基于BP神经网络木材纹理分类的研究

Research on the Classification of Wood Textures Based on the BP Neural Network

应用 BP 神经网络对常见的10种木材纹理进行了分类研究,获得了比较满意的效果。首先,应用灰度共生矩阵提取了木材的纹理特征参数;其次,在此特征参数体系下,应用 BP 神经网络对木材纹理进行了分类研究,识别率达89%。

This paper applied BP neural network to classify the 10 kinds of common wood textures. The results of the classification were reasonable. First, GLCM was used to extract the feature parameters of the wood textures. Then, the BP neural network was applied to classify the wood textures under the feature parameter system. The recognition rate was 89%.