基于颜色特征对木材分级的研究

给出几种基于颜色特征木材的分级算法,这几种方法通过主颜色特征,分别利用神经网络中的RBF网络、K-近邻和最近邻对木材样本图像进行处理,然后进行分类.木材样本图像是通过对木材加工处理采集而得到的;主颜色特征是反映一幅图像的基本面貌,并通过HSV颜色模型、量化、颜色直方图和绝对距离而得到的.选取3种主颜色方案数据,经过Matlab7.0平台进行程序设计,对木材样本图像数据进行训练和测试,并加以验证.实验结果表明,这几种方法能够较好地解决木材分级的问题.

Wood graduation research based on color characteristic

The several kinds methods based on the color characteristic wood graduation algorithm are given. These methods through the main color characteristic, use in the neural network separately the RBF network, the K-nearest neighbor and the nearest neighbor carry on processing to the wood sample image, then carries on classified the method. The wood sample image is obtained by woodworking processing gathering. The main color characteristic is reflected an image the basic appearance, and through the HSV color model, the quantification, the color histogram and is away from absolutely obtains. The three kind of main color plan data is selected, the progra- mming after Matlab7.0 platform is carried on, the training and the test to the wood sample image data are made, and confirmed. The experimental result indicated that, these methods could solve the wood graduation problem well.