Abstract：Aiming at the low accuracy in localization and sensitivity to noise in traditional edge detection algorithm, an algorithm of sub-pixel edge detection based on Gauss integral curve fitting is proposed. The gray value of discrete points on pixel edge normal section line is calculated with surface interpolation, and fitting Gauss integral curve, achieving accurate location of sub-pixel edge by searching the mean point of Gauss integral curve. Experiment with the gauge block line edge, and comparing with the existed sub-pixel edge detection algorithm, the experiment show that algorithm of sub-pixel edge detection based on Gauss integral curve fitting has high location accuracy, and can reach 1 μm, it also has high reliability and speed, and can be used for high precision measurement.
［1］潘兵,续伯钦,陈丁,等.数字图像相关中亚像素位移测量的曲面拟合法［J］.计量学报,2005,26(2):128-134.
［2］Zhao P, Ni G Q, Pu Z B. Simultaneous perimeter measurement for multiple planar objects［J］. Optics and Laser Technology, 2009, 41(2): 186-192.
［3］刘国栋,刘炳国,陈凤东,等.亚像素定位算法精度评价方法的研究［J］.光学学报,2009,29(12):3446-3451.
［4］李帅,卢荣胜,史艳琼,等.基于高斯曲面拟合的亚像素边缘检测算法［J］.工具技术,2011,45(7):79-82.
［5］段振云,张思佳,赵文辉.基于机器视觉的齿轮齿廓偏差测量［J］.计量技术,2014,(6):7-9.
［6］马睿,曾理,卢艳平.改进的基于Facet模型的亚像素边缘检测［J］.应用基础与工程科学学报,2009,17(2):296-302.
［7］王凯,张定华,黄鹤龄,等.基于3–D Facet模型的亚体素边缘检测算法研究［J］.机械科学与技术,2005,24(7):865-868.
［8］Carsten Steger, Markus Ulrich, Christian Wiedemann. Machine vision algorithms and application ［M］. Beijing: Tsinghua University Press, 2008: 204-228.
［9］于起峰,尚洋.摄像测量学原理与应用研究［M］.北京:科学出版社,2009.
［10］贺忠海,王宝光.利用曲线拟合方法的亚像素提取算法［J］.仪器仪表学报,2003,24(2):195-197.
2016, Vol. 37 
