基于卷积神经网络的小麦产量预估方法

doi:10.3969/j.issn.1004-1524.2020.12.16

浙江农业学报 ›› 2020, Vol. 32 ›› Issue (12): 2244-2252.DOI: 10.3969/j.issn.1004-1524.2020.12.16

基于卷积神经网络的小麦产量预估方法

鲍烈^a^,^b(), 王曼韬^a^,^b^,^*(), 刘江川^a^,^b, 文波^a^,^b, 明月^a^,^b

a.四川农业大学信息工程学院,四川雅安 625000
b.四川农业大学农业信息工程重点实验室,四川雅安 625000

收稿日期:2020-07-21 出版日期:2020-12-25 发布日期:2020-12-25
通讯作者: 王曼韬
作者简介:*王曼韬,E-mail: wangmantao@sicau.edu.cn
鲍烈(1994—),男,四川遂宁人,硕士,主要从事图像处理、深度学习研究。E-mail: mrbl12138@163.com
基金资助:
国家重点研发计划(2018YFD0501005);四川省教育厅青年基金(18ZB0467);四川农业大学国家级创新训练项目(201910626048)

Estimation method of wheat yield based on convolution neural network

BAO Lie^a^,^b(), WANG Mantao^a^,^b^,^*(), LIU Jiangchuan^a^,^b, WEN Bo^a^,^b, MING Yue^a^,^b

a. College of Information Engineering, Ya’an 625000, China
b. Key Laboratory of Agricultural Information Engineering of Sichuan Province,Sichuan Agricultural University, Ya’an 625000, China

Received:2020-07-21 Online:2020-12-25 Published:2020-12-25
Contact: WANG Mantao

摘要/Abstract

摘要：

小麦产量是评估农业生产力的重要指标之一,针对小麦产量人工预估困难,提出将卷积神经网络运用于小麦产量预估,为农业生产力的预估提供参考,指导农业生产管理决策。利用无人机分别在河南省新乡、漯河两地进行图片采集,并以之构建麦穗数据集,分为正样本(麦穗)和负样本(叶子和背景)。针对小麦常规的生理形态和生长环境,设计卷积神经网络识别模型,以图像金字塔构建多尺度滑动窗口,以非极大值抑制(NMS)去除重叠率较高的目标框,实现对单位面积内麦穗的计数,并利用随机采样的方式对大田麦穗进行单位面积图像采样,以采样图像中麦穗数量的平均值作为产量预估基准,进一步实现麦穗产量预估。随机抽取100幅不同小麦图片进行测试,与人工计数结果进行对比,准确率达到97.30%,漏检率为0.34%,误检率为2.36%,误差率为2.70%。试验结果表明,此方法能够克服环境中的多种噪声干扰,能够在不同光照条件下对麦穗进行计数和产量预估。

关键词: 图像处理, 深度学习, 卷积神经网络, 图像金字塔, 产量预估

Abstract:

Wheat yield is one of the important indices to evaluate agricultural productivity. In order to solve the problem of wheat yield prediction, convolution neural network was applied. It could provide reference for the estimation of agricultural productivity and guide the decision-making of agricultural production management. We constructed the wheat dataset from the images which were collected by unmanned aerial vehicle in Xinxiang and Luohe of Henan Province, and divided them into positive samples (wheat) and negative samples (leaf and background). The convolutional neural network recognition model was designed for normal physiological morphology and growth environment of wheat, and the image pyramid was used to construct multi-scale sliding windows. Finally, Non-maximum suppression (NMS) was used to remove the object frame with high overlap rate to realize counting of wheat ears per unit area. We randomly assigned 100 different wheat images to test. Compared with manual counting result, the accuracy rate reached 97.30%, the missed detection rate was 0.34%, the wrong detection rate was 2.36%, and the error rate was 2.70%. The experimental results showed that this method could overcome many kinds of noise interference in the environment, and could count wheat ear and estimate the yield of wheat accurately.

