基于特征级数据融合木材纹理分类的研究

为了提高对木材纹理识别的精度，提出了一种基于融合灰度共生矩阵与高斯-马尔可夫随机场纹理参数的特征级数据融合木材纹理模式识别方法。首先，分别获取了以上两种木材纹理特征参数；然后，使用模拟退火算法将两种不同类型的纹理特征量在特征层上进行了融合。利用融合后的特征对木材纹理样本进行识别，BP神经网络分类器的识别率达到97.00%，表明数据融合后的特征参数对木材纹理识别是十分有效的。

Research on classification of wood texture based on feature level data fusion

In order to enhance the precision of wood texture recognition,a kind of wood surface texture recognition method is proposed,which uses Gray Level Co-occurence Matrix(GLCM) and Gaussian-Markov Random Field(GMRF) on feature level data fusion.First,two sorts of feature parameters above are extracted respectively.Next,two sorts of texture features are fused on the feature level by simulated annealing algorithm.With the fused features,the recognition rate of the integrated BP neural network to the wood textural ...