SIFT算法在木材纹理分类上的应用  

Application of SIFT to Wood Texture Classification

作　　者：陈宇[1] 臧美英 李红波[1] 

机构地区：[1]东北林业大学信息与计算机工程学院,黑龙江哈尔滨150040

出　　处：《哈尔滨理工大学学报》2016年第4期7-12,共6页Journal of Harbin University of Science and Technology

基　　金：黑龙江省自然科学基金(F201347);中央高校基本科研业务费专项资金项目(DL12CB02);黑龙江省教育厅科学技术研究项目(12513016);黑龙江省博士后基金;国家948项目(2011-4-04);哈尔滨市科技创新人才专项资金项目(2013RFQXJ100)

摘　　要：为解决地板块纹理分类难度大的问题,提出了一种基于SIFT(scale-invariant feature transform),尺度不变特征转换的地板块纹理分类方法,该方法首先采用SIFT算法提取地板块图像特征值,并采用K-means聚类算法降低关键点数目,得到用于分类的特征行向量,最后构造DEELM分类器进行分类.实验结果表明,该方法不仅减少了描述图片的关键点数目便于分类,而且提高了木材纹理分类的准确性,为地板块纹理分类的研究提供了一个新的思路.In order to solve the problems of the classification of wood floor by the texture, a SIFT model on wood texture classification is proposed. This method uses the sift algorithm to extract ground plate image characteris- tic value firstly and then uses the k-means clustering algorithm to decrease the number of key points, obtaining the feature Row vector that used in the classification, and finally, constructing DEELM classifier to classify. The experi- mental results show that this method not only reduces the number of key points which describe the pictures, but also improves the precision of wood texture classification, providing a new thought for the research into the classification of wood floor＇ s texture.

关 键 词：木材纹理分类 SIFT算法 K—means算法 差分演化优化极限学习机 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]