介绍了一种光笔式便携三坐标视觉测量系统，提出了被测特征点像面坐标的提取方法。依据数字图像处理理论并结合测量系统中被测特征点的成像特点，给出了图像预处理过程及实际处理结果；采用基于灰度质心法的径向截面扫描法对图像边缘进行亚像素级检测；应用平面内任意位置椭圆的最小二乘曲线拟合法对椭圆形光斑中心的像面坐标进行了提取，实验结果表明该方法的提取精度为0.1个像素。

A light-pen portable 3D coordinate vision measurement system is introduced, and a method for picking up the image coordinates of the measured feature points is proposed. Based on the theory of digital image processing and combining with the image characteristics, the practical methods and results of image preprocessing are presented. Applying the radial section scanning method based on gray centroid, the image edges are tested with a subpixel precision. The image coordinates of ellipse spots centers are picked up using the Least Square fitting method of an ellipsis which is located arbitrarily in a plane, and experimental results show that the prediction of this method can reach 0.1pixel/subpixel.