Please wait a minute...
img

官方微信

遥感技术与应用  2023, Vol. 38 Issue (3): 566-577    DOI: 10.11873/j.issn.1004-0323.2023.3.0566
农业遥感专栏     
基于多时相Sentinel-2影像的棉花雹灾时序变化遥感监测
齐文栋1,郑学昌2,3,何黎明4,卢珍2,3,顾晓鹤5,周艳兵5()
1.北京尚德智汇科技有限公司,北京 100088
2.遥感科学国家重点实验室北京师范大学地理科学学部,北京 100875
3.北京师范大学地理科学学部遥感科学与工程研究院,北京 100875
4.中国农业再保险股份有限公司,北京 100073
5.北京市农林科学院信息技术研究中心,北京 100097
Remote Sensing Monitoring of Temporal Variation in Cotton Hail Disaster based on Multi-temporal Sentinel-2 Image
Wendong QI1,Xüechang ZHENG2,3,Liming HE4,Zhen LU2,3,Xiaohe GU5,Yanbing ZHOU5()
1.Beijing Sun-Golden Technology Company Limited. Beijing 100088,China
2.State Key Laboratory of Remote Sensing Science,Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences,Beijing 100875,China
3.Institute of Remote Sensing Science and Engineering,Faculty of Geographical Sciences,Beijing Normal University,Beijing 100875,China
4.China Agricultural Reinsurance Company Limited. Beijing 100073,China
5.Information Technology Research Center of Beijing Academy of Agricultural and Forestry Sciences,Beijing 100097,China
 全文: PDF(5420 KB)   HTML
摘要:

近年来全球变暖导致强对流天气日益加剧,冰雹灾害已成为农业生产的主要灾害之一。开展棉花冰雹灾情遥感评估对防灾减损、保险理赔、种植结构调整均具有重要意义。以2019年8月23日新疆准噶尔盆地西南部的奎屯河流域的棉花雹灾为研究对象,基于野外实测样本和雹灾前后多时相Sentinel-2遥感影像数据,分析雹灾前后的多种植被指数的动态变化规律,筛选能有效表征雹灾灾情的敏感植被指数差值特征组合,利用逻辑回归、决策树、梯度提升决策树、随机森林4种机器学习算法自动提取棉花雹灾的受灾范围与灾情等级,并利用野外实测样本进行精度对比分析。结果表明:单一植被指数中NDVI对雹灾的指示效果最佳,总体精度为84.39%,Kappa系数为0.75;多时相植被指数差值组合对雹灾的指示性显著优于单一植被指数;结合雹灾前后的植被指数差值时序特征,8月30日与8月20日差值对雹灾的指示性明显强于8月25日与8月20日的差值,说明雹灾灾情等级遥感监测有必要考虑灾后棉花植株的自我恢复能力,待灾情稳定后监测为宜;利用灾前灾后多种植被指数差值组合和随机森林分类算法的棉花雹灾灾情等级监测效果最佳,总体精度达到了89.51%,Kappa系数为0.83。基于多时相Sentinel-2影像能有效评估棉花雹灾的受灾范围以及灾情程度。

关键词: 遥感雹灾机器学习Sentinel?2棉花    
Abstract:

In recent years, global warming has led to an increase in strong convective weather, and hail disaster has become one of the main disasters in agricultural production. Carrying out remote sensing assessment of cotton hail disaster is of great significance for disaster prevention and mitigation, insurance claims and planting structure adjustment. Taking the cotton hail disaster in the Kuitun River Basin in the southwest of Junggar Basin, Xinjiang, on 23 August, 2019 as the research object, with the support of field measured samples, the multi-temporal Sentinel-2 remote sensing images before and after the hail disaster were obtained. We analyzed the dynamic changes of various vegetation indexes before and after the hail disaster, and screened the sensitive vegetation index difference feature combinations which can effectively characterize the hail disaster. The range and grade of cotton hail disaster were automatically extracted by using machine learning algorithms such as logical regression, decision tree, gradient lifting decision tree and random forest, and the accuracy was compared and analyzed via field measured samples. The results showed that NDVI was the best indicator of hail disaster among single vegetation index, with an overall accuracy of 84.39% and a Kappa coefficient of 0.75. The combination of multi-temporal vegetation index differences was significantly more indicative for hail disaster than that of single vegetation index. Compared with the time series characteristics of vegetation index differences before and after hail disaster, the indicative of hail disaster between August 30 and August 20 was obviously stronger than that between August 25 and August 20, which indicated that it was necessary to consider the self-recovery ability of cotton plants after hail disaster grade for remote sensing monitoring. The combination of the pre- and post-disaster vegetation indices and the random forest classification algorithm was the most effective methods in monitoring the level of cotton hail disaster level, with an overall accuracy of 89.51% and a Kappa coefficient of 0.83. In conclusion, the extent and degree of cotton hail disaster can be effectively evaluated based on multi-temporal Sentinel-2 image.

