摘要
为提高多肉植物分类效率和识别精度,融合多肉植物颜色特征和纹理特征,提出一种基于多肉植物图像复合特征的多肉植物分类算法。将颜色特征和纹理特征组成的复合特征作为WPA-SVM的输入,多肉植物类别作为WPA-SVM的输出,建立WPA-SVM多肉植物分类识别模型。与SVM、ELM和BPNN对比发现,研究结果表明,WPA-SVM可以有效提高多肉植物分类识别的精度,为多肉植物识别研究和应用提供了新的方法和途径。
In order to improve the classification efficiency and recognition precision of succulent plants,a succulent plant classification algorithm based on the composite features of succulent plant images is proposed.The composite characteristics of color characteristics and texture features are used as input of WPA-SVM,and the succulent plant category is used as output of WPA-SVM to establish a WPA-SVM succulent plant classification recognition model.Compared with SVM,ELM,and BPNN,the research results show that WPA-SVM can effectively improve the accuracy of succulent plant classification and identification,and provide new methods and approaches for succulent plant identification and application.
作者
王守富
孟昭磊
何利华
杨杰锋
WANG Shoufu;MENG Zhaolei;HE Lihua;YANG Jiefeng(Landscape Architecture College, Hubei Ecological Engineering Vocational and Technical College, Wuhan, Hubei 430200, China)
出处
《微型电脑应用》
2020年第6期29-32,36,共5页
Microcomputer Applications
基金
湖北省教育厅科学技术研究项目(B2017551)。
关键词
狼群算法
支持向量机
极限学习机
神经网络
多肉植物
wolf pack algorithm
support vector machine
extreme learning machine
neural networks
succulent plants
