摘要
将高斯—马尔可夫随机场(GMRF)引入木材纹理的研究,提取了二阶与五阶特征参数,并对二阶特征参数做了详细分析,得出通过θ2可以判断纹理的主方向,而结合θ1、θ2、θ3、θ4能够区分开木材的弦切和径切纹理。将五阶特征参数组成的特征向量输入给BP神经网络分类器,其分类识别率约为85%,表明了高阶GMRF参数对木材纹理描述的有效性。
This paper introduced GMRF into wood texture research, extracted 2-rank and 5-rank feature parameter, and analysed them in detail, So the conclusion followed was： θ2 could judge the main direction; radial and tangential texture could be distinguished by θ1.θ2.θ3.θ4. 5-rank feature parameters were put into BP neural network classifier, and the classification and recognition ratio basically reached to 85,0%, which indicated that wood texture could be described effectively by GMRF parameters.
出处
《林业机械与木工设备》
2006年第11期25-27,共3页
Forestry Machinery & Woodworking Equipment
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(2004AFXXJ020)