Key words: Remote sensing    Hail disaster    Machine learning    Sentinel-2    Cotton
收稿日期: 2022-09-05 出版日期: 2023-07-11
ZTFLH:  S127  
基金资助: 国家自然科学基金项目(42271319)
通讯作者: 周艳兵     E-mail: zhouyb@nercita.org.cn
作者简介: 齐文栋(1987-),男,山西大同人,工程师,主要从事农业遥感研究及应用研究。E?mail:qiwendong@sun?golden.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
齐文栋
郑学昌
何黎明
卢珍
顾晓鹤
周艳兵

引用本文:

齐文栋,郑学昌,何黎明,卢珍,顾晓鹤,周艳兵. 基于多时相Sentinel-2影像的棉花雹灾时序变化遥感监测[J]. 遥感技术与应用, 2023, 38(3): 566-577.

Wendong QI,Xüechang ZHENG,Liming HE,Zhen LU,Xiaohe GU,Yanbing ZHOU. Remote Sensing Monitoring of Temporal Variation in Cotton Hail Disaster based on Multi-temporal Sentinel-2 Image. Remote Sensing Technology and Application, 2023, 38(3): 566-577.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2023.3.0566        http://www.rsta.ac.cn/CN/Y2023/V38/I3/566

图1  研究区位置图
波段号波段中心波长/nm波段宽度/nm空间分辨率/m
1Aerosols4432060
2Blue4906510
3Green5603510
4Red6653010
5Red edge 17051520
6Red edge 27401520
7Red edge 37832020
8NIR84211510
8aRed edge 48652020
9Water vapor9452060
10Cirrus13753060
11SWIR-116109020
12SWIR-2219018020
表1  Sentinel-2多光谱影像的主要参数
物候播种出苗期苗期蕾期花铃期吐絮期
时期4月中上旬5月中下旬6月中下旬