Key words: image processing, deep learning, convolutional neural network, pyramid of images, yield estimation

中图分类号:

S512.1

鲍烈, 王曼韬, 刘江川, 文波, 明月. 基于卷积神经网络的小麦产量预估方法[J]. 浙江农业学报, 2020, 32(12): 2244-2252.

BAO Lie, WANG Mantao, LIU Jiangchuan, WEN Bo, MING Yue. Estimation method of wheat yield based on convolution neural network[J]. Acta Agriculturae Zhejiangensis, 2020, 32(12): 2244-2252.

图/表 14

参考文献 26

[1]	LU Y, YI S J, ZENG N Y, et al. Identification of rice diseases using deep convolutional neural networks[J]. Neurocomputing, 2017,267:378-384.
[2]	JOHANNES A, PICON A, ALVAREZ-GILA A, et al. Automatic plant disease diagnosis using mobile capture devices, applied on a wheat use case[J]. Computers and Electronics in Agriculture, 2017,138:200-209.
[3]	王献锋, 丁军, 朱义海. 一种改进的深度置信网络在棉花病虫害预测中的应用[J]. 棉花学报, 2018,30(4):300-307. DOI URL
	WANG X F, DING J, ZHU Y H. Application of modified deep belief network in forecasting cotton diseases and insect pests[J]. Cotton Science, 2018,30(4):300-307.(in Chinese with English abstract)
[4]	张建华, 祁力钧, 冀荣华, 等. 基于粗糙集和BP神经网络的棉花病害识别[J]. 农业工程学报, 2012,28(7):161-167.
	ZHANG J H, QI L J, JI R H, et al. Cotton diseases identification based on rough sets and BP neural network[J]. Transactions of the Chinese Society of Agricultural Engineering, 2012,28(7):161-167.(in Chinese with English abstract)
[5]	LU H, CAO Z G, XIAO Y, et al. TasselNet: counting maize tassels in the wild via local counts regression network[J]. Plant Methods, 2017,13(1):1-17. DOI URL
[6]	UBBENS J, CIESLAK M, PRUSINKIEWICZ P, et al. The use of plant models in deep learning: an application to leaf counting in rosette plants[J]. Plant Methods, 2018,14:6-15. DOI URL PMID
[7]	TANG J L, WANG D, ZHANG Z G, et al. Weed identification based on K-means feature learning combined with convolutional neural network[J]. Computers and Electronics in Agriculture, 2017,135:63-70.
[8]	DOS SANTOS FERREIRA A, MATTE FREITAS D, GONÇALVES DA SILVA G, et al. Weed detection in soybean crops using ConvNets[J]. Computers and Electronics in Agriculture, 2017,143:314-324.
[9]	MEHDIPOUR GHAZI M, YANIKOGLU B, APTOULA E. Plant identification using deep neural networks via optimization of transfer learning parameters[J]. Neurocomputing, 2017,235:228-235.
[10]	DYRMANN M, KARSTOFT H, MIDTIBY H S. Plant species classification using deep convolutional neural network[J]. Biosystems Engineering, 2016,151:72-80.
[11]	刘涛, 孙成明, 王力坚, 等. 基于图像处理技术的大田麦穗计数![J]. 农业机械学报, 2014,45(2):282-290.
	LIU T, SUN C M, WANG L J, et al. In-field wheatear counting based on image processing technology[J]. Transactions of the Chinese Society for Agricultural Machinery, 2014,45(2):282-290.(in Chinese with English abstract)
[12]	赵锋. 基于颜色特征和改进Adaboost算法的麦穗识别的研究[D]. 保定: 河北农业大学, 2014.
	ZHAO F. Research on wheat ear recognition based on color features and improved Adaboost algorithm[D]. Baoding: Hebei Agricultural University, 2014.(in Chinese with English abstract)
[13]	刘哲, 黄文准, 王利平. 基于改进K-means聚类算法的大田麦穗自动计数[J]. 农业工程学报, 2019,35(3):174-181.
	LIU Z, HUANG W Z, WANG L P. Field wheat ear counting automatically based on improved K-means clustering algorithm[J]. Transactions of the Chinese Society of Agricultural Engineering, 2019,35(3):174-181.(in Chinese with English abstract)
[14]	李毅念, 杜世伟, 姚敏, 等. 基于小麦群体图像的田间麦穗计数及产量预测方法[J]. 农业工程学报, 2018,34(21):185-194.
