基于自适应神经模糊推理系统的树叶特征的提取以及分类  

The Feature Extraction and Classification of Leaves Based ANFIS

作　　者：李永亮[1] 向长城[1] 刘国勋[1] 

机构地区：[1]湖北民族学院理学院,湖北恩施45000

出　　处：《湖北民族学院学报（自然科学版）》2015年第4期388-392,420,共6页Journal of Hubei Minzu University(Natural Science Edition)

基　　金：湖北省自然科学基金项目(2015 CFB264)

摘　　要：通过植物的几何特征:长宽比,矩形度,面积凹凸比再根据傅里叶描述子和得到的傅里叶系数来提取树叶的一些形状特征:圆形度、细长度、密集度、凹凸度,再利用广泛应用于处理与分析具有复杂细节特征的分形维数来对植物特征进行提取,最后采用自适应神经模糊推理系统(Adaptive Network-based Fuzzy Inference System,ANFIS)基于sugeno模糊来对树叶进行分类,并对树叶的分类进行分析,结果显示:随着适应神经网络的训练次数的增加,识别的误差越来越低,最后达到了比较好的效果,能够准确的进行分类.Research on feature extraction and classification of the leaves is a hot field. The difficulty of this kind of problems is that according to the image information leaves extracts a set of features, and then make the corresponding classification methods. First, for feature extraction, this paper adopts the corre?sponding feature geometry,Fourier descriptors and the fractal dimension of the feature description leaves. Finally the leaves are classitied by adopting ANFIS based sugeno fuzziness. Classification recognition rate is higher and can be applied to accurate classification.

关 键 词：几何特征 傅里叶描述子 分形维数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]