基于分形维数的皮革图像分类研究  

The Research of Leather Image Classification Based on Fractal Dimension

作　　者：朱凯[1] 金建刚[1] 

机构地区：[1]华北水利水电学院软件学院,郑州450011

出　　处：《河南师范大学学报（自然科学版）》2013年第1期159-162,共4页Journal of Henan Normal University(Natural Science Edition)

基　　金：河南省基础与前沿技术研究计划项目(092300410234)

摘　　要：为了对皮革图像进行分类,引入分形维数这一新特征,并通过毯子法计算皮革图像的分形维数,通过与实验所得的阈值进行比较实现牛皮和羊皮图像的分类.试验表明,该方法对牛皮和羊皮图像能进行有效的分类.Leather is typical natural texture image which has fine and complex structure. In order to classification leather image, the new feature called fractal dimension is used in this paper. Blanket dimension is used to calculate the fractal dimension of leather image. The fractal dimension is compared with threshold to discriminate cowhide image and sheepskin image. According to the result of experimentation, the algorithm can effectively discriminate cowhide and sheepskin.

关 键 词：皮革纹理 图像分类 分形维数 毯子维 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]