一种彩色PCB图像的边缘检测算法研究  

Research on Edge Detection Algorithm of Color PCB Image

作　　者：杜颜颜[1] 杨帆[1] 王晓颖[1] 

机构地区：[1]河北工业大学信息工程学院,天津300401

出　　处：《电视技术》2011年第13期113-115,共3页Video Engineering

摘　　要：在以图像处理技术为核心的PCB自动光学检测系统中,图像的边缘检测占有重要的地位,并且彩色图像提供的边缘信息更丰富。提出了一种检测彩色PCB图像边缘的方法。首先将彩色图像由RGB空间转换到HSI空间,然后应用基于中心样条滤波和4个模板方向的Sobel算子计算梯度幅值的改进Canny算子,分别对HSI的三分量检测其边缘,并将其结果融合成最终的图像边缘。实验结果表明,算法能成功检测出彩色PCB图像的边缘,提供更多的边缘细节,具有边缘定位精度高、去噪效果好的优点。The edge detection is important in the PCB＇s auto optical inspection system based on image processing, and color image can supply more useful edge information, so a color PCB image edge detection method is proposed. The color image is transformed from the RGB space to the HSV space, then the improved Canny algorithm based on B-spline filtering and Sobel operator with 4 template direction calculating gradient amplitude is applied to detect its edge of the three components of HSI separately, finally, the image edge is obtained by fusing the three results above. The results show that the algorithm successfully detect the edge of color PCB image, more edge details are obtained, the paper＇s advantage is having higher precision and wipping off yawp is more effectively.

关 键 词：彩色PCB图像 边缘检测 改进Canny HSI空间 B样条函数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]