ASSESSMENT OF SRTM, ACE2 AND ASTER-GDEM USING RTK-GPS

Transcription

1 ASSESSMENT OF SRTM, ACE2 AND ASTER-GDEM USING RTK-GPS Hsing-Chung Chang, Xiaojing Li, Linlin Ge School of Surveying and Spatial Information Systems The University of New South Wales, Sydney, NSW 2052, Australia Phone: , Fax: Abstract To date, the majority of digital elevation models (DEMs) are generated using photogrammetric methods. Recent remote sensing techniques, such as radar interferometry (InSAR), have also been used to generate high quality DEMs over a great area. Spaceborne high resolution DEMs with almost global coverage, such as SRTM, ACE2 and ASTER-GDEM, were made available to the public in the last few years. This paper aims to assess the quality of global DEMs using real-time kinematic GPS (RTK-GPS) data, together with airborne laser scanner (ALS) data. The quality of the DEMs for coastal region in Sydney was also compared. 1. Introduction A digital elevation model (DEM) is a digital cartographic representation of the elevation of earth surface. The applications of DEMs include flood risk assessment, landslides modelling, flight planning, rectification of remote sensing imagery, urban planning, military uses, to name only a few. High quality DEMs are also essential for measuring land deformations. To date, the majority of DEMs are generated using photogrammetric methods. Besides, new remote sensing techniques, such as radar interferometry (InSAR), (Zebker and Goldstein 1986; Ackermann 1999; Schiewe 2005) can also be used to generate high quality DEMs. Photogrammetry is a passive system which detects reflected solar radiation from ground surface and records the returns digitally or on films. Unlike photogrammetry, radar is an active system that equips its own energy source for illuminating the land. Differential InSAR (DInSAR) has been used for monitoring or mapping the ground deformation due to earthquakes, volcanic activities, underground mining, etc. In DInSAR processing, the knowledge of local topography is important in order to separate the interferometric phase signals induced by the ground movement from topography. The detectable phase signal of land surface displacement is dependent on the accuracy of the DEM used (Nolan and Fatland 2003). The errors in the DEM will propagate into the DInSAR results. Therefore, the quality of DEM used in DInSAR processing is important. Conventionally, high resolution DEMs were only available at regional scales. Then there were some high resolution global DEMs becoming more public 1

2 accessible in the past few years. This paper aims to assess the quality of global DEMs of ACE2, ASTER and SRTM, using real-time kinematic GPS (RTK-GPS), together with airborne laser scanner (ALS) data. 2. Global Digital Elevation Models Global coverage DEMs were made available in coarse posting resolutions since For examples, ETOPO5, TerrainBase and JGP95E DEMs (Arabelos 2000) have at a pixel size of 5 arc-minute, and the GTOPO30 (USGS 2009) and GLOBE (Hastings and Dunbar 1998) DEMs have at a resolution of 30 arcsecond. Although these DEMs provided the basic topographic information, the scientific studies related to land surface at regional scale was constrained by the coarse resolution of the DEMs. Finally, the global DEMs in high resolutions, i.e. less than 100m, were made available to the science community and general public users in the last decade, such as shuttle radar topography mission (SRTM), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), and altimetry DEMs. 2.1 SRTM Synthetic aperture radar (SAR) measures both amplitude and phase information of the back-scattered radar signals from imaging pixels on the Earth s surface. Radar interferometry (InSAR) exploits the phase information recorded in two SAR images to derive geodetic information of terrain. The phase difference between two SAR images acquired at slightly different satellite positions contains the information of topography with respective to the SAR imaging geometry and wavelengths. InSAR has demonstrated its capability of generating DEMs (Zebker and Goldstein 1986; Rosen et al. 2000). An 11 day space Shuttle Radar Topography Mission (SRTM) was successfully flown in February This mission used InSAR with C- (5.6cm) and X- (3cm) bands of the microwave to create the first DEM of entire earth between the latitude ranging from 60 N to 57 S. SRTM used two antennas, separated 60 m apart, to image the earth s surface instantaneously (Rabus et al. 2003). The C- band antenna has an imaging swath width of 225 km while the X-band antenna was only limited to a swath of 45 km. Therefore the coverage of X-band is limited and does not provide in global coverage. As the fact that X-band wavelength cannot penetrate the vegetation and C-band wavelength will be reflected at the top of the canopies, the elevation measured by SRTM is also referred to as a digital surface model (DSM) which represents the height of the ground surface objects including vegetations. The C-band SRTM DEM data can be downloaded from the USGS EROS Data center. By combining the data from both ascending and descending orbits, the topography data with a post spacing of 1 arc-second (approximately 30m) is released for the coverage within United States and 3 arc-second (approximately 90m) resolution data are available elsewhere. The C-band SRTM data for Australia, New Zealand and South Pacific Islands were released in October

