Please wait a minute...
img

官方微信

遥感技术与应用  2021, Vol. 36 Issue (5): 1083-1091    DOI: 10.11873/j.issn.1004-0323.2021.5.1083
数据与图像处理     
引入NDVI改进SRTM DEM的新方法
刘焕军2,3(),马雨阳1,杨昊轩2,姜芸2(),巩超4,吕航4
1. 中国地质大学(武汉)地理与信息工程学院,湖北 武汉 430074
2. 东北农业大学公共管理与法学院,黑龙江 哈尔滨 150030
3. 中国科学院东北地理与农业生态研究所,吉林 长春 130012
4. 吉林省土壤肥料总站,吉林 长春 130033
A New Method of Introducing NDVI to Improve SRTM DEM
Huanjun Liu2,3(),Yuyang Ma1,Haoxuan Yang2,Yun Jiang2(),Chao Gong4,Lü Hang4
1. School of Geography and Information Engineering,China University of Geosciences,Wuhan 430073,China
2. College of Public Administration and law,Northeast Agricultural University,Harbin 150030,China
3. Northeast Institute of Geography and Agricultural Ecology,Chinese Academy of Sciences,Changchun 130012,China
4. Jilin Provincial soil and Fertilizer Station,Changchun 130033,China
 全文: PDF(2542 KB)   HTML
摘要:

精准/智慧农业需要田块尺度高精度地形数据,而现有耕地范围地形测绘数据不能满足需求。为获得高精度数字高程模型(DEM),以SRTM DEM为基础进行改进,获取2016年6~9月SPOT 6多光谱数据,测量实际高程,将实际高程通过克里金空间插值获得分辨率为6 m的DEM;将SRTM DEM格网的栅格值和植物生长季节的归一化植被指数(Normalized Difference Vegetation Index,NDVI)作为输入量,建立逐步多元回归和BP神经网络模型,以实测的地面高程和无人机获取的DEM数据作为验证,与空间插值方法和资源三号获取DEM进行对比。结果表明:①引入生长季节NDVI的线性回归模型的精度达到96.0%,RMSE为1.12;BP神经网络模型精度达到98.7%,RMSE降为0.86;②生长季节NDVI的时空变化是坡度、坡位等地形因子作用的结果;③利用BP神经网络改进的SRTM DEM获得更高空间分辨率的DEM精度高于资源三号生成的DEM,与无人机DEM的空间趋势相似。可为田间变量管理、精准管理分区、土壤分类与精细制图等提供支持。

关键词: SRTM DEMBP神经网络NDVI多元线性回归克里金插值    
Abstract:

Precision agriculture and intelligent agriculture need high precision terrain factor data of field scale, but the existing topographic mapping data of arable land can not meet the demand. In order to establish a high spatial resolution Digital Elevation Model (DEM), obtained SPOT-6 multispectral data in June, July, August and September 2016, SRTM DEM and the actual elevation of the study area were measured. High resolution DEM(6 m) data are obtained by Kriging spatial interpolation based on the measured elevation. Taking the grid value of SRTM DEM and Normalized Difference Vegetation Index (NDVI) as inputs, the multiple linear regression and BP neural network reconstruction model are established, The verification is based on the measured ground elevation and the DEM data obtained by UAV. Compared with the UAV DEM and ZY-3 DEM. The results show that: ①The accuracy of the linear regression model introduced into NDVI time series is 96%. The RMSE is 1.12; The accuracy of BP neural network model is as high as 98.7%, and RMSE is reduced to 0.86. ②The temporal and spatial variation of NDVI in growing season is the result of topographic factors such as slope and slope position. ③The improved SRTM DEM based on BP neural network achieves higher spatial resolution, which is similar to the spatial trend of UAV DEM, and its accuracy is higher than that of ZY-3 DEM. It can provide data support for field variable management, precise management zoning, soil classification and fine mapping.

