基于Canny-Harris角点检测的光学元件表面疵病检测
Surface Defect Detection of Optical Components Based on Canny-Harris Corner Detection
王赛赛 1周阿维1
作者信息
- 1. 西安工程大学机电工程学院,陕西西安 710600
- 折叠
摘要
为提高光学元件表面疵病检测的准确性,提出了一种基于Canny-Harris角点检测的光学元件表面疵病检测方法.该方法首先利用基于迭代式阈值的Canny算子对图像进行边缘检测,得到候选角点;接着,利用基于局部窗口角点密度的可变窗口非极大值抑制方法去除候选角点中的伪角点,以得到检测精度更高的Canny-Harris角点检测算法;然后,利用Canny-Harris角点检测算法进行图像拼接,并利用边界测定法提取出拼接图像中的疵病特征;最后,通过计算疵病的形状因子,实现对疵病的分类.实验结果表明,该方法能够成功实现子孔径图像的精确拼接,并准确地检测出光学元件表面疵病信息.
Abstract
To improve the accuracy of surface defect detection of optical components,a method for de-tecting surface defects of optical components based on Canny-Harris corner point detection is proposed.Firstly,the Canny operator based on iterative threshold is used to detect the edge of the image to obtain candidate corner points,so as to reduce the calculation amount of the algorithm;then,the variable window non-maximum suppression method based on local window corner point density is used to remove the pseudo corner points in the candidate corner points to obtain the correct corner points;then,the detected corner points are used for image stitching,and the defective features in the stitched images are extracted by boundary determination method;finally,defect classification is achieved through the calculation of the de-fect's shape factor.Experimental results show that this method can successfully achieve the accurate stitc-hing of sub-aperture images and accurately detect the surface defect information of optical components.
关键词
光学元件/疵病检测/Canny-Harris算法/图像拼接/疵病识别Key words
optical component/defect detection/Canny-Harris algorithm/image stitching/defect recog-nition引用本文复制引用
基金项目
陕西省自然科学基础研究计划资助项目(2018JQ5180)
出版年
2024
国家科技期刊平台
NSTL
万方数据