Journal of Computer Applications ›› 2018, Vol. 38 ›› Issue (11): 3241-3245.DOI: 10.11772/j.issn.1001-9081.2018041309

Previous Articles     Next Articles

Plant image recoginiton based on family priority strategy

CAO Xiangying1, SUN Weimin1, ZHU Youxiang1, QIAN Xin2, LI Xiaoyu1, YE Ning1   

  1. 1. School of Information Technology, Nanjing Forestry University, Nanjing Jiangsu 210037, China;
    2. Housing and Real Estate Promotion Center of Jiangsu Provincial Department of Housing and Urban Rural Development, Nanjing Jiangsu 210009, China
  • Received:2018-04-29 Revised:2018-05-31 Online:2018-11-10 Published:2018-11-10
  • Supported by:
    This work is partially supported by the National Key Research and Development Program (2016YFD0600101), the National Natural Science Foundation of China (31570662, 31500533, 61401214), the Undergraduate Research and Innovation Program Project in Jiangsu Province (201710298029Z), the Jiangsu Provincial Department of Housing and Urban-Rural Development (2016ZD44), the Priority Academic Program of Jiangsu Higher Education Institutions.


曹香滢1, 孙卫民1, 朱悠翔1, 钱鑫2, 李晓宇1, 业宁1   

  1. 1. 南京林业大学 信息科学技术学院, 南京 210037;
    2. 江苏省住房和城乡建设厅住宅与房地产业促进中心, 南京 210009
  • 通讯作者: 业宁
  • 作者简介:曹香滢(1997-),女,江苏南通人,主要研究方向:深度学习;孙卫民(1997-),男,江苏盐城人,主要研究方向:深度学习;朱悠翔(1994-),男,福建福州人,主要研究方向:深度学习;钱鑫(1978-),男,江苏南京人,硕士,主要研究方向:房地产信息化;李晓宇(1998-),女,湖北荆州人,主要研究方向:深度学习;业宁(1967-),男,江苏江宁人,教授,博士,CCF会员,主要研究方向:数据挖掘。
  • 基金资助:

Abstract: Plant recognition includes two kinds of tasks:specimen recognition and real-environment recognition. Due to the existence of background noise, real-environment plant image recognition is more difficult. To reduce the weight of Convolutional Neural Networks (CNN), to improve over-fitting, to improve the recognition rate and generalization ability, a method of plant identification with Family Priority (FP) was proposed. Combined with the lightweight CNN MobileNet model, a plant recognition model Family Priority MobileNet (FP-MobileNet) was established by means of migration learning. On the single background plant dataset flavia, the MobileNet model achieved 99.8% of accuracy. For the more challenging real-environment flower dataset flower102, when the number of samples in the training set was greater than that in the test set FP-MobileNet achieved 99.56% of accuracy. When the number of samples in the training set was smaller than that in the test set, FP-MobileNet still obtained 95.56% of accuracy. The experimental results show that the accuracies of FP-MobileNet under two different data set partitioning schemes are both higher than those of the pure MobileNet model. In addition, FP-MobileNet weighs only occupy 13.7 MB with high recognition rate. It takes into account both accuracy and delay, and is suitable for promotion to mobile devices that require a lightweight model.

Key words: Family Priority (FP) strategy, real-environment plant image, plant image recognition, deep learning, Convolutional Neural Network (CNN)

摘要: 植物识别领域的研究包括单一背景和自然环境植物图像识别,由于背景噪声的存在,自然环境植物图像识别难度更大。针对如何降低卷积神经网络(CNN)的权重大小、如何改善过拟合、如何提高模型对自然环境植物的识别率和泛化能力的问题,提出科优先(FP)的植物识别方法。与轻量卷积神经网络MobileNet模型结合,利用迁移学习的方法,建立基于MobileNet的科优先(FP-MobileNet)植物识别模型。单纯使用MobileNet模型在单一背景植物数据集flavia上获得了99.8%的识别率;对于更具挑战的自然环境花卉数据集flower102,在训练集样本数量大于测试集时FP-MobileNet获得了99.56%识别率,在训练集样本数量小于测试集时FP-MobileNet仍获得了95.56%的识别率。实验结果表明,两种数据集划分方案下FP-MobileNet的识别率均高于单纯的MobileNet模型;并且FP-MobileNet模型在获得较高识别率的同时,权重仅占13.7 MB,兼顾了精度和延迟,适合推广到需要轻量模型的移动设备。

关键词: 科优先策略, 自然环境植物图像, 植物图像识别, 深度学习, 卷积神经网络

CLC Number: