基于深度学习和GEE的作物遥感分类

doi:10.6048/j.issn.1001-4330.2024.11.022

新疆农业科学 ›› 2024, Vol. 61 ›› Issue (11): 2807-2814.DOI: 10.6048/j.issn.1001-4330.2024.11.022

基于深度学习和GEE的作物遥感分类

赵昊楠¹(), 马海燕²^,³, 阿斯娅·曼力克²^,³, 田聪²^,³, 徐俊²^,³, 潘竟¹, 孙宗玖¹, 郑逢令²^,³()

1.新疆农业大学草业学院,乌鲁木齐 830052
2.新疆畜牧科学院草业研究所,乌鲁木齐 830057
3.新疆畜牧科学院天山北坡草地生态环境野外定位观测研究站,乌鲁木齐 830057

收稿日期:2024-03-11 出版日期:2024-11-20 发布日期:2025-01-08
通信作者: 郑逢令(1974-),男,新疆乌鲁木齐人,副研究员,博士,硕士生导师,研究方向为植被遥感和空间信息技术,(E-mail)xjzheng@sina.com
作者简介:赵昊楠(1999-),男,河北张家口人,硕士研究生,研究方向为植被遥感和空间信息技术,(E-mail)zhn113355@gmail.com
基金资助:
新疆维吾尔自治区自然科学基金项目“基于多源遥感时序数据的人工草地识别与估产研究-以苜蓿为例”(2023D01A75);新疆维吾尔自治区奶产业技术体系(XJARS-11);国家自然科学基金项目(31860679)

Crop remote sensing classification study in Qitai County based on deep learning and Google earth engine (GEE)

ZHAO Haonan¹(), MA Haiyan²^,³, Asiya Manlike²^,³, TIAN Cong²^,³, XU Jun²^,³, PAN Jing¹, SUN Zongjiu¹, ZHENG Fengling²^,³()

1. College of Grassland Science, Xinjiang Agricultural University, Urumqi 830052,China
2. Grassland Research Institute of Xinjiang Academy of Animal Sciences, Urumqi 830057,China
3. Xinjiang Academy of Animal Science Field Orientation Observation and Research Station of Grassland Ecological Environment on the Northern Slope of Tianshan Mountains, Urumqi 830057,China

Received:2024-03-11 Published:2024-11-20 Online:2025-01-08
Supported by:
Sponsored by Natural Science Foundation of Xinjiang Uygur Autonomous Region"Research on Identification and Yield Estimation of Artificial Grassland Based on Multi-Source Remote Sensing Time-Series Data - A Case of Alfalfa"(2023D01A75);Supported by the earmarked fund for XJARS(XJARS-11);The National Natural Science Foundation of China(31860679)

摘要/Abstract

摘要：

【目的】利用遥感数据和深度学习方法准确获取农作物的种植结构和分布。【方法】通过实地调查新疆奇台县获取样本集,借助Google Earth Engine云平台获取Sentinel-2号和Sentinel-1号影像,利用Google Colab进行深度学习算法的模型训练和验证,调整和优化深度学习的相关参数来提升分类精度,并且比较了深度学习、随机森林和支持向量机3种分类方法的精度。【结果】深度学习的分类精度最高,总体精度达到94.6%。【结论】利用深度学习算法可实现奇台县农作物种植结构的精准监测。

关键词: 农作物; 遥感分类; 深度学习; 谷歌地球引擎; Google Colab

Abstract:

【Objective】 This study aims to accurately acquire the crop planting structure and distribution using remote sensing data and deep learning methods in view of the complex crop planting structure.【Methods】 A sample set was obtained through field investigations.Sentinel-2 and Sentinel-1 images were acquired using the Google Earth Engine cloud platform; Model training and validation for deep learning algorithms were conducted using Google Colab; Classification accuracy was improved by adjusting and optimizing relevant parameters of deep learning.Additionally, the accuracy of three classification methods—deep learning, random forest, and support vector machine—was compared.【Results】 The deep learning approach achieved the highest classification accuracy, with an overall accuracy of 94.6%.【Conclusion】 The utilization of deep learning algorithms enables precise monitoring of crop planting structure in Qitai County.

Key words: crop; remote sensing classification; deep learning; Google Earth Engine; Google Colab

中图分类号:

S11+7

赵昊楠, 马海燕, 阿斯娅·曼力克, 田聪, 徐俊, 潘竟, 孙宗玖, 郑逢令. 基于深度学习和GEE的作物遥感分类[J]. 新疆农业科学, 2024, 61(11): 2807-2814.

