Abstract: The image classification algorithm of the microscopic cell of wood based on the texture feature is proposed.The modulus extreme density of the Nonsubsampled Contourlet transform is used to extract the f0eature of cell image classification.The K-nearest neighbor classification method is used to carry out the classification of timber microscopic cell image.Experimental results indicate that the average recognition accuracy is above 85%.The proposed method can effectively achieve the classification of timber microscopic cell image.

Key words: texture feature, classification, nonsubsampled Contourlet transform, wood cell

摘要： 提出一种基于纹理的木材显微细胞图像分类算法。通过非下采样的Contourlet变换模极值密度提取图像纹理特征，并采用K近邻分类方法进行分类，实现对木材显微细胞图像的分类。实验结果表明：平均识别正确率在85%以上。提出的方法能有效地实现对木材显微细胞图像的分类。

关键词: 纹理特征, 分类, 非下采样Contourlet变换, 木材细胞