基于亚像素特征点提取的螺纹检测粗糙聚类算法

本文研究了一种应用于高速图像检测的基于亚像素特征点提取的螺纹检测粗糙聚类算法。其基本思想是:采用分阶段的高精度亚像素特征点提取方法,将图像边缘特征离散为亚像素级特征点,利用粗糙集中的不可分辨概念和近似集合概念,对图像亚像素级特征点进行粗糙聚类,以便区分图像中多个螺纹零件,确定螺纹小径不可分辨类。在此基础上,给出了螺纹几何参数测量的步骤和计算规则,根据计算结果对螺纹零件进行基于图像特征的判别和处理。这种基于亚像素特征点提取的螺纹检测粗糙聚类方法具有较高的检测精度。

A Roughness Clustering Algorithm for Threads Detection Based on Sub-Pixel Feature Extraction

In order to improve accuracy and the speed of thread detection,a roughness clustering algorithm based on the sub-pixel feature points extracted in thread digital images is proposed. Firstly,the high accuracy sub-pixel feature points extracted by the stage method is adopted to disperse image edge features and to obtain sub-pixel feature points. Then,these sub-pixel feature points are aggregated by the roughness clustering algorithm so as to partition different thread parts in a thread image as well as the th...