Accuracy assessment of the tandem-X DEM in the Northwestern Region of Peninsular Malaysia using GPS-levelling

This study provided a unique accuracy assessment of the TanDEM-X DEM and two other DEMs; SRTM and AW3D30; using three sets of GPS data around northwestern Peninsular Malaysia. Firstly, the absolute vertical accuracy of all DEMs provided by Department Survey and Mapping Malaysia (DSMM) was examined b...

Full description

Bibliographic Details
Published in:ASM Science Journal
Main Author: Pa'suya M.F.; Bakar A.F.A.; Din A.H.M.; Aziz M.A.C.; Samad M.A.A.; Mohamad M.I.
Format: Article
Language:English
Published: Akademi Sains Malaysia 2019
Online Access:https://www.scopus.com/inward/record.uri?eid=2-s2.0-85072225812&partnerID=40&md5=6421caf6cc92c18ed56d21de1e9cb456
Description
Summary:This study provided a unique accuracy assessment of the TanDEM-X DEM and two other DEMs; SRTM and AW3D30; using three sets of GPS data around northwestern Peninsular Malaysia. Firstly, the absolute vertical accuracy of all DEMs provided by Department Survey and Mapping Malaysia (DSMM) was examined by 7755 ground control points (GCPs), which were distributed across the northwestern region. The root-mean-square error (RMSE) of 3.980m and mean error of 2.464m indicated a good agreement of TanDEM-X DEM and GCPs, followed by AW3D30 (RMSE=5.206m) and SRTM (RMSE=6.684m). Secondly, the vertical accuracy assessment of TanDEM-X DEM and the others two DEMs was conducted by comparing the elevation of GPS data around four different land-covers; coastal, hills, foothills and agriculture to study the effects of different land-covers to DEMs accuracy. Again, at the coastal, foothills and agriculture region, TanDEM-X DEM showed the lowest RMSE of 0.66m, 3.8m and 0.41m, respectively. Surprisingly, the accuracy of TanDEM-X DEM at the hill region was lower than those of AW3D30 and SRTM DEM with RMSE of 55.51m. This study also managed to assess the best geoid model to convert the elevation of TanDEM-X DEM to the orthometric height. Three geoid models were tested using local mean sea level at 38 Benchmark around the Perlis region. The results showed that the Malaysia geoid model, MyGEOID, gave the lowest RMSE of 2.277m, followed by EGM2008 (RMSE=2.504m) and EGM96 (RMSE=2.675m). © Academy of Sciences Malaysia 2019.
ISSN:18236782