Key words: SRTM DEM    BP neural network    NDVI    Multiple linear regression    Kriging interpolation
收稿日期: 2021-06-23 出版日期: 2021-12-07
ZTFLH:  TP75  
基金资助: 国家自然科学基金项目(41671438);东北农业大学“学术骨干”项目
通讯作者: 姜芸     E-mail: huanjunliu@yeah.net;remotesense@163.com
作者简介: 刘焕军(1981-),男,黑龙江牡丹江人,教授,主要从事宵业遥感、地理信息系统等方面的研究。E?mail:huanjunliu@yeah.net
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  
刘焕军
马雨阳
杨昊轩
姜芸
巩超
吕航

引用本文:

刘焕军,马雨阳,杨昊轩,姜芸,巩超,吕航. 引入NDVI改进SRTM DEM的新方法[J]. 遥感技术与应用, 2021, 36(5): 1083-1091.

Huanjun Liu,Yuyang Ma,Haoxuan Yang,Yun Jiang,Chao Gong,Lü Hang. A New Method of Introducing NDVI to Improve SRTM DEM. Remote Sensing Technology and Application, 2021, 36(5): 1083-1091.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2021.5.1083        http://www.rsta.ac.cn/CN/Y2021/V36/I5/1083

图1  研究区位置与采样点分布图
图2  6~9月NDVI空间分布
模型构建 精度验证
模型类型 模型公式 决定系数(R 2) 均方根误差(RMSE)/m
克里金插值 实测点高程值空间内插 0.94 1.15
多元回归 Y=-5.25*6月NDVI+8.09*8月NDVI+0.94*SRTM DEM+10.22 0.96 1.12
BP神经网络 模型输入量:SRTM DEM(6)、6月、7月、8月、9月NDVI 0.98 0.86
SRTM DEM(6) 14.82
SRTM DEM(30) 14.77
资源三号DEM 3.69
无人机DEM 0.10
表1  高精度DEM模型及精度评价
图3  研究区坡度、坡向与太阳辐射
图4  研究区坡位分布图
图5  不同坡位上的NDVI在不同月份的分布特征
图6  不同月份的NDVI极值在不同坡位的分布特征
图7  研究区全年气温降雨量
图8  DEM空间格局
1 Kulp S A , Strauss B H . Coastal DEM: a global coastal digital elevation model improved from SRTM using a neural network[J]. Remote Sensing of Environment 2018,206:231-239. DOI:10.1016/j.rse.2017.12.026 .
doi: 10.1016/j.rse.2017.12.026
2 Freeman A .Calibration, synthetic aperture radars[A]. Njoku E G. Encyclopedia of remote sensing, encyclopedia of earth Sciences[M]. New York:Springer New York, 2014. .
3 Li Zhenhong , Li Peng , Ding Dong ,et al . Research progress of global high resolution digital elevation models[J]. Geomatics and Information Science of Wuhan University, 2018, 43(12): 1927-1942.
3 李振洪,李鹏,丁咚,等 . 全球高分辨率数字高程模型研究进展与展望[J]. 武汉大学学报(信息科学版),2018,43(12):1927-1942.
4 Wang Jiayao , Cui Tiejun , Miao Guoqiang . Digital elevation model and data structure[J]. Hydrographic Surveying and Charting, 2004,24(3): 1-4.
4 王家耀,崔铁军,苗国强 . 数字高程模型及其数据结构[J]. 海洋测绘,2004,24(3):1-4.
5 Hu Peng , Yang Chuanyong , Wu Yanlan ,et al . New methods, principles and applications of digital elevation model[M]. Beijing: Surveyingand Mapping Press, 2007: 56-58.
5 胡鹏,杨传勇,吴艳兰,等 . 新数字高程模型理论、方法、标准和应用[M]. 北京:测绘出版社, 2007:56-58.
6 Gobakken T , Bollandsas O M , Næsset E . Comparing biophysical forest characteristics estimated from photogrammetric matching of aerial images and airborne laser scanning data[J].Scandinavian Journal of Forest Research, 2015, 30(1): 73-86. DOI:10.1080/02827581.2014.961954 .
doi: 10.1080/02827581.2014.961954
7 Błaszczak-Bąk W , Janowski A , Kaminski W , et al . Application of the msplit method for filtering airborne laser scanning data-sets to estimate digital terrain models[J]. International Journal of Remote Sensing, 2015,36(9): 2421-2437.
8 Chen X , Liu Y , Tang X , et al . Some critical issues on airborne LiDAR mapping software in coastal survey[J]. International Society for Optics and Photonics, 2009, 7492:74920F. DOI:10.1117/12.838623 .
doi: 10.1117/12.838623
9 Nnesset E . Vertical height errors in digitalterrain models derived from airborne laser scanner data in a boreal-alpine ecotone in norway[J]. Remote Sensing, 2015,7(4):4702-4725. DOI:10.3390/rs70404702 .
