基于形态学的机器视觉玻璃切割边缘提取  

Edge Detection of Cutting-Glass Using Machine Vision Based on Morphology

作　　者：王蔚 王晓凯[1] 龚真 刘慧锋 WANG Wei;WANG Xiaokai;GONG Zhen;LIU Huifeng(College of Physics and Electronic Engineering,Shanxi University,Taiyuan 030006,China;Shanxi Teamwork Photoelectric Industries Co.Ltd.,Taiyuan 030006,China)

机构地区：[1]山西大学物理电子工程学院,山西太原030006 [2]山西迪迈沃科光电工业有限公司,山西太原030006

出　　处：《测试技术学报》2020年第1期22-27,共6页Journal of Test and Measurement Technology

基　　金：山西省重点研发计划（高新技术领域）资助项目（201803D121102）

摘　　要：图像边缘提取是基于机器视觉的光学玻璃精密切割的前提和基础.通过Sobel、Roberts、Prewitt、LOG、Canny传统算法分别对光学玻璃切割中的图像进行边缘提取,发现这些算法均无法有效解决边缘线型连续性和去噪问题.在此基础上,引入了数学形态学理论,对图像依次进行闭合运算、Otsu阈值处理、轮廓提取,从而获得了连续、平滑、无噪点的光学玻璃切割轮廓曲线图,且峰值信噪比相对较高,仿真实验证明了该方法的有效性.The image edge detection is the premise and foundation of the optical glass cutting based on machine vision.The image edges were detected by five algorithms,such as Sobel,Roberts,Prewitt,LOG and Canny operators.However,these traditional algorithms could not effectively solve the problems of edge line continuity and denoising.Therefore,mathematical morphology theory was introduced,and then the image was successively calculated as follows,including closure operation,Otsu threshold processing and contour extraction,which can obtain continuous,smooth and noise-free curves of cutting contour for the optical glass and PSNR is relatively higher than that of before.The simulation results proves the validity of the proposed method.

关 键 词：机器视觉 图像边缘提取 形态学 光学玻璃切割 

分 类 号：X832[环境科学与工程—环境工程]