ZHAO Haonan, MA Haiyan, Asiya Manlike, TIAN Cong, XU Jun, PAN Jing, SUN Zongjiu, ZHENG Fengling. Crop remote sensing classification study in Qitai County based on deep learning and Google earth engine (GEE)[J]. Xinjiang Agricultural Sciences, 2024, 61(11): 2807-2814.

图/表 9

参考文献 22

[1]	王爱芳, 张运, 黄静, 等. 作物遥感分类研究进展[J]. 测绘与空间地理信息, 2021, 44(10): 80-83, 88.
	WANG Aifang, ZHANG Yun, HUANG Jing, et al. Recent progresses in research of crop classification by using remote sensing[J]. Geomatics & Spatial Information Technology, 2021, 44(10): 80-83, 88.
[2]	赵红伟, 陈仲新, 刘佳. 深度学习方法在作物遥感分类中的应用和挑战[J]. 中国农业资源与区划, 2020, 41(2): 35-49.
	ZHAO Hongwei, CHEN Zhongxin, LIU Jia. Deep learning for crop classification of remote sensing data: applications and challenges[J]. Chinese Journal of Agricultural Resources and Regional Planning, 2020, 41(2): 35-49.
[3]	Onojeghuo A O, Blackburn G A, Wang Q M, et al. Mapping paddy rice fields by applying machine learning algorithms to multi-temporal Sentinel-1A and Landsat data[J]. International Journal of Remote Sensing, 2018, 39(4): 1042-1067.
[4]	Lin Z X, Zhong R H, Xiong X G, et al. Large-scale rice mapping using multi-task spatiotemporal deep learning and sentinel-1 SAR time series[J]. Remote Sensing, 2022, 14(3): 699.
[5]	张乾坤, 蒙继华, 任超. 构建地块二维表征及CNN模型的作物遥感分类[J]. 遥感学报, 2022, 26(7): 1437-1449.
	ZHANG Qiankun, MENG Jihua, REN Chao. Crop classification based on two-dimensional representation and CNN model from remote sensing[J]. National Remote Sensing Bulletin, 2022, 26(7): 1437-1449.
[6]	黄翀, 侯相君. 基于Bi-LSTM模型的时间序列遥感作物分类研究[J]. 中国农业科学, 2022, 55(21): 4144-4157. DOI
	HUANG Chong, HOU Xiangjun. Crop classification with time series remote sensing based on Bi-LSTM model[J]. Scientia Agricultura Sinica, 2022, 55(21): 4144-4157. DOI
[7]	Zhong L H, Gong P, Biging G S. Efficient corn and soybean mapping with temporal extendability: a multi-year experiment using Landsat imagery[J]. Remote Sensing of Environment, 2014, 140: 1-13.
[8]	Luo C, Liu H J, Lu L, et al. Monthly composites from Sentinel-1 and Sentinel-2 images for regional major crop mapping with Google Earth Engine[J]. Journal of Integrative Agriculture, 2021, 20(7): 1944-1957. DOI
[9]	Wang D, Liu C A, Zeng Y, et al. Dryland crop classification combining multitype features and multitemporal quad-polarimetric RADARSAT-2 imagery in Hebei Plain, China[J]. Sensors, 2021, 21(2): 332.
[10]	Kang Y P, Meng Q Y, Liu M, et al. Crop classification based on red edge features analysis of GF-6 WFV data[J]. Sensors, 2021, 21(13): 4328.
[11]	常布辉, 王军涛, 罗玉丽, 等. 河套灌区沈乌灌域GF-1/WFV遥感耕地提取[J]. 农业工程学报, 2017, 33(23): 188-195.
	CHANG Buhui, WANG Juntao, LUO Yuli, et al. Cultivated land extraction based on GF-1/WFV remote sensing in Shenwu irrigation area of Hetao Irrigation District[J]. Transactions of the Chinese Society of Agricultural Engineering, 2017, 33(23): 188-195.
[12]	马凯, 罗泽. 基于卷积神经网络的青海湖区域遥感影像分类[J]. 计算机系统应用, 2018, 27(9): 137-142.
	MA Kai, LUO Ze. Classification of remote sensing images in Qinghai Lake based on convolutional neural network[J]. Computer Systems & Applications, 2018, 27(9): 137-142.
[13]	杨泽航, 王文, 鲍健雄. 融合多源遥感数据的黑河中游地区生长季早期作物识别[J]. 地球信息科学学报, 2022, 24(5): 996-1008. DOI
	YANG Zehang, WANG Wen, BAO Jianxiong. Identifying crop types in early growing season in the middle reaches of Heihe River by fusing multi-source remote sensing data[J]. Journal of Geo-Information Science, 2022, 24(5): 996-1008.
[14]	Teluguntla P, Thenkabail P S, Oliphant A, et al. A 30-m landsat-derived cropland extent product of Australia and China using random forest machine learning algorithm on Google Earth Engine cloud computing platform[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 144: 325-340.
[15]	李前景, 刘珺, 米晓飞, 等. 面向对象与卷积神经网络模型的GF-6 WFV影像作物分类[J]. 遥感学报, 2021, 25(2): 549-558.
	LI Qianjing, LIU Jun, MI Xiaofei, et al. Object-oriented crop classification for GF-6 WFV remote sensing images based on Convolutional Neural Network[J]. National Remote Sensing Bulletin, 2021, 25(2): 549-558.
[16]	杨庆振, 郭敏, 范新成. 基于随机森林算法的高光谱遥感作物分类[J]. 测绘与空间地理信息, 2023, 46(4): 149-151, 154.
	YANG Qingzhen, GUO Min, FAN Xincheng. Hyper-spectral remote sensing crop classification based on random forest algorithm[J]. Geomatics & Spatial Information Technology, 2023, 46(4): 149-151, 154.
[17]	王利军, 郭燕, 贺佳, 等. 基于决策树和SVM的Sentinel-2A影像作物提取方法[J]. 农业机械学报, 2018, 49(9): 146-153.
	WANG Lijun, GUO Yan, HE Jia, et al. Classification method by fusion of decision tree and SVM based on sentinel-2A image[J]. Transactions of the Chinese Society for Agricultural Machinery, 2018, 49(9): 146-153.
[18]	Yao J X, Wu J, Xiao C Z, et al. The classification method study of crops remote sensing with deep learning, machine learning, and google earth engine[J]. Remote Sensing, 2022, 14(12): 2758.
[19]	薛乃丹. 基于高分卫星遥感影像的农作物分类方法研究[D]. 沈阳: 沈阳农业大学, 2022.
	XUE Naidan. Study on Crop Classification Method Based on GF Satellite Remote Sensing Image[D]. Shenyang: Shenyang Agricultural University, 2022.
[20]	张国顺. 基于高分辨率影像的北疆农作物分类研究[D]. 石河子: 石河子大学, 2020.
	ZHANG Guoshun. Crop Classification of Northern Xinjiang Based on High Resolution Imagery[D]. Shihezi: Shihezi University, 2020.
[21]	杨舒婷. 基于深度学习的高分辨率遥感影像农作物分类算法研究[D]. 长春: 吉林大学, 2021.
	YANG Shuting. Research on Crop Classification Algorithms Based on Deep Learning Using High-resolution Remote Sensing Imagery[D]. Changchun: Jilin University, 2021.
[22]	马永建. 基于CNN的高分辨率遥感影像典型农作物分类方法研究[D]. 石河子: 石河子大学, 2020.
	MA Yongjian. Study on Typical Crops Classification with High-resolution Remote Sensing Images Based on CNN[D]. Shihezi: Shihezi University, 2020.

