中文摘要：

通过利用实数9/7小波变换,对木材纹理进行多尺度分解,利用特征参数分析了木材纹理频率分布特征,比较了针叶材与阔叶材、径向与弦向切面的木材纹理的差异。实验结果表明,利用实数9/7小波变换,分析各子图像能量值的标准差,筛选出木材纹理最佳分解尺度为3,同时还发现小波能量分布比例和EHL/ELH值可作为木材纹理方向性参数。

英文摘要：

Based on the 9/7 wavelet transforms, multi-scale decomposition was conducted on the the wood texture. The frequency distribution features of wood texture were analyzed by using characteristic coefficients. Wood texture differences between coniferous and broad-leaved woods, radial section and tangential section of woods were analyzed. From the results of analyzing the standard deviations of sub-image energies by the 9/7 wavelet transforms, it was found the the optimal decomposition scale for wood texture was 3, and the distribution ratio of subsection energy and EHL/ELH could be used to declare texture direction of wood.