摘要
提出了基于经典Hough变换的部分矿物薄片图像线特征检测方法.首先使用Laplace算子得到图像的特征边缘,然后用Otsu对梯度图像进行二值化,接着去除小连通域的噪声,采用与目标大小相关的自适应变换参数对图像进行Hough变换得到直线特征.实验表明该方法可以检测出部分矿物的线特征.
A method for line feature detection of some minerals and rocks slice image is introduced,which is based on classic Hough transform（HT）.The edge feature of image is obtained by using Laplace operator measurement.Then the method applies Otsu Binarization algorithm to gain the edge image,and denoising algorithm to remove small connected domain.At last,on the basis of it,the straight line feature of image was gotten by using Hough Transform method with adaptive transform parameters associated with the size of target.The proposed method can extract line feature of part of minerals and rocks identified by experiment.
出处
《微电子学与计算机》
CSCD
北大核心
2015年第8期120-124,共5页
Microelectronics & Computer
基金
国家自然科学基金项目(61372174)
