基于深度学习的玉米抽雄期判识
|
摘要
农作物自动化观测是农业现代化和自动化进展的重要标志和不可缺少的一部分。目前作物生长自动化观测主要通过获取作物生长图像,进而对图像进行处理分析来获取作物生长特征等信息,利用传统的图像分割、特征点检测等方法检测作物的生长特征误差较大。近年来,深度学习迅速发展,且在多个领域得到广泛应用。在ImageNet数据集和大量的玉米作物图像基础上,对深度学习中常用的CNN进行训练和微调,充分利用迁移学习的优势,得到识别模型,对玉米雄穗进行识别,进而对玉米抽雄期进行判识。实验证明深度学习方法在作物生长特征识别上有良好的效果,比传统方法有显著提高。
Automatic observation of crop is one of the core components of modern agriculture and automation progress. At present, automatic crop growth observation is mainly by obtaining crop growth images and then processing and analyzing the images to obtain growth characteristics information of crop. Traditional methods such as image segmentation and feature point detection have obvious error when used to detect crop growth characteristic. In recent years, deep learning has developed rapidly and has been widely used in many fields. In this paper, based on ImageNet data set and a large number of images of corn crops, the CNN which frequently used in deep learning is trained and fine-tuned. The advantage of transfer learning is fully utilized to obtain the recognition model, to identify the male ear of maize and then to judge the tasseling period of maize. The experimental results show that the deep learning method has a good effect on the recognition of crop growth characteristics, and is significantly improved than the traditional method.
Automatic observation of crop is one of the core components of modern agriculture and automation progress. At present, automatic crop growth observation is mainly by obtaining crop growth images and then processing and analyzing the images to obtain growth characteristics information of crop. Traditional methods such as image segmentation and feature point detection have obvious error when used to detect crop growth characteristic. In recent years, deep learning has developed rapidly and has been widely used in many fields. In this paper, based on ImageNet data set and a large number of images of corn crops, the CNN which frequently used in deep learning is trained and fine-tuned. The advantage of transfer learning is fully utilized to obtain the recognition model, to identify the male ear of maize and then to judge the tasseling period of maize. The experimental results show that the deep learning method has a good effect on the recognition of crop growth characteristics, and is significantly improved than the traditional method.
引文
[1] 孙红,李民赞,周志艳,等.基于光谱技术的水稻稻纵卷叶螟受害区域检测[J].光谱学与光谱分析,2010,30(4):1080-1083.
[2] 祖琴,张水发,曹阳,等.结合光谱图像技术和SAM分类法的甘蓝中杂草识别研究[J].光谱学与光谱分析,2015,35(2):479-485.
[3] 白敬,徐友,魏新华,等.基于光谱特性分析的冬油菜苗期田间杂草识别[J].农业工程学报,2013,29(20):128-134.
[4] 王璨,武新慧,李志伟.基于卷积神经网络提取多尺度分层特征识别玉米杂草[J].农业工程学报,2018,34(5):144-151.
[5] 杨建姣,齐迹,朱凤武.基于遗传神经网络的田间杂草识别的研究[J].中国农机化学报,2016,37(9):223-226.
[6] 李荣春,陶洪斌,张竹琴,等.基于图像处理技术的夏玉米群体长势监测研究[J].玉米科学,2010,18(2):128-132.
[7] 刘永娟.基于计算机视觉技术的玉米发育期识别研究[D].无锡:江南大学,2017.
[8] 陆明,申双和,王春艳,等.基于图像识别技术的夏玉米生育期识别方法初探[J].中国农业气象,2011,32(3):423-429.
[9] 吴茜.基于图像处理技术的棉花发育期自动观测研究[D].武汉:华中科技大学,2013.
[10] 孙华生,黄敬峰,彭代亮.利用MODIS数据识别水稻关键生长发育期[J].遥感学报,2009,13(6):1130-1146.
[11] 张芸德,刘蓉,刘明,等.基于深度学习特征的玉米生长期识别[J].电子测量技术,2018,41(16):79-84.
[12] 韩悦婷,行鸿彦,金红伟.基于opencv的玉米出苗期和三叶期自动检测系统的设计[J]电子测量与仪器学报.2017,31(10):1574-1581.
[13] KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[C].Proceedings of Advances in Neural Information Prossing Systerms,Cambrigdge,2012:1106-1114.
[14] GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich features hierarchies for accurate object detection and semantic segmentation [C].Processing of the 2014 IEEE Conference on Computer Vision and Pattern Recognition,2014:580-587.
[15] HUANG G B,LEE H,ERIK G.Learning hierarchical re-presentation for face verification with convolution deep belief networks[C].CVRP 12:Proceeding of the 2012 IEEE Conference on Computer Vision and Pattern Recognition,2012:2518-2525.
[16] SHAN S L,KHALIL-HANI M,RADZI S A,et al.Gender classification:a convolutional neural network approach[J].Turkish Journal of Electrical Engineering & Computer Sciences,2016,24(3):1248-1264.
[2] 祖琴,张水发,曹阳,等.结合光谱图像技术和SAM分类法的甘蓝中杂草识别研究[J].光谱学与光谱分析,2015,35(2):479-485.
[3] 白敬,徐友,魏新华,等.基于光谱特性分析的冬油菜苗期田间杂草识别[J].农业工程学报,2013,29(20):128-134.
[4] 王璨,武新慧,李志伟.基于卷积神经网络提取多尺度分层特征识别玉米杂草[J].农业工程学报,2018,34(5):144-151.
[5] 杨建姣,齐迹,朱凤武.基于遗传神经网络的田间杂草识别的研究[J].中国农机化学报,2016,37(9):223-226.
[6] 李荣春,陶洪斌,张竹琴,等.基于图像处理技术的夏玉米群体长势监测研究[J].玉米科学,2010,18(2):128-132.
[7] 刘永娟.基于计算机视觉技术的玉米发育期识别研究[D].无锡:江南大学,2017.
[8] 陆明,申双和,王春艳,等.基于图像识别技术的夏玉米生育期识别方法初探[J].中国农业气象,2011,32(3):423-429.
[9] 吴茜.基于图像处理技术的棉花发育期自动观测研究[D].武汉:华中科技大学,2013.
[10] 孙华生,黄敬峰,彭代亮.利用MODIS数据识别水稻关键生长发育期[J].遥感学报,2009,13(6):1130-1146.
[11] 张芸德,刘蓉,刘明,等.基于深度学习特征的玉米生长期识别[J].电子测量技术,2018,41(16):79-84.
[12] 韩悦婷,行鸿彦,金红伟.基于opencv的玉米出苗期和三叶期自动检测系统的设计[J]电子测量与仪器学报.2017,31(10):1574-1581.
[13] KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[C].Proceedings of Advances in Neural Information Prossing Systerms,Cambrigdge,2012:1106-1114.
[14] GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich features hierarchies for accurate object detection and semantic segmentation [C].Processing of the 2014 IEEE Conference on Computer Vision and Pattern Recognition,2014:580-587.
[15] HUANG G B,LEE H,ERIK G.Learning hierarchical re-presentation for face verification with convolution deep belief networks[C].CVRP 12:Proceeding of the 2012 IEEE Conference on Computer Vision and Pattern Recognition,2012:2518-2525.
[16] SHAN S L,KHALIL-HANI M,RADZI S A,et al.Gender classification:a convolutional neural network approach[J].Turkish Journal of Electrical Engineering & Computer Sciences,2016,24(3):1248-1264.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |