论文部分内容阅读
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.
A texture segmentation method based on wavelet and fuzzy theory is proposed. Firstly, the image is decomposed by higher-order wavelet to obtain a series of sub-images with different resolutions. Then the fuzzy clustering method is used to extract the texture from the lowest resolution image Clustering, the lower level of segmentation result is extended and then applied to a higher resolution, until the highest resolution, so get an original image of the initial segmentation; Finally, the introduction of space constraint algorithm, the thickness of the original image Split results. Since the correlation between image pixels is considered, the accuracy of the segmentation is improved. Simulation results show that this method is more effective.