改进的基于单一尺度的医学图像边缘检测  

Improved Single Scale Edge Detection for Medical Image

作　　者：张萌萌[1,2] 杨扬[1] 杨志辉[2] 李夏[2] 白慧慧[3] 

机构地区：[1]北京科技大学信息工程学院,北京100083 [2]北方工业大学信息工程学院,北京100144 [3]北京交通大学信息科学研究所,北京100044

出　　处：《太原理工大学学报》2011年第4期329-333,共5页Journal of Taiyuan University of Technology

基　　金：国家自然科学基金(60903066;60972085);北京市自然科学基金(4102049);教育部新教师基金(20090009120006);北京市属高等学校人才强教深化计划基金(PHR201008187)

摘　　要：针对经典取边缘算法的缺点和高斯多尺度边缘检测中尺度选择的复杂性等问题,提出了改进的单一尺度边缘检测方法,并将该方法应用到医学图像的边缘检测中。该方法首先用平滑理论,对图像进行平滑,将图像中一些无用的细节信息平滑掉,抑制噪声和高频干扰成分;因为边缘细节也被平滑掉,所以再利用模糊增强算子加大边缘两侧灰度的差异,然后利用基于高斯核的单一尺度过零点边缘检测方法提取图像的边缘;最后,将该算法与经典的sobel,canny算子进行比较。实验结果表明,这种方法较好解决了图像边缘的提取精度和图像噪声的抑制能力之间的矛盾。Aiming at the shortcoming of the classical edge algorithm and the complexity of the Gaussian multi-scale edge detection, this paper proposed an improved single scale edge detection method applied it to detect the edges of medical images. The method uses the smoothing theory to smooth the image so as to smooth out some useless details and suppress noise and high-frequency interference components. However, it also smooth out some details of the edge. So the fuzzy en- hancement operator was added to enhance the difference between both sides of the edge, and then the zero-crossing edge detection based on a single scale is adopted to extract the image edge. And finally, this algorithm was compared with the classical sobel and canny operators. The experimental results show that this method solved the contradiction between the accuracy of image edge extraction and the suppression of image noise.

关 键 词：模糊增强 图像平滑 过零点 边缘检测 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]