doi: 10.3390/rs70404702
10 Hansen E H , Gobakken T , Naesset E . Effects of pulse density on digital terrain models and canopy metrics using airborne laser scanning in a tropical rainforest[J]. Remote Sensing, 2015, 7(7) : 8453-8468. DOI:10.3390/rs70708453 .
doi: 10.3390/rs70708453
11 Bhardwaj A , Jain K , Chatterjee R S . Generation of high-quality digital elevation models by as similation of remote sensing-based DEMs[J]. Journal of Applied Remote Sensing 2019, 13. DOI:10.1117/1.JRS.13.4.044502 .
doi: 10.1117/1.JRS.13.4.044502
12 Ma Y Y , Liu H J , Jiang B W , et al . An innovative approach for improving the accuracy of digital elevation models for cultivated land[J]. Remote Sensing,2020,12(20):DOI:org/10.3390/rs12203401.
13 Wendi D , Liong S Y , Sun Y B ,et al . An innovative approach to improve SRTM DEM using multispectral imagery and artificialneural network[J]. Journal of Advances in Modeling Earth Systems,2016,8:691-702. DOI:10.1002/2015MS000536 .
doi: 10.1002/2015MS000536
14 Ajibola I I , Mansor S , Pradhan B , et al . Fusion of UAV-based DEMs for vertical component accuracy improvement[J]. Measurement,2019,147:106795. DOI:10.1016/j.measurement.2019.07.023 .
doi: 10.1016/j.measurement.2019.07.023
15 Perry M A , Niemann J D . Analysis and estimation of soil moisture at the catchment scale using EOFs - ScienceDirect[J]. Journal of Hydrology, 2007,334(3–4):388-404. DOI:10.1016/j.jhydrol.2006.10.014 .
doi: 10.1016/j.jhydrol.2006.10.014
16 Chang Chun , Wang Xinyuan , Yang Ruixia , et al . A quantitative characterization method for alpine vegetation zone based on DEM and NDVI[J]. Geographical Research,2015,34(11): 2113-2123.
16 常纯,王心源,杨瑞霞,等 . 基于DEM-NDVI的高山植被带定量刻划[J]. 地理研究,2015,34(11):2113-2123.
17 Wei Wei , Zhao Jun , Wang Xufeng ,et al . Studyon the spatial relationships between NDVI, DEM and vegetation cover in the Tianzhu alpine grasslands based on TM/ETM[J]. Arid area research, 2008,25(3):394-401.
17 魏伟,赵军,王旭峰,等 . 天祝高寒草原区NDVI,DEM与地表覆盖的空间关系[J].干旱区研究,2008,25(3):394-401.
18 Hu Yunfeng , Xu Zhiying . Characteristic scales and the multi-scale correlation between DEM and NDVI factors: a case study in the mongolian plateau[J]. Geography Science, 2014, 34(12): 1511-1517.
18 胡云锋,徐芝英 . 蒙古高原地形与植被指数的特征尺度及多尺度耦合关系. 地理科学[J],2014,34(12):1511-1517.
19 Tokuchi N , Takeda H , Yoshida K , et al . Tpographical variations in a plant-soil system along aslope on mt ryuoh, japan[J]. Ecological Research, 1999, 14: 361-369.
20 Lewis K J , Trummer L M , Thompson R D . Incidence of tomentosus root disease relative to spruce density and slope position in south-central alaska[J]. Forest Ecology & Management,2004,194(1-3):159-167. DOI:10.1016/j.foreco.2004. 02.027 .
doi: 10.1016/j.foreco.2004. 02.027
21 Ji Xinyang , Luo Lie , Wang Xinyuan , et al . Identification and change analysis of mountainaltitudinalzone based on DEM-NDVI land cover classification in Tianshan bogda natural heritage site[J]. Journal of Geo-information Science, 2018, 20(9): 1350-1360.
21 冀欣阳,骆磊,王心源,等 . 基于“DEM-NDVI-土地覆盖分类”的天山博格达自然遗产地山地垂直带提取与变化分析[J].地球信息科学学报,2018,20(9):1350-1360.