7月上旬~

8月下旬

9月上旬~10月下旬
表2  研究区棉花种植物候信息
数据源分辨率/m时相物候特征示例
Sentinel-2105月7日

幼苗,

有地膜覆盖

Sentinel-2107月31日

花铃期,

植被信息强

Sentinel-2108月20日

花铃期,

植被信息强

表3  棉花识别选用影像特征
图2  2019年研究区棉花种植及外业调查分布
受灾程度实地照片农技专家鉴定
未受灾叶片、棉铃完好,轻微倒伏,对产量影响较小。
轻度受灾叶片、棉铃有一定程度受损,预计减产20%左右。
中度受灾叶片受损较严重,大量脱落,棉铃部分掉落,预估减产40%以上。
重度受灾叶片受损严重,大量脱落,棉铃掉落严重,预估减产70%以上。
表 4  雹灾等级鉴定样例
植被指数计算公式描述参考文献
NDVINDVI=ρ842-ρ665ρ842+ρ665反映农作物长势和营养信息的重要参数之一[15]
RVIRVI=ρ842ρ665绿色植物的灵敏指示参数,用于检测和估算植物生物量[16]
DVIDVI=ρ842-ρ665对土壤背景的变化极为敏感[17]
EVIEVI=2.5×(ρ842-ρ665)ρ842+6×ρ665-7.5×ρ490+1降低了土壤背景和大气的干扰,可以较客观的反映植被生长状况[18]
NDRENDRE=ρ842-ρ705ρ842+ρ705植被度盖度较大时对植被变化较敏感[19]
表5  遥感植被指数计算公式
特征维计算公式
备注:VI分别代表NDVI、RVI、DVI、EVI、NDRE等遥感指数
遥感时序特征1VI820-VI825
遥感时序特征2VI820-VI830
表6  遥感时序特征计算列表
图3  8月20日、8月25日、8月30日Sentinel-2影像缩略图
图4  各受灾等级样本不同时期光谱反射率曲线图
算法精度指标NDVIRVIDVIEVINDREALL
LRAcc57.14%58.24%58.24%58.24%57.14%58.24%
Kappa0.000 00.294 10.034 90.034 90.000 00.294 1
DTCAcc73.63%62.64%69.23%72.53%68.13%77.93%
Kappa0.572 50.395 30.505 40.555 60.482 90.647 5
RFCAcc73.63%62.64%69.24%72.53%68.15%80.84%
Kappa0.572 50.395 40.505 50.555 60.483 30.694 9
GBDTAcc73.63%62.64%69.23%72.53%68.13%80.22%
Kappa0.572 50.395 30.505 40.555 60.482 90.680 0
表7  基于时序特征1对应的各遥感指标精度
算法精度指标NDVIRVIDVIEVINDREALL
LRAcc59.34%72.53%59.34%61.54%57.14%74.73%
Kappa0.216 40.533 70.086 30.287 00.160 80.574 3
DTCAcc80.22%61.54%68.13%67.03%73.63%81.41%
Kappa0.684 90.395 20.508 20.457 00.595 30.693 7
RFCAcc80.25%61.53%68.14%67.07%73.66%83.54%
Kappa0.685 30.395 10.508 30.457 60.595 90.733 4
GBDTAcc80.22%61.54%68.13%67.03%73.63%87.69%
Kappa0.684 90.395 20.508 20.457 00.595 30.803 3
表8  基于时序特征2对应的各遥感指标精度
算法精度指标NDVIRVIDVIEVINDREALL
LRAcc63.74%76.92%60.44%62.64%61.54%83.52%
Kappa0.322 60.615 40.225 30.327 80.270 30.727 3
DTCAcc83.18%70.23%72.93%71.50%80.17%85.85%
Kappa0.728 30.520 90.552 60.535 20.684 80.770 8
RFCAcc82.05%69.01%72.00%72.03%84.15%89.51%
Kappa0.712 40.505 10.553 80.542 70.747 10.831 2
GBDTAcc84.39%69.31%71.17%72.53%80.51%87.15%
Kappa0.750 40.508 60.535 20.551 00.685 00.795 3
表9  基于时序特征1和2组合对应的各遥感指标精度
验证灾情等级预测灾情等级
未受灾轻度受灾中度受灾重度受灾
未受灾2 1015501490
轻度受灾781 1102120
中度受灾003 53268
重度受灾008519 549
表10  最优模型精度评估混淆矩阵(200次实验)
图5  2019年研究区棉花雹灾等级空间分布图
1 ZHANG Dan, XIN Haihong. Spatial and temporal distribution of hail disasters in Shaanxi from 1950 to 2000[J]. Journal of Shanxi Meteorology, 2014(2): 20-23.
1 张丹, 邢海虹. 1950-2000年陕西冰雹灾害的时空分布规律[J]. 陕西气象, 2014(2): 20-23.
2 YANG Feng, LI Maosong, WANG Yanjiao, et al. Application of remote sensing technique to monitor agricultural disasters[J]. Journal of Natural Disasters, 2006, 15(6): 131-136.
2 闫峰, 李茂松, 王艳姣, 等.遥感技术在农业灾害监测中的应用[J]. 自然灾害学报, 2006, 15(6): 131-136.
