结合小波变换和数学形态学的图像分割算法  

Image Segmentation Algorithm based on Wavelet Transform and Mathematical Morphology

作　　者：刘鑫[1] 王坤[1] 

机构地区：[1]中国民航大学航空自动化学院,天津300300

出　　处：《科技与创新》2015年第16期6-8,共3页Science and Technology & Innovation

基　　金：中国民航大学校级项目基金(资助号:08QD17X;2010ky E07;2010QD02S);国家自然科学基金(编号:U1333111);天津市自然基金(编号:12JCQNJC00600);中国民航机务维修科研基地资助

摘　　要：飞机蒙皮孔图像的形成、传输或变换过程中会受到多种因素的干扰,为了精确地提取蒙皮边缘,采用了一种结合小波变换和数学形态学的边缘检测算法。在小波域中,对低频子图像采用改进的数学形态学算法进行边缘检测,使用不同类型的结构元素对现有的检测算子进行改进,然后对高频子图像采用小波模极大值法进行边缘检测,最后采用一定的融合规则对高低频边缘子图像进行融合。实验结果表明,与传统边缘算法相比,该方法用于图像边缘提取不但能有效去除噪声干扰,还能突出边缘细节,边缘定位连续准确。Aircraft skin pore image formation, transmission and transformation process will interfered by many factors, in order to accurately extract skin edge, using a combination of wavelet transform and mathematical morphology edge detection algorithm. In the wavelet domain, the low frequency sub image using the improved mathematical morphology algorithm for edge detection, to improve the existing detectors using different types of structural elements and of high frequency sub images by wavelet modulus maxima method of edge detection, finally, the fusion rules of high and low frequency edge sub image fusion. Experimental results show that the proposed method is effective in removing the noise and edge details, and the edge location is continuous and accurate, compared with the traditional edge detection method.

关 键 词：图像分析 小波变换 边缘检测 结构元素 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]