22 Gao Zhiqiang , Liu Jiyuan . The study on driving factors and models of NDVI change based on remote sensing and GIS in China[J]. Climatic and Environmental Research,2000,5(2): 155-164.
22 高志强,刘纪远 . 基于遥感和GIS的中国植被指数变化的驱动因子分析及模型研究[J]. 气候与环境研究,2000,5(2):155-164.
23 Cantón Y , Del Barrio G , Sole-Benet A , et al . Topographic controls on the spatial distribution of ground cover in the tabernas badlands of SE spain[J]. Catena,2004,55(3):341-365.
24 Ma Yuyang , Guan Haixiang , Yang Haoxuan , et al . A new method to improve the accuracy of digital elevation model in northeast china by using terrain, soil and crop information[J]. Scientia Agricultura Sinica,2021,54(8):1715-1727.
24 马雨阳,官海翔,杨昊轩,等 . 利用地形、土壤和作物信息辅助提高东北漫岗地数字高程模型精度的新方法[J]. 中国农业科学,2021,54(8):1715-1727.
25 Tang Guoan , Yang Xin . ArcGIS experimental course of spatial analysis of GIS[M]. Beijing: Science Press,2012.
25 汤国安,杨昕 . ArcGIS地理信息系统空间分析实验教程[M]. 北京:科学出版社,2012.
26 Zhong W , Deng Y C , Machado J A , et al . Strength prediction of similar materials to ionic rare earthores based on orthogonal test and back propagation neural network[J]. Soft Computing,2019,23: 9429-9437. DOI:10.1007/s00500-019-03833-7 .
doi: 10.1007/s00500-019-03833-7
27 Pao Y H . Adaptive pattern recognition and netural networks, reading[J].MA:Addison-Wesley,1989. DOI:10.1007/s00500-019-03833-7 .
doi: 10.1007/s00500-019-03833-7
28 Peretto P . An introduction to the modeling of neural, cambridge[M]. Cambridge University Press, 1992.
29 Ahmad M , Salam F . Error back-propagation learning using polynomial energy function[C]∥ IEEE International Conference on Systems Engineering. IEEE, 1992.
30 Vogl T P , Mangis J K , Rigler A K , et al . Accelerating the convergence of the back-propagation method,1988,59:257-263.DOI:10.1007/BF00332914
doi: 10.1007/BF00332914
31 Dou X , Wang X , Liu, H J, et al . Prediction of soil organic matter using multi-temporal satellite images in the Songnen plain,China[J]. Geoderma,2019,365:113896. DOI:10.1016/j.geoderma.2019.113896 .
doi: 10.1016/j.geoderma.2019.113896
32 Ghandehari M , Buttenfield B P , Farmer C J Q . Comparing the accuracy of estimated terrain elevations across spatial resolution[J]. International Journal of Remote Sensing, 2019, 40(13): 5025-5049. DOI:10.1080/01431161.2019.1577581 .
doi: 10.1080/01431161.2019.1577581
33 Pan Hongbo , Zhang Guo , Tang Xinming . Accuracy analysis and verification of ZY-3 products[J]. Journal of Surveying and Mapping,2013,42(5):738-744,751.
33 潘红播,张过,唐新明,等 .资源三号测绘卫星影像产品精度分析与验证[J].测绘学报,2013,42(5):738-744,751.
[1] 李毅,任超,张志刚,梁月吉,潘亚龙. 基于多元线性回归的GPS-IR积雪深度反演研究[J]. 遥感技术与应用, 2020, 35(6): 1312-1319.
[2] 李汝嫣,颉耀文,姜转芳. 面向对象的天然绿洲与人工绿洲区分[J]. 遥感技术与应用, 2020, 35(4): 873-881.
[3] 焦雪敏,张赫林,徐富宝,王岩,彭代亮,李存军,徐希燕,范海生,黄运新. 青藏高原1982~2015年FPAR时空变化分析[J]. 遥感技术与应用, 2020, 35(4): 950-961.
[4] 岳昂,曾庆伟,王怀警. 于桥水库蓝藻水华遥感长时序监测研究[J]. 遥感技术与应用, 2020, 35(3): 694-701.
[5] 纪甫江,蒙继华,方慧婷. 基于CASA-WOFOST耦合模型的大豆单产遥感估算研究[J]. 遥感技术与应用, 2020, 35(2): 406-415.
[6] 刘俊,孟庆岩,葛小三,刘顺喜,陈旭,孙云晓. 基于BP神经网络的夏玉米多生育期叶面积指数反演研究[J]. 遥感技术与应用, 2020, 35(1): 174-184.
[7] 李娜,张升伟,何杰颖. 基于FY-3C MWHTS的台风降水反演算法研究[J]. 遥感技术与应用, 2019, 34(5): 1091-1100.
[8] 李嘉玲, 董东林, 林刚, 汪箫悦, 王健, 吴朝阳. 基于NDVI数据的江苏省植被物候变化及其影响因子分析[J]. 遥感技术与应用, 2019, 34(2): 367-376.
[9] 汪子豪, 秦其明, 孙元亨.