	LI Y N, DU S W, YAO M, et al. Method for wheatear counting and yield predicting based on image of wheatear population in field[J]. Transactions of the Chinese Society of Agricultural Engineering, 2018,34(21):185-194.(in Chinese with English abstract)
[15]	范梦扬, 马钦, 刘峻明, 等. 基于机器视觉的大田环境小麦麦穗计数方法[J]. 农业机械学报, 2015,46(增刊):234-239.
	FAN M Y, MA Q, LIU J M, et al. Counting method of wheatear in field based on machine vision technology[J]. Transactions of the Chinese Society for Agricultural Machinery, 2015,46(Suppl.):234-239.(in Chinese with English abstract)
[16]	张领先, 陈运强, 李云霞, 等. 基于卷积神经网络的冬小麦麦穗检测计数系统[J]. 农业机械学报, 2019,50(3):144-150.
	ZHANG L X, CHEN Y Q, LI Y X, et al. Detection and counting system for winter wheat ears based on convolutional neural network[J]. Transactions of the Chinese Society for Agricultural Machinery, 2019,50(3):144-150.(in Chinese with English abstract)
[17]	张建华, 孔繁涛, 吴建寨, 等. 基于改进VGG卷积神经网络的棉花病害识别模型[J]. 中国农业大学学报, 2018,23(11):161-171.
	ZHANG J H, KONG F T, WU J Z, et al. Cotton disease identification model based on improved VGG convolution neural network[J]. Journal of China Agricultural University, 2018,23(11):161-171.(in Chinese with English abstract)
[18]	陈含, 吕行军, 田凤珍, 等. 基于Sobel算子边缘检测的麦穗图像分割[J]. 农机化研究, 2013,35(3):33-36.
	CHEN H, LYU X J, TIAN F Z, et al. Wheat panicle image segmentation based on Sobel operator-edge detection[J]. Journal of Agricultural Mechanization Research, 2013,35(3):33-36.(in Chinese with English abstract)
[19]	LIU W B, WANG Z D, LIU X H, et al. A survey of deep neural network architectures and their applications[J]. Neurocomputing, 2017,234:11-26.
[20]	马永强, 王顺利, 孙伟, 等. 基于高斯金字塔和拉普拉斯金字塔融合的图像对比度增强算法研究[J]. 信息与电脑(理论版), 2018(4):38-40.
	MA Y Q, WANG S L, SUN W, et al. Research on image contrast enhancement algorithm based on fusion of Gaussian pyramid and laplacian pyramid[J]. China Computer & Communication, 2018(4):38-40.(in Chinese with English abstract)
[21]	朱伟, 刘健, 竺明月, 等. 基于高斯-拉普拉斯金字塔的DR图像增强改进算法研究[J]. 中国医疗器械杂志, 2019,43(1):10-13. PMID
	ZHU W, LIU J, ZHU M Y, et al. Research on improved algorithm of DR image enhancement based on Gauss-Laplacian pyramid[J]. Chinese Journal of Medical Instrumentation, 2019,43(1):10-13.(in Chinese with English abstract) URL PMID
[22]	ZORAN S Š, KOVAČEK K I. Sliding window object detection without spatial clustering of raw detection responses [J]. IFAC Proceedings Volumes, 2012,45(22):114-119.
[23]	赵文清, 严海, 邵绪强. 改进的非极大值抑制算法的目标检测[J]. 中国图象图形学报, 2018,23(11):1676-1685.
	ZHAO W Q, YAN H, SHAO X Q. Object detection based on improved non-maximum suppression algorithm[J]. Journal of Image and Graphics, 2018,23(11):1676-1685.(in Chinese with English abstract)
[24]	张婷婷. 基于小麦群体图像的麦穗计数方法研究[D]. 合肥: 安徽大学, 2020.
	ZHANG T T. Study on wheat ear counting method based on wheat population image[D]. Hefei: Anhui University, 2020.
[25]	高云鹏. 基于深度神经网络的大田小麦麦穗检测方法研究[D]. 北京: 北京林业大学, 2019.
	GAO Y P. Study on detection method of wheat ear in field based on deep neural network[D]. Beijing: Beijing Forestry University, 2019.(in Chinese with English abstract)
[26]	李鹏. 基于无人机图像的麦穗识别技术研究[D]. 郑州: 河南农业大学, 2019.
	LI P. Research on wheat ear recognition technology based on UAV image[D]. Zhengzhou: Henan Agricultural University, 2019.(in Chinese with English abstract)

