中文摘要：

基于图像图形学方法，实现了对木材纹理灰度特性的分析，并利用计算机自动检测出木材纹理形状、角度、纹理周期长度、线宽度和间距等特征量值。试验结果证实，BWMORPH为最适于木材类中弱纹理的形状轮廓检测，并生成新的纹理骨骼线图像；对纹理骨骼线图像进行Radon变换后，可得到0～180°纹理线条在相应角度上投影变换域的积分值，从而绘制出纹理角度的二维曲线图，纹理曲线图所反映的木材纹理方向性规律与人们平常对木材纹理的印象相吻合；通过将纹理图像二值化后再横向扫描的方式，能够得到纹理的峰—谷周期图，从中能够准确计算出纹理的周期长度，对应于木材的生长轮宽度，并可进一步求出纹理的线宽度和纹理的间距，分别对应木材的早、晚材宽度。

英文摘要：

Assisted by digital image processing technique, it analyzed textural gray characters of wood, and detected such features of wood as textural shape, directionality, period run-length, textural width, and textural spacing. Results had proved that BWMORPH was best suitable function for detecting such weak textural shape as wood, and could generate a skeletal texture image. After Radon transform on the skeletal texture image, a two dimensional curve plot was drawn to represent textural directionality of wood. By scanning pixel value across binary textural region, it obtained apex vs vale periodic plot, and calculated textural run-length, textural.width, textural spacing which were respectively corresponding with growth-ring width, latewood width and earlywood width.