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  

A New Method to Improve the Accuracy of Digital Elevation Model in Northeast China by Using Terrain,Soil and Crop Information

作　　者：马雨阳 官海翔 杨昊轩 邵帅 邵逸群 刘焕军[1,2] MA YuYang;GUAN HaiXiang;YANG HaoXuan;SHAO Shuai;SHAO YiQun;LIU HuanJun(College of Public Administration and Law,Northeast Agricultural University,Harbin 150030;Northeast Institute of Geography and Agroecology,Chinese Academy of Sciences,Changchun 130012;College of Resources and Environment,Northeast Agricultural University,Harbin 150030)

机构地区：[1]东北农业大学公共管理与法学院,哈尔滨150030 [2]中国科学院东北地理与农业生态研究所,长春130012 [3]东北农业大学资源与环境学院,哈尔滨150030

出　　处：《中国农业科学》2021年第8期1715-1727,共13页Scientia Agricultura Sinica

基　　金：国家自然科学基金(41671438);东北农业大学“学术骨干”项目。

摘　　要：【目的】SRTM DEM是可免费访问公开可用的数字高程模型,但是当前SRTM DEM的垂直精度不能满足精细农业对地形数据的需求,提高其垂直精度,为精准农业等领域提供数据基础。【方法】以黑龙江省海伦东兴农机合作社为研究区,采集实际地面高程数据,获取SPOT-6、Sentinel-2A遥感影像和SRTM DEM。提取归一化湿度指数(NDMI)、归一化植被指数(NDVI)、土壤亮度(TCB)、潜在太阳辐射(PSR)等变量分析地形对其影响关系。利用极限学习机(ELM)和反向传播神经网络(BPNN)提高SRTM DEM水平空间分辨率和垂直精度。使用实际地面高程点进行精度验证,与基于无人机和光学立体像对(ZY-3)生成的DEM进行对比。【结果】SRTM、NDVI、NDMI、TCB与改进后高程的灰色关联度在90%以上,是改进SRTM DEM的重要辅助信息。在整个研究区,BPNN方法的RMSEP为0.98,R^(2) P为0.98,ELM的RMSEP为1.00,R^(2) P为0.90。在平坦区,BPNN方法的RMSEP为0.84,ELM的RMSEP为1.00;在起伏区,BPNN方法的RMSEP为0.99,ELM的RMSEP为0.94。该方法获得的DEM的垂直精度高于ZY-3光学立体像对生成的DEM的垂直精度,为提高SRTM的水平空间分辨率和垂直精度提供了新思路。【结论】引入SRTM、NDVI、NDMI、TCB辅助信息有利于提高SRTM DEM的空间分辨率和垂直精度,获得高精度的DEM。BPNN方法获得的数字高程模型的精度整体上高于ELM方法,BPNN方法更加适用于平坦区高精度DEM的获取,ELM方法更加适用于起伏区。【Objective】SRTM DEM is a publicly available DEM accessible at no cost.However,it is well known that SRTM DEM has a large vertical deviation.In order to improve the accuracy of SRTM DEM in cultivated land,the effects of topography on the temporal and spatial distribution of soil physical and chemical properties and crop growth were analyzed to mine the factors of interaction with topography,so as to obtain a digital elevation model for precision agriculture.【Method】This paper took Helen Dongxing Agricultural Machinery Cooperative in Heilongjiang Province as the study area,the actual ground elevation data were collected,and SPOT-6,Sentinel-2A remote sensing images and SRTM DEM were obtained.The variables,such as normalized differential moisture index (NDMI), normalized difference vegetation index (NDVI), Tasseled Cap Brightness (TCB) and potentialsolar radiation (PSR), were extracted, and the effects of topography on them were analyzed. Extreme Learning Machine (ELM) andback Propagation Neural Network (BPNN) were used to improve horizontal spatial resolution and vertical accuracy of SRTM DEM.The accuracy was verified with the actual ground elevation point and compared with the DEM generated by the UAV and the ZY-3.【Result】The correlation degree between SRTM, NDVI, NDMI, TCB and the improved elevation were more than 90%, which wereimportant factors for improving SRTM DEM. In the whole study area, the RMSEP of BPNN method was 0.98, the RMSEP of R^(2)Pwas 0.98, and the RMSEP of Elm was 1.00, R^(2)P was 0.90. In the flat area, the RMSEP of BPNN method was 0.84 and the RMSEP ofElM was 1.00. In the fluctuation area, the RMSEP of BPNN method was 0.99, and the RMSEP of ELM was 0.94. The verticalaccuracy of DEM obtained by this method was higher than that of DEM generated by ZY-3. It provided a new idea for improving thespatial resolution of SRTM. 【Conclusion】 The introduction of auxiliary information of SRTM, NDVI, NDMI and TCB was helpfulto improve the spatial resolution and vertical accuracy of SRTM DEM,

关 键 词：SRTM DEM 神经网络 多光谱影像 灰色关联分析 方差膨胀因子 

分 类 号：S127[农业科学—农业基础科学]