编辑推荐 0

Metrics

阅读次数

全文

HTML			PDF

最新录用	在线预览	正式出版	最新录用	在线预览	正式出版
0	0	4	0	0	7

来源	本网站	其他网站

次数	9	2
比例	82%	18%

摘要

最新录用	在线预览	正式出版

0	0	65

来源	本网站	其他网站

次数	30	35
比例	46%	54%

数据源 The data source	波段数 Band	空间分辨率 Spatial resolution	重访周期日 Revisit cycle (d)
Sentinel-1	VV	10 m	12
Sentinel-1	VH	10 m	12
Sentinel-2	B1(Aerosols)	60 m	10
	B2(Blue)	10 m
	B3(Green)	10 m
	B4(Red)	10 m
	B5(Red Edge 1)	20 m
	B6(Red Edge 2)	20 m
	B7(Red Edge 3)	20 m
	B8(NIR)	10 m
	B8A(Red Edge 4 )	20 m
	B9(Water vapor)	60 m
	B10(Cirrus)	60 m
	B11(短波红外)	20 m
	B12(短波红外)	20 m

数据源 The data source	波段数 Band	空间分辨率 Spatial resolution	重访周期日 Revisit cycle (d)
Sentinel-1	VV	10 m	12
Sentinel-1	VH	10 m	12
Sentinel-2	B1(Aerosols)	60 m	10
	B2(Blue)	10 m
	B3(Green)	10 m
	B4(Red)	10 m
	B5(Red Edge 1)	20 m
	B6(Red Edge 2)	20 m
	B7(Red Edge 3)	20 m
	B8(NIR)	10 m
	B8A(Red Edge 4 )	20 m
	B9(Water vapor)	60 m
	B10(Cirrus)	60 m
	B11(短波红外)	20 m
	B12(短波红外)	20 m