3 Li Na, Liu Huanjun, Huang Wenjiang, et al. Hail monitoring and evaluation by remote sensing based on HJ-1A/1B CCD [J]. Transactions of the Chinese Society of Agricultural Engineering, 2011, 27(5): 237-243.
3 李娜, 刘焕军, 黄文江, 等. 基于HJ-1A/1B CCD数据的雹灾监测与评价[J]. 农业工程学报, 2011, 27(5): 237-243.
4 ZHANG Feng. Drought monitoring and temporal and spatial analysis of Zhejiang Province based on MODIS data[D]. Urumqi: Xinjiang Agricultural University, 2009.
4 张峰.基于MODIS数据的浙江省旱情监测及其时空分析[D]. 乌鲁木齐: 新疆农业大学, 2009.
5 QIU Sujuan. The information consultation risk analysis and monitor of crop alien organism[D]. Fuzhou: Fujian Agriculture and Forestry University, 2008.
5 邱素娟. 农作物外来有害生物信息系统及风险评估与监控体系[D]. 福州: 福建农林大学, 2008.
6 Chen Zhongxin, Ren Jianqiang, Tang Huajun, et al. Progress and perspectives on agricultural remote sensing research and applications in China[J]. Journal of Remote Sensing, 2016, 20(5):748-767.
6 陈仲新, 任建强, 唐华俊, 等. 农业遥感研究应用进展与展望[J]. 遥感学报, 2016,20(5):748-767.
7 JIA Dewei, ZHOU Lei, HUANG Canhui,et al.Spatial distribu-tion information extraction of winter wheat hail disaster based on MODIS data: A case study of Pingdingshan City, Henan Province[J].Journal of Areal Research and Development,2018,37(6):134-138.
7 贾德伟,周磊,黄灿辉,等.基于MODIS数据的冬小麦雹灾空间分布信息提取研究——以河南省平顶山市为例[J]. 地域研究与开发,2018,37(6):134-138.
8 LIU Dan, YING Shiping, YU Chenglong, et al. Survery of monitoring hail disaster using HJ-1 A/B moonlet[J]. Journal of Northwest A&F University, 2012, 40(9): 128-132.
8 刘丹, 殷世平, 于成龙, 等. 利用环境减灾卫星遥感监测冰雹灾害初探[J]. 西北农林科技大学学报(自然科学版), 2012, 40(9): 128-132.
9 HU Wen. Hail monitoring of Heilongjiang by remote sensing and analysis of spatial and temporal characteristic[D]. Harbin:Northeast Agricultural University,2015.
9 胡文.黑龙江省雹灾遥感监测及时空特征分析[D].哈尔滨:东北农业大学,2015.
10 LI Hongli. Impact of agricultural meteorological disasters on cotton growth in Xinjiang and preventive measures[J]. Journal of Smart Agriculture. 2022, 2(10): 22-24.李红丽. 新疆农业气象灾害对棉花生长的影响及防范措施[J]. 智慧农业导刊, 2022, 2(10): 22-24.
11 LUAN Yarui. Cause of two hail events in Kuitun-Manas River Basin[C]∥ Proceedings of the 2022 Science and Technology Annual Conference of the Chinese Society of Environmental Sciences (1). 2022: 432-438.
11 栾亚睿. 奎玛流域两次冰雹天气成因分析[C]∥ 中国环境科学学会2022年科学技术年会论文集(一), 2022: 432-438.
12 WANG Xiuzhen, WANG Renchao, HUANG Jingfeng. Derivative spectrum remote sensing and its application in measurement of rice agronomic parameters of rice[J]. Transactions of the Chinese Society of Agricultural Engineering, 2002(1):9-14,13.
12 王秀珍, 王人潮, 黄敬峰. 微分光谱遥感及其在水稻农学参数测定上的应用研究[J]. 农业工程学报, 2002(1):9-14,13.
13 WNAG Xiuzhen, LI Jianlong, TANG Yanlin. Approach the action of derivative spectra for determining agronomic parameters of cotton[J]. Journal of South China Agricultural University,2004,25(2):17-21.
13 王秀珍, 李建龙, 唐延林. 导数光谱在棉花农学参数测定中的作用[J]. 华南农业大学学报,2004,25(2):17-21.
