum entropy adaptive threshold algorithm was proposed, which replaced the artificial threshold with the adaptive threshold. The experimental results show that the algorithm is better than the traditional algorithm in sealing the edge of weld defect contour and has strong antinoise ability.

关 键 词：改进Canny算子 形态学滤波 轮廓边缘 阈值自适应 

分 类 号：TP23[自动化与计算机技术—检测技术与自动化装置]