【摘要】 针对传统Canny算法在识别直缝焊管焊缝边缘时易受噪声干扰、强弱边缘连接不佳等问题，通过分析直缝焊管图像特点，提出一种改进的Canny算法检测直缝焊管焊缝位置方法。首先在进行非极大值抑制中引入了梯度方向局部均值偏差；其次对非极大值抑制之后的点利用大津法计算阈值；最后通过实际采集焊管图片强弱边缘分布特点将强边缘边界所连边界作为附加连接条件。通过实验验证改进的非极大值相对于传统方法抑制能够有效的抑制噪声点并保留边缘信息，在直缝焊管焊缝边缘位置识别中改进的Canny算法相对于传统边缘检测算法能够抑制噪声干扰，保留更多的焊管边缘与焊缝边缘，因此较好的满足实际使用需求。更多还原

【Abstract】 Aiming at the problems that the traditionalCanny operator is susceptible to noise and poor edge connection when recognizing the weld seam edge of the straight seam welded pipe,an improved Canny operator is proposed to detect the weld seam of the straight welded pipe by analyzing the characteristics of the straight welded pipe image location method. First,the local mean deviation in the gradient direction was introduced in the non-maximum suppression;secondly,the threshold value was calculated by the Otsu method for the points after the non-maximum suppression;finally,the strong and weak edge distribution characteristics of the actual collection of welded pipe pictures are used as additional connection conditions. Experiments prove that the improved non-maxima suppression method can effectively suppress noise points and retain edge information compared with traditional methods.Compared with the traditional edge detection operator,the improved Canny operator in the welding seam edge position recognition of the straight seam welded pipe can suppress the noise interference and keep more welded pipe edge and weld edge,so it can better meet the actual use needs.更多还原