基于小波和模糊理论的纹理分割方法

Texture Segmentation Method Based on Wavelet and Fuzzy Theory

提出了一种基于小波和模糊理论的纹理分割方法，该方法首先对图象进行高阶小波分解，得到一系列分辨率不同的子图象；然后采用模糊聚类方法从最低分辨率图象进行聚类，将低一级的分割结果扩展后再应用于较高级分辨率，一直到最高分辨率为止，这样就得到一个原始图象的初始分割；最后引入空间约束算法，得到原始图象的粗细分割结果．由于考虑了图象象素之间的相互关系，故提高了分割的准确性．仿真结果表明该方法是比较有效的．

An approach based on wavelet and fuzzy is proposed. (1) It processes pyramid hierarchical wavelet for an image, the result is a set of images which have different sizes, then the local feature is extracted by using different windows on different images. (2) From the lowest resolution (the smallest size), it processes clustering based on fuzzy clustering algorithm. The result of the low resolution is expanded and used in the high resolution, and the process goes on until it reaches the highest; so a coarse segmentation of the original image is obtained. (3) It introduces spatial constraints and a label relaxation algorithm is used to correct every pixel's label of image. The result is a fine segmentation of the original image. Considering the relationship of image's neighboring pixels, this method improves the accuracy of segmentation. The proposed method is tested on a standard Bordatz image and proved as an effective method.