• CN:11-2187/TH
  • ISSN:0577-6686

›› 2004, Vol. 40 ›› Issue (6): 179-182.

• 论文 • 上一篇    下一篇

基于视觉图像的机加工零部件亚像素边缘定位

张永宏;胡德金;徐俊杰   

  1. 上海交通大学机械与动力工程学院
  • 发布日期:2004-06-15

SUBPIXEL EDGE LOCATION OF MACHINE PARTS BASED ON THE VISION IMAGES

Zhang Yonghong;Hu Dejin;Xu Junjie   

  1. School of Mechanical and Power Engineering, Shanghai Jiaotong University
  • Published:2004-06-15

摘要: 首先介绍了LOG算子的快速定位方法以及二维理想图像边缘模型的空间矩亚像素细分算法的基本原理;在此基础上给出了符合实际图像采样的三级灰度图像边缘模型的空间矩算子,并推导了该算子的原理误差公式。最后利用试验对空间矩算子亚像素定位方法的有效性、精度进行了研究,并给出了该方法的亚像素定位与实测尺寸的对比结果。结果表明该方法比传统的边缘检测算子的边缘定位精度有较大提高。

关键词: LOG算子, 空间矩算子, 亚像素边缘定位

Abstract: The Laplacian of Gaussian operator rapid localization and the principle of subpixel edge operator based on two-dimensional spatial moment are presented at first. A three gray levels edge model based on spatial moment is given which is closed to practical image sampling. The error formula is deduced. Finally, the experiment is carried out. The validity and precision of subpixel edge spatial moment operator is studied and the subpixel edge detection results are compared to practical size. The results indicate that the subpixel edge detection operator based on spatial moment is much better than traditional edge detection operator.

Key words: Laplacian of gaussian operator, Spatial moment operator, Subpixel edge localization

中图分类号: