用于大豆品种识别的叶片深度特征学习方法
作者:
作者单位:

1.南京财经大学信息工程学院;2.南京财经大学 信息工程学院;3.澳大利亚悉尼大学

作者简介:

通讯作者:

基金项目:

江苏省自然科学基金(BK20181414); 江苏省高校优秀科技创新团队项目(2017-15); 江苏省高校自然科学研究重大项目(18KJA52004); 智能机器人湖北省重点实验室开放基金(HBIR202001)


Leaf Deep Features Learning Method for Soybean Cultivar Recognition
Author:
Affiliation:

1.School of Information Engineering, Nanjing University of Finance &Economics;2.College of Information Engineering,Nanjing University of Finance and Economics;3.School of Biomedical Engineering, The University of Sydney

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
    摘要:

    大豆有许多品种(cultivar),它们的叶片图像模式的差异非常细微,因此很难通过叶片特征将大豆品种区分开。虽然在使用叶片图像模式进行植物种类(species)识别方面的研究已经取得了巨大的进步,然而,作为一项非常细粒度的模式识别问题,大豆品种的识别与分类研究尚未引起足够的重视。传统的手工叶片图像分析方法一般无法刻划不同大豆品种的叶片特征的细微差异,因此识别率很低。本文尝试使用深度学习来提取具有强的辨识能力的叶片特征,以解决大豆的品种识别问题。我们提出了一种新颖的深度学习模型,称为目标转换注意力网络(TAN)。该方法首先通过注意力机制提取细粒度的叶片图像特征,然后使用仿射变换纠正叶片姿势。我们构建了一个由240个大豆品种组成的大豆叶片品种图像数据集,每个品种有10个样本,以此数据集验证叶片图像模式中品种信息的可用性,并验证了所提出的深度学习模型对大豆品种识别的有效性。令人鼓舞的是实验结果证实了叶片图像模式在区分栽培大豆品种方面的有效性,并证明了所提出的方法优于流行的叶片手工特征提取方法和深度学习方法.

    Abstract:

    Soybeans include many varieties (cultivars) and their cultivars have very subtle differences in leaf patterns which makes it very tough to distinguish them from leaf features. Great progress has been made on using leaf image patterns for plant species recognition. However, as a general very fined-grained pattern recognition problem, soybean cultivar recognition has not yet received considerable attention. Traditional handcrafted leaf image analysis methods are limited to capture the subtle differences of leaf features among different cultivars. In this paper, we make the attempt of using deep learning to harvest discriminatory leaf features for soybean cultivar recognition. A novel deep learning model, named transformation attention network (TAN), is proposed in this work. It first focuses on extracting fine-grained leaf features via attention mechanism and then rectifies the leaf posture using affine transformations. We constructed a soybean leaf cultivar dataset which consists of 240 soybean cultivars with 10 samples per cultivar to examine the availability of cultivar information in leaf patterns and validate the effectiveness of the proposed deep learning model for soybean cultivar recognition. The encouraging experimental results confirm the effectiveness of leaf image patterns for distinguishing cultivars and demonstrate the better performance of the proposed method over the state-of-the-art handcrafted methods and deep learning methods for soybean cultivar recognition.

    参考文献
    相似文献
    引证文献
引用本文

游嘉伟,王斌,曾瑞.用于大豆品种识别的叶片深度特征学习方法.计算机系统应用,,():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
历史
  • 收稿日期:2021-04-15
  • 最后修改日期:2021-05-08
  • 录用日期:2021-05-11
  • 在线发布日期:
您是第位访问者
版权所有:中国科学院软件研究所 京ICP备05046678号-3
地址:北京海淀区中关村南四街4号 中科院软件园区 7号楼305房间,邮政编码:100190
电话:010-62661041 传真: Email:csa (a) iscas.ac.cn
技术支持:北京勤云科技发展有限公司

京公网安备 11040202500063号