摘要
分析了非相干光照明条件下微小型零件的边缘光学特征,传统的边缘检测方法没有考虑微小型零件的边缘光学特征,不能满足显微视觉测量方法的精度要求。针对微小型零件边缘检测的问题,提出一种基于最小二乘拟合的亚像素边缘检测方法。首先找出边缘区域灰度值最大和最小的点,根据灰度值最大和最小的点计算出阈值,然后将灰度值最大和最小点之间的点用直线方程回归,根据阈值的大小求出微小型零件的尺寸。试验结果表明,这种新的方法能够准确地识别出微小型零件的边缘,满足显微视觉测量精度的要求。
The optical feature of the micro accessory under incoherent beams is analyzed. The optical feature of the micro accessory can＇ t meet the measure precision of the microscope measurement system, for it is not considered by using traditional image edge detecting method. With the view of the difficulty, an sub pixel edge detection method based on least square fit detecting is proposed. Firstly, the maximum and the minimum are found in the edge region, and the threshold are detected. The second step is to disposed the pixel with linear regression method. Then dimension of the micro accessory is calculated by the threshold. Experiment results show that the new method can detect the micro accessory edges exactly and can meet the measure precision of the microscope measurement system.
出处
《制造业自动化》
北大核心
2007年第2期27-28,74,共3页
Manufacturing Automation
基金
基础科研项目(20050329014)
关键词
微小型零件
边缘检测
亚像素
阈值
micro accessory
edge detection
sub pixel
threshold