植被指数 Vegetation indexes	计算公式 Calculation formula
NDVI	(B8-B4)/(B8+B4)
NDVIRE1	(B8A-B5)/(B8A+B5)
NDVIRE2	(B8A-B6)/(B8A+B6)
NDVIRE3	(B8A-B6)/(B8A+B7)

植被指数 Vegetation indexes	计算公式 Calculation formula
NDVI	(B8-B4)/(B8+B4)
NDVIRE1	(B8A-B5)/(B8A+B5)
NDVIRE2	(B8A-B6)/(B8A+B6)
NDVIRE3	(B8A-B6)/(B8A+B7)

项目 Items	5 000Epochs		10 000Epochs
项目 Items	神经元个数 Number of neurons	总体精度 Overall accuracy	神经元个数 Number of neurons	总体精度 Overall accuracy
三个隐藏层 Three hidden layers	64、32、16	0.909	64、32、16	0.934
	128、64、32	0.913	128、64、32	0.946
	256、128、64	0.925	256、128、64	0.941
四个隐藏层 Four hidden layers	128、64、32、16	0.929	128、64、32、16	0.938
	256、128、64、32	0.931	256、128、64、32	0.940
	512、256、128、64	0.933	512、256、128、64	0.943

基于深度学习和GEE的作物遥感分类

Crop remote sensing classification study in Qitai County based on deep learning and Google earth engine (GEE)

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 22

相关文章 7

编辑推荐 0

Metrics

作物种类 Crop types	MLP		RF		SVM
作物种类 Crop types	UA	PA	UA	PA	UA	PA
小麦 Wheat	0.895	0.921	0.793	0.890	0.591	0.753
玉米 Corn	0.946	0.934	0.908	0.899	0.851	0.869
甜菜 Beet	0.963	0.985	0.955	0.977	0.933	0.966
西瓜 Watermelon	0.954	0.980	0.929	0.975	0.811	0.951
其他作物 Other crops	0.901	0.753	0.833	0.595	0.698	0.440

	小麦 Wheat	玉米 Corn	甜菜 Beet	西瓜 Watermelon	其他作物 Other crops	戈壁 Gobi	建筑道路 Building and roads	草地 Grassland
小麦Wheat	257	0	0	0	4	6	6	6
玉米Corn	2	228	0	2	11	1	0	0
甜菜Beet	0	0	259	2	1	0	0	1
西瓜Watermelon	0	0	0	247	5	0	0	0
其他作物Other crops	28	13	10	6	201	4	1	4
戈壁Gobi	0	0	0	2	0	243	6	1
建筑道路 Building and roads	0	0	0	0	0	1	247	0
草地Grassland	0	0	0	0	1	0	0	243

[1]	秦叶康阳, 李嘉欣, 靳瑰丽, 刘文昊, 马建, 李文雄, 陈梦甜. 基于UAV和CNN ResNet 18参数调节的伊犁绢蒿荒漠草地植物识别性能分析[J]. 新疆农业科学, 2024, 61(10): 2547-2556.
[2]	王野;孙东升. 新疆主要农作物比较优势分析——基于国内资源成本法[J]. , 2015, 52(8): 1555-1562.
[3]	孙挺;刘新平;朱爱荣. 农作物秸秆利用的影响因素分析及其效果评价[J]. , 2013, 50(12): 2330-2336.
[4]	房世杰;张淳;任红松. 生物节水专利现状及其对新疆干旱区的启示[J]. , 2011, 48(12): 2343-2347.
[5]	李国萍. 阿勒泰地区土壤有效钼对农作物的影响及对策[J]. , 2008, 45(z1): 204-206.
[6]	张炳坤;仝莉;郭文超. 浅析新疆昌吉州农作物主要有害生物防治存在的问题与对策[J]. , 2007, 44(z2): 158-161.
[7]	李万云;李韬. 激光诱变育种技术的研究与开发应用前景[J]. , 2006, 43(z1): 57-60.

分类算法 Classification algorithm	Overall Accuracy	Kappa
MLP	0.946	0.941
RF	0.911	0.898
SVM	0.842	0.819

分类算法 Classification algorithm	Overall Accuracy	Kappa
MLP	0.946	0.941
RF	0.911	0.898
SVM	0.842	0.819