14 HUANG Linsheng, JIANG Jing, HUANG Wenjiang, et al. Wheat yellow rust monitoring method based on Sentinel-2 image and BPNN model[J]. Transactions of the Chinese Society of Agricultural Engineering, 2019, 35(17): 178-185.
14 黄林生, 江静, 黄文江, 等. Sentinel-2影像和BP神经网络结合的小麦条锈病监测方法[J]. 农业工程学报, 2019, 35(17): 178-185.
15 ROUSE J W, HAAS R H, SCHELL J A, et al. Monitoring the vernal advancement and retrogradation (green wave effect) of natural vegetation[R]. Nasa Type III, Final Report. Greenbelt, MD, USA, 1974.
16 PEARSON R L, MILLER L D. Remote mapping of standing crop biomass for estimation of productivity of the shortgrass prairie[J].Remote Sensing of Environment,1972,45(2):7-12.
17 JORDAN C F. Derivation of leaf-area index from quality of light on the forest floor[J]. Ecology, 1969, 50(4): 663-666.
18 HUETE A, JUSTICE C, LEEUWEN W V. MODIS Vegetation Index (MOD13) algorithm theoretical basis document[DS]. NASA Goddard Space Flight Center,Greenbelt,MD, USA, 1999.
19 FITZGERALD G. J. et al. Spectral and thermal sensing for nitrogen and water status in rainfed and irrigated wheat environments[J]. Precision Agriculture, 2006, 7(4): 233-248.
20 WANG Fei, YANG Shengtian, DING Jianli, et al. Environmental sensitive variable optimization and machine learning algorithm using in soil salt prediction at oasis[J]. Transactions of the Chinese Society of Agricultural Engineering, 2018, 34(22):102-110.
20 王飞, 杨胜天, 丁建丽, 等. 环境敏感变量优选及机器学习算法预测绿洲土壤盐分[J]. 农业工程学报, 2018,34(22):102-110.
21 WU Wei, ZHANG Yuan, LI Qiangzi, et al. A hierarchical classification and iterative model based method for remote sensing classification of land cover[J]. Remote Sensing Technology and Application, 2019, 34(1): 68-78.
21 吴薇, 张源, 李强子, 等. 基于迭代CART算法分层分类的土地覆盖遥感分类[J]. 遥感技术与应用, 2019, 34(1): 68-78.
22 JING Xia, ZHANG Jie, WANG Jiaojiao, et al. Comparison of machine learning algorithms for remote sensing monitoring of rice yields[J]. Spectroscopy and Spectral Analysis, 2022, 42(5):1620-1627.
22 竞霞, 张杰, 王娇娇, 等. 水稻产量遥感监测机器学习算法对比[J]. 光谱学与光谱分析, 2022, 42(5):1620-1627.
23 YANG Jinyun, ZHANG Sha, BAI Yun, et al. SPEI simulation for monitoring drought based machine learning integrating multi-source remote sensing data in Shandong[J]. Chinese Journal of Agrometeorology,2021,42(3):230-242.
23 杨晋云, 张莎, 白雲, 等. 基于机器学习融合多源遥感数据模拟SPEI监测山东干旱[J]. 中国农业气象, 2021,42(3):230-242.
24 FANG Miao, ZHANG Jinlong, XU Zhen. Landslide susceptibility zoning study in Lanzhou City based on GIS and Logistic regression model[J]. Remote Sensing Technology and Application, 2011, 26(6): 845-854.
24 方苗, 张金龙, 徐瑱. 基于GIS和Logistic回归模型的兰州市滑坡灾害敏感性区划研究[J]. 遥感技术与应用, 2011, 26(6): 845-854.
