基于图像的线轮廓度评价算法研究

根据微型零件的成像特点，为了实现其轮廓的高精度测量，提出了一种基于离散点的轮廓度评价算法——被测轮廓与理论轮廓离散点间的最小距离法：首先提取出被测轮廓的边缘点信息，其次依据理论轮廓计算出一系列间距极小的坐标点并建立坐标系将被测轮廓点与理论轮廓点对应，最后计算出每个被测轮廓点到最近理论轮廓点的距离作为该点的轮廓度误差。实验结果证明，精度优于1.5个像素，此方法可有效的提高了线轮廓度的评价精度和效率。

Research on the Evaluation Algorithm of Curve Profile Based on the Image

For achieving high precision measurement of linear profile for micro-parts,an evaluation algorithm of linear profile based on discrete points is proposed according to the imaging characteristics.Firstly,an edge detector is used to obtain the edge of the micro-parts.Then,a series of points with tiny interval can be calculated by the expressions of designed profile and a system coordinate is build for corresponding the points of actual profile and the points of designed profile.Finally,the minimum distance between every point of actual profile and the point of designed profile is taken as the linear profile error of the point.The experimental results show that the error is less than 2 pixel,so the proposed algorithm can significantly improve the precision and efficiency in the measurement of linear profile.