基于多特征融合的药用植物标本识别

来源期刊:中南大学学报(自然科学版)2011年第z1期

论文作者:叶锋 蔡光东 郑子华 亓晓旭 尹鹏

文章页码:656 - 660

关键词:标本识别;多特征融合;边缘方向直方图;LBP;SIFT

Key words:specimen recognition; multi-features fusion; edge orientation histogram; LBP; SIFT

摘    要:提出一种基于多特征融合的药用植物标本识别的新方法。首先,从标本图片的底层纹理特征入手,研究标本图像的边缘方向直方图、LBP以及SIFT特征;然后提出基于准确率类别加权的融合模型以确定出最好的融合策略。最后,比较了多特征融合方法和单一特征方法的优劣。通过对21种植物1 470份药用植物标本的测试,证实了本文提出方法的有效性。

Abstract: A new method for medicinal plants specimen recognition was presented based on multi-feature fusion. Firstly, the bottom feature of the specimen image, such as edge orientation histogram, LBP and SIFT features was studied and extracted. Then, the accuracy-class-weighted algorithm was proposed to determine the best integration strategy. Furthermore, the comparison between the multi-feature fusion methods and a single feature method was done. Finally, the performance of this system was evaluated, which consisted of 21 kinds of plants including 1 470 specimens of medicinal plants. The results confirm the effectiveness of the proposed method.

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号