摘要
光学元件表面缺陷检测依赖高精密仪器配合人工检测,存在检测精度低的难题,为此提出基于数字图像特征分割和角点检测的光学元件表面缺陷检测算法。首先采用激光扫描方法采集光学元件表面图像,引入最小二乘匹配滤波方法降噪,然后通过纹理尺度自适应分割方法进行表面缺陷的边缘轮廓特征分割,结合Harris角点检测算法实现对表面缺陷检测。结果表明,提出的算法对光学元件表面缺陷检测轮廓分割和标识准确性高,准确检测概率高于传统算法。
Accurate detection of surface defect optical element has relied on high precision instruments with manual testing, in order to solve the problem of low detection accuracy, put forward a optical element surface defect detection algorithm based on characteristics of digital image segmentation and comer detection. Firstly, laser scanning method is adopted to improve optical element surface image capture which introducing least square matching filter method for noise reduction processing, texture segmentation method for edge contour features of surface defect segmentation, combined with Harris comer detection algorithm of the optical element surface defect detection. Test results show that the algorithm is adopted to improve the optical element surface defect detection, to the edge of the defective parts contour segmentation and identification accuracy is high, accurate angular point feature extraction with good performance, accurate detection probability is higher than the traditional algorithm.
出处
《激光杂志》
北大核心
2017年第1期47-50,共4页
Laser Journal
基金
新疆自治区高校科研计划项目(XJEDU2011I49)
关键词
光学元件
缺陷检测
图像降噪
边缘轮廓
optical components
defects
detection
image denoising
edge profile