摘要
提出了一种基于二维经验模式分解(BEMD)的纹理图像分类改进方法。采用BEMD算法将纹理图像分解为两层二维固有模态函数(IMF)和一个余量之和,结合灰度共生矩阵(GL-CM)对这两层IMF各提取5个纹理特征参数,组成一个扩展的10维特征向量,然后根据扩展的特征向量,采用最小距离分类器(MDC)进行纹理图像分类。仿真结果证明了该方法进行纹理图像分类的有效性。
An improved texture image classification method based on bidimensional empirical mode decomposition is proposed. Firstly, the texture image is deco to two bidimensional Intrinsic Mode Function（IMF） and one residue by using BEMD algorithm, and then five texture feature parameters are extracted from each of the two IMF combined with Gray Level Co-occurrence Matrix, by which an extended ten dimensional feature vector is formed. Finally, texture image classification is done using minimum distance classifier according to the extended feature vector. Simulation results verified the effectiveness of the method for texture image classification.
出处
《黑龙江大学自然科学学报》
CAS
北大核心
2012年第5期680-685,690,共7页
Journal of Natural Science of Heilongjiang University
基金
国家自然科学基金资助项目(61077079
60802059)
高等学校博士学科点专项基金资助项目(20102304110013)
哈尔滨市优秀学科带头人基金资助项目(2009RFXXG034)
