摘要
为了解决木材的分级问题,提出了利用神经网络和木材表面颜色特征对木材进行分级的方法。利用竞争神经网络对114种不同树种进行粗分,然后再用BP神经网络算法对此进行验证,又对四种木材进行分类验证。从计算机分类分级仿真效果上看,达到了预期的目的。
In order to solve the problem of grading the timber, the authors have put forward a kind of classification method by utilizing neural network and wood surface color characteristics, The competitive neural network was used to divide thickly 114 kinds of different species of trees and then BP neural networks algorithm was applied to exemplify the resuits. And four kinds of timber were further classified. From the results of classifying and grading simulation on computer, the anticipated goal was achieved.
出处
《森林工程》
2006年第1期18-20,共3页
Forest Engineering
基金
黑龙江省自然科学基金项目(C2004-03)哈尔滨市自然科学基金项目(2004AFX XJ 0 20)
关键词
神经网络
BP神经网络算法
颜色
颜色矩
neural network
Back Propagation Algorithms
color
color matrix