样本 Samples	总数据集 Total data set	训练集 Training set	验证集 Validation set	测试集 Test set
麦穗Wheat	31 505	27 505	3 000	1 000
叶子与背景	29 987	25 987	3 000	1 000
Leaf and background

样本 Samples	总数据集 Total data set	训练集 Training set	验证集 Validation set	测试集 Test set
麦穗Wheat	31 505	27 505	3 000	1 000
叶子与背景	29 987	25 987	3 000	1 000
Leaf and background

计数方式 Method	总麦穗数量 Total wheat quantity	正确框数 Number of correct boxes	误检数 Number of mistakenly identified boxes	漏检数 Number of missed boxes	误差数 Number of wrong boxes
人工计数Manual	12 530	12 530	0	0	0
Wheat-Net	12 632	12 288	298	43	341

计数方式 Method	总麦穗数量 Total wheat quantity	正确框数 Number of correct boxes	误检数 Number of mistakenly identified boxes	漏检数 Number of missed boxes	误差数 Number of wrong boxes
人工计数Manual	12 530	12 530	0	0	0
Wheat-Net	12 632	12 288	298	43	341

模型 Model	漏检率 Missed detection rate/%	误检率 Wrong detection rate/%	误差率 Error rate/%	准确率 Accuracy rate/%	每张所需时间 Time per sheet/s
VGG-16	2.38	2.65	5.03	94.97	0.433
Wheat-Net	0.34	2.36	2.70	97.30	0.115

基于卷积神经网络的小麦产量预估方法

Estimation method of wheat yield based on convolution neural network

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 26

相关文章 15

编辑推荐

Metrics

本文评价

编号 No.	人工统计 Artificial statistics	机器统计 Machine statistics	正确框数 Number of correct boxes	误检数 Number of mistakenly identified boxes	漏检数 Number of missed boxes	误差数 Number of wrong boxes	误差率 Error rate/%
1	134	142	133	9	1	10	7.46
2	121	124	120	4	1	5	4.13
3	132	130	129	1	3	4	3.03
4	127	126	125	1	2	3	2.36
5	118	122	118	4	0	4	3.39
6	117	114	114	0	3	3	2.56
7	92	95	92	3	0	3	3.26
8	107	104	102	2	5	7	6.54
9	107	107	106	1	1	2	1.87
10	133	130	129	1	4	5	3.76

拍摄时间 Shooting time	人工统计 Artificial statistics (total)	机器统计 Machine statistics	正确框数 Number of correct boxes	误检数 Number of mistakenly identified boxes	漏检数 Number of missed boxes	误差数 Number of wrong boxes	误差率 Error rate/%
傍晚Nightfall	7 431	7 631	7 080	221	21	242	3.25
中午Noon	6 984	7 251	6 755	298	31	329	4.71

方法 Method	漏检率 Missed detection rate/%	误检率 Wrong detection rate/%	误差率 Error rate/%	准确率 Accuracy rate/%	每张所需时间 Time per sheet/s
Hourglass Network	5.09	3.54	8.63	91.37	—
YOLOv3	0.90	12.03	12.88	87.12	0.120
Mask R-CNN	1.50	1.50	3.00	97.00	0.940
基于颜色和纹理特征	—	—	3.45	96.55	—
Based on color and texture features
Wheat Net	0.34	2.36	2.70	97.30	0.115

