复杂背景下钢索图像的纹理分割与边界识别

Texture Segmentation and Boundary Recognition of Wire Rope Images in Complicated Background

针对复杂背景下钢索图像难以准确分割的问题,提出一种基于纹理分析的钢索图像分割与边界识别方法.采用基于模糊Hough变换的纹理方向检测方法确定钢索走向,利用边缘方向密度直方图作为纹理特征,对与钢索纹理方向相应的边缘方向赋予不同权重,抑制纹理分割中背景的干扰,对钢丝绳图像进行聚类分割,采用检测平行直线的方法确定其边界,并根据算法参量对边界进行修正.在实验中,对比了边缘方向密度直方图特征与灰度共生矩阵、局部二值模式在钢索图像纹理分割中的结果与计算时间,结果表明边缘方向密度直方图特征计算速度快、受背景干扰小,分割准确率高.本文方法无须预先训练,受背景干扰小,可以准确地识别出钢索并确定其边界,能满足钢丝绳视觉检测的要求.

In order to detect the boundary of wire rope exactly in visual inspection of surface defection, a novel texture segmentation and boundary recognition method is proposed. The fuzzy Hough transform is used to detect the direction of rope texture. A new texture feature is proposed based on the Edge Direction Density Histogram （EDDH）. The EDDH is weighted based on the direction of rope texture in Fuzzy C-Means clustering segmentation. The boundary of wire rope is recognized preliminary to detect parallel lines in the map of membership by Hough transform. Moreover, an adaptive boundary correction method is presented to confirm the boundary exactly. The results of texture segmentation and boundary recognition by the proposed method are compared with the results of Gray Level Co-occurrence Matrix （GLCM） and Local Binary Pattern （LBP） respectively. The experimental results show that the proposed method can realize wire rope image segmentation and boundary recognition effectively, and its performance is better than the GLCM′s and LBP’s.