基于BP神经网络的地表温度空间降尺度方法 [J]. 遥感技术与应用, 2018, 33(5): 793-802.

[10] 茹皮亚·西拉尔, 杨辽. 新疆棉花物候时空变化遥感监测及气温影响分析[J]. 遥感技术与应用, 2018, 33(5): 923-931.
[11] 张滔,唐宏. 基于Google Earth Engine的京津冀2001~2015年植被覆盖变化与城镇扩张研究[J]. 遥感技术与应用, 2018, 33(4): 593-599.
[12] 汪航,师茁. 基于MODIS时间序列数据的春尺蠖虫害遥感监测方法研究—以新疆巴楚胡杨为例[J]. 遥感技术与应用, 2018, 33(4): 686-695.
[13] 苗茜,王昭生,王荣,黄玫,孙佳丽. 基于NDVI数据评估O3污染对华北地区夏季植被生长的影响[J]. 遥感技术与应用, 2018, 33(4): 696-702.
[14] 周玉科,刘建文. 基于MODIS NDVI和多方法的青藏高原植被物候时空特征分析[J]. 遥感技术与应用, 2018, 33(3): 486-498.
[15] 杨朦朦,汪汇兵,欧阳斯达,范奎奎,戚凯丽. 基于双树复小波分解的BP神经网络遥感影像分类[J]. 遥感技术与应用, 2018, 33(2): 313-320.