25 CHEN Jun, SHEN Runping, LI Bolun, et al. The development of plastic greenhouse index based on Logistic regression analysis[J].Remote Sensing For Land and Resources,2019,3: 43-50.
25 陈俊,沈润平,李博伦,等.基于Logistic回归分析的塑料大棚遥感指数构建[J]. 国土资源遥感,2019,3:43-50.
26 Quinlan J R. C4.5: Programs for machine learning [S]. Morgan, Kauff-man, 1993.
27 CHEN Liping, WU Wenbo. Remote sensing image classification based on object-oriented technique and C 4.5 algorithm[J]. Remote Sensing Information,2013,28(2):116-120.
27 陈丽萍,武文波.基于决策树C4.5算法的面向对象分类方法研究[J]. 遥感信息, 2013, 28(2):116-120.
28 YU Qiao, JIANG shujuan, ZHANG Yanmei, et al. The impact study of class imbalance on performance of software defect prediction models[J]. Chinese Journal of Computers, 2018,41(4):809-824.
28 于巧, 姜淑娟, 张艳梅, 等. 分类不平衡对软件缺陷预测模型性能的影响研究[J].计算机学报, 2018,41(4):809-824.
29 Breiman L. Random forests[J]. Machine Learning, 2001, 45(1): 5-32.
30 ZHOU P C, CHENG G, YAO X W, et al. Machine learning paradigms in high-resolution remote sensing image interpretation[J]. Journal of Remote Sensing,2021,25(1):182-197.
30 周培诚, 程塨, 姚西文, 等. 高分辨率遥感影像解译中的机器学习范式[J]. 遥感学报, 2021,25(1):182-197.
31 Friedman J H. Stochastic gradient boosting[J]. Computational Statistics & Data Analysis. 2002, 38(4): 367-378.
32 YANG Zhichao, ZHAO Sen. A hyperspectral classification method based on 3-D Gabor texture feature and GBDT classifier[J]. Journal of Chengdu University of Technology (Science & Technology Edition). 2021, 48(6): 747-753.杨志超, 赵森. 基于3-D Gabor纹理特征和GBDT分类器的高光谱分类方法[J]. 成都理工大学学报(自然科学版), 2021, 48(6): 747-753.
[1] 黄仲良,何敬,刘刚,李政. 面向GEE平台的遥感影像分析与应用研究进展[J]. 遥感技术与应用, 2023, 38(3): 527-534.
[2] 蒙继华,郑贺刚,王松雪,叶金. 基于卫星遥感预测作物真菌毒素的可行性分析[J]. 遥感技术与应用, 2023, 38(3): 535-543.
[3] 刘胜威,彭代亮,陈俊杰,胡锦康,楼子杭,冯旭祥,程恩惠. 基于Sentinel-2影像的冬小麦收获面积测算[J]. 遥感技术与应用, 2023, 38(3): 544-557.
[4] 齐文栋,何黎明,王安鹏,顾晓鹤,周艳兵. 低温冷害胁迫下的多熟水稻产量遥感监测[J]. 遥感技术与应用, 2023, 38(3): 558-565.
[5] 范沛薇,郝蒙蒙,江东,丁方宇. 基于高分遥感影像的露天矿区土地利用动态监测——以平朔矿区为例[J]. 遥感技术与应用, 2023, 38(2): 274-284.
[6] 米军,张肖,刘良云. 1985~2020年上海市不透水面时空特征演化分析[J]. 遥感技术与应用, 2023, 38(2): 297-307.
[7] 董世英,吴田军,焦思佳. 地块级航空高光谱遥感土地覆盖制图及其精度评估[J]. 遥感技术与应用, 2023, 38(2): 353-361.
[8] 协子昂,张超,冯绍元,张富仓,蔡焕杰,唐敏,孔纪迎. 植被物候遥感监测研究进展[J]. 遥感技术与应用, 2023, 38(1): 1-14.
[9] 赵会芹,于博,陈方,王雷. 基于高分辨率卫星遥感影像滑坡提取方法研究现状[J]. 遥感技术与应用, 2023, 38(1): 108-115.
[10] 邴芳飞,金永涛,张文豪,徐娜,余涛,张丽丽,裴莹莹. 基于机器学习的遥感影像云检测研究进展[J]. 遥感技术与应用, 2023, 38(1): 129-142.
[11] 张妮娜,张珂,李运平,李曦,刘涛. 中国南方典型湿润山区植被类型的无人机多光谱遥感机器学习分类研究[J]. 遥感技术与应用, 2023, 38(1): 163-172.
[12] 刘志君,崔丽娟,李伟,窦志国,左雪燕,雷茵茹,潘旭,李晶,赵欣胜,翟夏杰. 基于高光谱的辽河口盐地碱蓬生态化学计量特征反演研究[J]. 遥感技术与应用, 2023, 38(1): 239-250.
[13] 凌菊,李爱农,靳华安. 基于DART模型和随机森林估算山地叶面积指数[J]. 遥感技术与应用, 2023, 38(1): 39-50.
[14] 张腾,游冬琴,闻建光,唐勇. 二向反射遥感反演最优角度采样方法研究[J]. 遥感技术与应用, 2023, 38(1): 66-77.
[15] 乔树亭,叶回春,黄文江,黄珊瑜,刘荣豪,郭安廷,钱彬祥. 基于Sentinel-1/2影像的水稻种植面积提取方法研究[J]. 遥感技术与应用, 2023, 38(1): 78-89.