基于直方图和颜色矩方法的木材表面颜色特征的表达

Expression on the Basis of the Timber Surface Color Characteristic of the Histogram and Color Moment Method

简述了彩色图像处理中颜色空间的选择,分别用直方图和颜色矩方法表达木材表面的颜色特征。从H分量直方图可知,H分量在色彩上有较好的分类性,反映出木材彩色特征的变化。用颜色矩特征值作为BP神经网络的输入,对东北常见树种按颜色进行了分类,分级正确率达到了96.7%。

This paper discussed the choice of the color space in multicolored image processing, and expressed the color feature of wood surface with the histogram and color moment method. The better classfication on color was got by analyzing H weight histogram, and H weight reflected the change of the colored characteristic of wood. Using the characteristic value of the color - square as the input of BP neural network, four kinds of common trees in the Northeast region have been classified according to the color, the correct rate of classification has reached 96.7 %.