大田编号 Field No.	m/g	t	a/m²	小区预估产量 Estimation yield/kg	人工称量 Artificial weighing/ (kg·m^-2)	预估产量 Algorithm estimation/ (kg·m^-2)	误差 Error/ (kg·m^-2)	误差率 Error rate/%
新乡1号大田 Xinxiang No.1	5.02	141	42 000	29 728.44	0.69	0.71	0.02	2.90
新乡2号大田 Xinxiang No.2	4.28	158	30 000	20 287.20	0.66	0.68	0.02	3.03
漯河1号大田 Luohe No.1	4.93	135	18 000	11 979.90	0.65	0.67	0.02	3.08
漯河2号大田 Luohe No.2	4.37	148	24 000	15 522.24	0.62	0.65	0.03	4.84

[1]	蒋丰千, 李旸, 余大为, 孙敏, 张恩宝. 基于Caffe卷积神经网络的大豆病害检测系统[J]. 浙江农业学报, 2019, 31(7): 1177-1183.
[2]	刘志, 贺正, 苗芳芳, 贾彪. 基于无人机的水肥一体化玉米出苗率估算方法与试验[J]. 浙江农业学报, 2019, 31(6): 977-985.
[3]	王彦翔, 张艳, 杨成娅, 孟庆龙, 尚静. 基于深度学习的农作物病害图像识别技术进展[J]. 浙江农业学报, 2019, 31(4): 669-676.
[4]	金秀, 卢杰, 傅运之, 王帅, 许高健, 李绍稳. 基于深度卷积神经网络的小麦赤霉病高光谱病症点分类方法[J]. 浙江农业学报, 2019, 31(2): 315-325.
[5]	步正延, 李臻峰, 宋飞虎, 李斌, 李静. 基于太赫兹成像技术的大豆叶片水分含量测定[J]. 浙江农业学报, 2018, 30(8): 1420-1426.
[6]	杨国亮, 许楠, 康乐乐, 龚曼, 洪志阳. 基于参数指数非线性残差神经网络的脐橙病变叶片识别[J]. 浙江农业学报, 2018, 30(6): 1073-1081.
[7]	顾正敏, 李臻峰, 宋飞虎, 张君生, 庄为. 基于光场相机的大豆冠层叶面积无损测量方法研究[J]. 浙江农业学报, 2018, 30(12): 2144-2152.
[8]	张善文, 黄文准, 尤著宏. 基于物联网和深度卷积神经网络的冬枣病害识别方法[J]. 浙江农业学报, 2017, 29(11): 1868-1874.
[9]	李颀, 胡艺聪, 武付闯. 基于图像处理的育苗箱环境控制系统[J]. 浙江农业学报, 2017, 29(11): 1912-1919.
[10]	赵君爱1，2，魏艳春3. 基于改进中值滤波的图像噪声去除算法的研究[J]. 浙江农业学报, 2015, 27(6): 1078-.
[11]	石铁，索雪松*，周靖仁，韩昌兴，高立艾. 现代农业设施的溯源管理的研究 ——基于图像识别的太阳能电池片分析[J]. 浙江农业学报, 2015, 27(2): 285-.
[12]	李振1，2，廖同庆1，冯青春2，3，4，5，张东彦1，2，3，4，5，王秀2，3，4，5，*. 基于机器视觉的蔬菜种子活力指数检测算法研究及系统实现[J]. 浙江农业学报, 2015, 27(12): 2218-.
[13]	胡孟晗;董庆利;*;刘宝林;张重阳;叶飞. 基于椭圆傅里叶描述子的香蕉形状识别[J]. , 2014, 26(1): 200-205.
[14]	姚立健;边起;雷良育;赵大旭. 基于BP神经网络的水果分级研究[J]. , 2012, 24(5): 0-930.
[15]	刘建军;姚立健;彭樟林. 基于机器视觉的山核桃等级检测技术[J]. , 2010, 22(6): 854-858.