3 2.2 ASTER-GDEM Since 1986, the SPOT satellites (SPOT-1 to 4) have provided cross-track satellite optical stereo images for generating DEMs. However, due to the difficulty of obtaining cloud free cross-track stereo pairs, the work of DEM generation for a global coverage is very challenging. Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), was designed to take along-track stereo image pairs. The sensors pointing nadir and backward to acquire the along-track stereo pair increased the chances of getting cloud free image pairs. ASTER, built by the Ministry of Economy, Trade and Industry (METI) of Japan, onboard the NASA Terra satellite was launched on 18 December 1999 (Hirano et al. 2003; ASTER GDEM Team 2009). Then SPOT-5, which was launched on 4 May 2002, also equipped the forward and backward looking sensors to acquire the along-track stereo pairs. The ASTER global digital elevation model (GDEM) was released by METI and NASA on 29 June ASTER GDEM covers the land surfaces between 83 0 N and 83 0 S with estimated accuracies of 20m and 30m at 95% confidence for vertical and horizontal data, respectively (ASTER GDEM Team 2009). The DEM is delivered in 1 x 1 degree tiles as GeoTIFF files with a resolution of 30m. 2.3 ACE2 ACE2 DEMs were created by synergistically merging the SRTM data with satellite radar altimetry. The altimeter heights generated from the ERS-1 Geodetic Mission were used in creation of the ACE Global DEM (ACE GDEM 2000). When the SRTM DEM was made available, the heights generated using the InSAR technique in C-band were validated at a global scale using the altimeter heights (Berry et al. 2007). The signal of altimeter is able to penetrate dense canopy and reach to the underlying terrain. That has been proved by comparing the altimetry data acquired by ERS-1 Geodetic Mission over the forest area in Amazon (Berry et al. 2007). Satellite altimetry was initially designed to measure the ocean surface as it is easier to measure the radar return signals from open oceans. The averaged sea level and ocean circulation variations have been better understood using satellite altimetry (Andersen and Knudsen 1996; Fu and Le Traon 2006). The ERS-1 mission collected large volumes of echoes over both ocean and land surfaces. The highly complex altimeter echoes were made useful to derive the geodetic information of land after re-tracking the waveforms (Berry et al. 1998). 3

4 3. Height Reference Data 3.1 RTK-GPS Real-time kinematic (RTK) GPS deliver instantaneous point coordinates with centimetre-level accuracy. There are many applications that take advantage of RTK technology, including topographic surveying, engineering construction, geodetic control, vehicle guidance and automation, etc. (Riley et al. 2000). RTK positioning uses a static GPS receiver as a reference station located at a known point. Another receiver is used as the rover which can move and survey any point of interest within a limited range from the reference station. Both receivers make observations of the GPS signals at the same time and a radio data link between the two receivers permits data to be sent from the reference to rover, where the calculation of coordinates is carried out. For the field survey at the test site, Appin, the reference receiver station was set up on top of a hill on a trig station. The rover receiver was mounted on a car roof so that the survey could be easily conducted by driving around the region within the radio link coverage, which is approximately 8km from the reference station in our set-up. 3.2 Airborne Laser Scanner Airborne laser scanner (ALS), or so-called light detection and ranging (LiDAR), uses a laser scanner mounted on an airplane to sample the height profiles of land. In general, ALS can measure several points per square meter, at vertical accuracy ranging from m (Skaloud et al. 2010), which are subject to the slope of terrain and density of vegetation cover. In this study, the ALS data acquired by ALS50 sensor of NSW Land and Property Management Authority (LPMA) is used as a reference. The average laser point density is 0.75 point per square metre. 4. Results The posting resolutions of ACE2, ASTER and SRTM DEMs are about 90m, 30m and 90m, respectively. There were 144 RTK-GPS points selected at intervals equal to or greater than 100m to assess these remote sensing global DEMs. The locations of the GPS points are overlaid on an aerial photograph and shown in Figure 1 (a). The heights of ACE2, ASTER and SRTM are shown in Figure 1 (b) (d). Despite the 30m grid size of ASTER DEM, there is a height error that can be visually identified across Nepean river. The heights of the DEMs were extracted based on the locations of GPS points with the aid of GIS. The height differences between the DEMs and RTK-GPS are shown in Figure 2. It can be noticed that elevations in both ACE2 and ASTER DEMs are underestimated while the elevations in SRTM DEM tends to be higher than the GPS data. The root-mean-square (RMS) errors for ACE2, ASTER and SRTM DEMs are 3.4m, 6.5m and 3.0m, respectively. The errors of ALS DEM are also given in the table for illustration purpose only as the height 4

