摘要
不变矩是模式识别中的一种重要方法,它具有平移不变性、比例不变性和旋转不变性等优点。本文将其引入到木材纹理的计算机视觉研究领域,提取了木材纹理的不变矩参数,并用提取的特征参数对木材纹理进行了分类研究,最近邻分类器的正确率为86.67%,获得了较高的分类正确率,从而验证了不变矩参数对木材纹理描述的有效性。
Invariant moments are important measure in the pattern recognition. Invariant moments are independent of position, scale and orientation. So this paper introduced it to wood texture computer vision research area, and extracted the invariant moments＇parameters of wood texture. The classification was carried on wood texture of samples based on invariant moments＇parameters, and the correct ratio of the nearest neighbor classifier reached 86.67% and confirmed that invariant moments＇parameters were effective to describe wood texture.
出处
《林业机械与木工设备》
2006年第6期29-30,共2页
Forestry Machinery & Woodworking Equipment
关键词
不变矩
木材
纹理
最近邻分类器
invariant moments
wood
texture
the nearest neighbor classifier
