Font Size: a A A

Analysis And Comparison Of Typical Open Global Digital Elevation Models In Shanxi Province

Posted on:2019-02-24Degree:MasterType:Thesis
Country:ChinaCandidate:W J WuFull Text:PDF
GTID:2310330569979684Subject:Surveying the science and technology
Abstract/Summary:PDF Full Text Request
DEM is the basic data describing ground elevation information,and it is very important to evaluate DEM data accuracy.SRTM DEM and ASTER GDEM data have been widely used since release.AW3D30 is the data released by Japan Aerospace Exploration Agency in 2016 with a resolution of 30 m.SRTM1 DEM,SRTM3 DEM,ASTER GDEM,and AW3D30 as typical open global DEM data sources,obtaining comparative analysis of their accuracy,which has important reference value for the selection of data sources related to the research and application of DEM data.Taking Shanxi Province as the research area,this paper uses ICESat/GLA14 data to compare the absolute vertical accuracy and relative vertical accuracy of the above four kinds of DEM data,and calculates the error distribution of these four kinds of data in different elevation levels and slope grades,and landform types and land use types.The terrain profile and river network were extracted by DEM data,and the topographic expression and hydrological application were compared.Research indicates:(1)Comparison of vertical accuracy: the RMSE of AW3D30,SRTM1 DEM,SRTM3 DEM and ASTER GDEM are 5.0m,6.2m,9.8m and 10.8m respectively.AW3D30 is the best and ASTER GDEM is the worst.With smaller slope values,the vertical accuracy of SRTM1 DEM and SRTM3 DEM is slightly better than that of AW3D30,or equivalent to AW3D30.Low altitudes and slopes,the vertical accuracy of SRTM3 DEM is higher than that of ASTER GDEM,and the ASTER GDEM is slightly better than SRTM3 DEM in areas with high altitude,high slope and undulation.(2)Comparison of relative accuracy: the RMSE of AW3D30,SRTM1 DEM,ASTER GDEM and SRTM3 DEM are 2.523%,3.399%,5.224% and 5.406% respectively,and AW3D30 is the best,SRTM1 DEM is second,and SRTM3 DEM is close to the ASTER GDEM.In the case of low slope,the relative accuracy of SRTM1 DEM and SRTM3 DEM is comparable.The relative accuracy of the SRTM3 DEM is better than that of the ASTER GDEM in the low-altitude,low-slope,plain,and terrace regions,but the ASTER GDEM is slightly superior to the SRTM3 DEM in areas with high altitude and slope and undulation.The AW3D30 has the smallest FSR of 4.37%,SRTM1 of 5.80%,SRTM3 of 8.27%,and ASTER GDEM of 12.15%.The variation rule of FSR of four DEM data with elevation levels and landform types is contrary to vertical error and relative error.It shows that the ability of DEM data recognition slope direction of points is lower in areas with low elevations and flat terrain than in areas with rough terrain.(3)Profile comparison: ASTER GDEM and AW3D30 have more abnormal values in the plain and platform areas,and ASTER GDEM is significantly over-expressed in the hilly regions.In undulating areas,SRTM3 DEM has more overestimation of valleys,and SRTM1 DEM and AW3D30 data are more accurate.(4)River network comparison: the SRTM1 DEM has the smallest river network overlay difference of 1.35%,SRTM3 DEM of 1.36%,AW3D30 of 1.56%,and ASTER GDEM of 2.75%.Therefore,in general,the accuracy of the SRTM1 DEM extraction river network is the best,and ASTER GDEM is the worst.However,in the flat area,the accuracy of SRTM1 DEM and SRTM3 DEM is better than that of AW3D30 and ASTER GDEM in river network extraction.In the region with large topographic relief,AW3D30 has the best extraction accuracy,and ASTER GDEM is the worst.The accuracy of four typical open DEM datasets was analyzed from two aspects of absolute vertical error and relative vertical error.The error distribution in different elevations,slopes,landforms,land use types,and accuracy in terrain profile extraction and water network extraction applications was compared.These results contribute to the selection of DEM data sources.
Keywords/Search Tags:DEM, ICESat/GLA14, vertical accuracy, relative accuracy, Shanxi Province
PDF Full Text Request
Related items