为了提取含噪声印制电路板（PCB）光板缺陷图像边缘信息，提出了一种基于混合法的图像边缘检测方法.在分析类间最大距离法图像分割基本原理的基础上提出了一种改进的类间最大距离法（IMDBC）；设计了结合中值滤波、IMDBC、改进的数学形态学边缘检测算子与LOG算子进行PCB光板缺陷图像边缘检测的混合方法.用CCD及显微镜成像系统获取4幅PCB光板缺陷图像，结果表明：用本文方法提取出的图像边缘信息清晰且较精确，噪声点较少，所得到的4幅图像优质系数是其它6种方法的1.0111~1.3586倍.

To extract the edge information of noisy Printed Circuit Board (PCB) defect image, an image edge detection method based on mixed method was proposed. The basic principle and advantages of the image segmentation method based on the Maximum Distance Between Categories (MDBC) were analyzed, and its improved segmentation method (IMDBC) was proposed. The image edge detection method based on the mixed method by combining with median filter, IMDBC, the improved mathematical morphology edge detection operator and LOG operator was discussed. The edge information extraction experiment by using four PCB defect images acquired by CCD and microscope imaging system was carried out. The experiment results show that it can extract better the PCB defect image edge information by using this method, and it has lesser noise points. In this method, the image excellent quality coefficients are from 1.0111 to 1.3586 times of other six methods mentioned in this paper.