基于离散小波变换多种纹理特征提取的草细胞图像分割比较

摘要

利用图像分割技术可有效检测草细胞纤维素的含量.基于离散小波变换的图像纹理特征提取方法已广泛应用于图像分割领域,而基于离散小波变换后的不同纹理特征提取方法对于图像分割效果具有不同的影响.本文首先介绍离散小波变换基本概念.然后基于离散小波变换提出了不同的纹理特征提取方法.最后针对草细胞图像分割结果对所列举的纹理特征提取方法做简单比较,从中选取了最有效的纹理特征提取方法.%Image segmentation technique can effectively detect the cellulose content of the grass cell. Texture feature extraction methods based on The Discrete Wavelet Transform(DWT)have been widely used in the area of image segmentation. But different texture feature extraction methods based on DWT have different impact on the results of image segmentation. The basic concepts of DWT are first briefly reviewed. And the next different texture feature extraction methods based on DWT are put forward. Finally those texture feature extraction methods listed are compared briefly according to the results of grass cells image segmentation, and the most effective texture feature extraction method is selected.