基于提升小波提取木材颜色特征的研究

Research of Wood Color Features Extraction Based on Lifting Wavelet

针对木材表面颜色自动分类的难题,在总结以往研究方法的基础上,提出一种新的颜色特征提取方法,即基于HSV颜色空间,运用提升小波变换来提取木材表面颜色信息,结合图像分块理论,最终形成了12个特征参数,然后运用BP神经网络,K-近邻和支持向量机对木材样本图像进行了分类仿真,最高的分类正确率达到了98.33%,实验结果验证了提出的颜色特征提取方法的有效性。

According to the problem of the wood surface color automatic classification, a new color feature extraction method is given in summary on the basis of the previous study. The method is based on HSV color model, uses lifting wavelet transform to extract wood surface color information, combines image block theory and forms 12 characteristic parameters eventually. Simulation experiments use the BP neural network, the K-nearest neighbor and the support vector machine （SVM） to classify the wood sample images. The highest classification accuracy is98.33%.The results of experiments validate the effectiveness of the new color feature extraction method.