不同对比度纹理结构特征的多粒度分解及描述  

Multi-granularity Decomposition and Description of Texture Features with Different Contrast

作　　者：冯霞 周勃[2,3] 崔卫红 

机构地区：[1]湖北省标准化与质量研究院,湖北武汉430061 [2]武汉大学,湖北武汉430079 [3]武汉市规划研究院,湖北武汉430014

出　　处：《计算机技术与发展》2017年第12期85-88,共4页Computer Technology and Development

基　　金：国家"973"重点基础研究发展计划项目(2012CB719903)

摘　　要：纹理是一种普遍存在的视觉现象,在图像处理、三维重建和模式识别中都具有非常重要的作用。由于纹理具有多粒度特性,基于单一尺度的结构基元已经不足以准确地描述纹理特征,而在数学形态学运算中,利用不同大小的结构元素迭代进行形态学运算,可以获得不同粒度下的结构信息。基于上述考虑,给出了一种基于形态学开运算的纹理结构的多粒度分解形式,并提出了一种整体对比度不同而结构相同的纹理结构特征的多粒度描述方法。为了验证该方法的有效性,以Brodatz数据库和MIT Vis Tex数据库的20幅纹理图像为实验数据,将提出的多粒度描述方法与基于灰度共生矩阵的纹理描述方法和基于Gabor滤波的纹理描述方法进行比较。实验结果表明,该方法优于上述两种方法。Texture is a kind of common visual phenomena, which plays a very important role in image processing,3D reconstruction and pattern recognition. Due to the multi-granularity characteristic of texture,structural primitives based on single size are not enough to de- scribe texture features accurately. However, based on different size structural elements for iterative morphological operations in the mathe- matical morphology, the structure information with different granularity can be obtained. On this basis, a multi-granularity decomposition of the texture structure based on morphological opening operation is put forward, and a multi-granularity description method of texture features with different overall contrasts and the same structure is proposed. In order to verify its effectiveness, with 20 texture images from Brodatz database and MIT VisTex database as experimental data, the proposed method is made a comparison with texture description method based on gray level co-occun^nce matrix and based on Gabor filtering. The experiment shows that it is better than the other two methods.

关 键 词：纹理结构特征 数学形态学 开运算 多粒度分解 多粒度描述 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]