一种改进的灰度矩亚像素边缘检测算法

An improved algorithm for subpixel edge detection using gray moment

在分析Tabatabai提出的灰度矩亚像素边缘检测算法的基础上，指出灰度矩算法存在边缘判断条件不够完善和未能考虑模板效应的问题，提出了改进方法，考虑Tabatabai的灰度矩算法产生很多虚假边缘，改进算法分析了各参数对结果的影响，对边缘判断条件进行完善。实验结果表明，所改进算法具有抗干扰性、边缘细化能力强，定位准确的特点，分辨精度可达0．06～0．08个像素。

Subpixel edge detection is an important research issues in image measurement. By analyzing the algorithm for subpixel edge detection presented by Tabatabai, two shortcomings of gray moment algorithm were revealed： difficulties in judging edges, and no consideration of the model effect. An improved algorithm was presented aiming at remedying the shortcomings. Considering there were many false edges in the image processed by Tabatabai＇s algorithm, we improved the edge judging rules of the algorithm by analyzing the influence of parameters on the result. The experimental results show that the improved algorithm has good anti-jamming and edge thinning abilities, and high locating accuracy with an error of 0.06 to 0.08 pixels.