5 accuracy of RTK-GPS may not be adequate to assess the quality of ALS DEMs. The summary of the DEM height errors is given in Table 1. The heights of DEMs over bushland were assessed using ALS data. The selected region of interest (ROI) for assessment is indicated by a red rectangle shown in Figure 3. The results showed a increase in RMS errors in all DEMs due to the heights of vegetation cover as shown in Table 2. The ACE2 DEM illustrated less RMS error than the SRTM DEM. This may be due to the height correction using the altimetry data as discussed earlier in this paper. Finally, the DEMs near the coastal area of Sydney were also compared. Figure 4 showed the subsets of DEMs in Sydney. ACE2 and ASTER DEMs illustrated less coverage along the coastline, while SRTM provides the best coverage. It may suggest that the mosaic between SRTM and altimetry data (i.e. ACE2 DEM) does not necessarily improve the quality of DEM for coastal regions. (a) (b) (c) Figure 1. (a) Aerial photograph of the test site, Appin, overlaid by the RTK GPS survey points in red, and the DEMs of the same region derived from (b) ACE2, (c) ASTER and (d) SRTM, respectively. (d) 5

6 Figure 2. Height differences between the 144 selected RTK-GPS points and the DEMs of ACE2, ASTER, SRTM and ALS. Table 1. Height errors of the selected remote sensing DEMs against RTK-GPS data. ACE2 ASTER SRTM ALS RMSE (m) mean (m) s.d. (m) Table 2. Height errors of the DEMs of ACE2, ASTER and SRTM against the one of ALS over bushland. ACE2 ASTER SRTM RMSE (m) mean (m) s.d. (m)

7 Figure 3. Heights of global DEMs over bushland were extracted from the region of interest (ROI) coloured in red and compared against the ALS data. (a) (b) (c) Figure 4. (a) Landsat-5 true colour image of Sydney and the subsets of DEMs of (b) ACE2, (c) ASTER and (d) SRTM. (d) 7

8 Concluding Remarks The quality of global digital elevation models of SRTM, ACE2 and ASTER were assessed using real-time kinematic GPS data. Among the three DEMs, SRTM has the smaller RMS error of 3.0m in vertical against GPS. Also, airborne laser scanner data were used to assess the quality of the global DEMs over a selected bushland. The results showed that ACE2 has the lowest RMS error of 8.5m than the SRTM and ASTER DEMs. So for the applications of DInSAR over a vegetated region, the selection between using ACE2 and SRTM DEMs as the reference of topography is important in order to minimise the topographic residual errors. Nevertheless, ASTER-GDEM has greater global coverage than SRTM or ACE2 DEMs, for the lands between N or S. For coastal regions, SRTM gave the best representation along the coastline in Sydney. Therefore, SRTM DEM may be more suitable for the applications near the coastal region. Acknowledgement This research is jointly supported by Cooperative Research Centre for Spatial Information (CRCSI), The University of New South Wales (UNSW) and NSW Land and Property Management Authority (LPMA). Authors would like to thank LPMA for providing the high resolution aerial photograph and ALS DEM of the test area. References ACE GDEM ACE (Altimeter Corrected Elevations) DEM. [Last access: 31 July 2010], at Ackermann, F., Airborne laser scanning--present status and future expectations. ISPRS Journal of Photogrammetry and Remote Sensing 54(2-3): pp Andersen, O. B. and P. Knudsen, Altimetric gravity field from the full ERS-1 geodetic mission. Physics and Chemistry of The Earth 21(4): pp Arabelos, D., Intercomparisons of the global DTMs ETOPO5, TerrainBase and JGP95E. Physics and Chemistry of the Earth, Part A: Solid Earth and Geodesy 25(1): pp ASTER GDEM Team, ASTER Global DEM Validation Summary Report. June Berry, P., R. Sanders, C. Leenmans and E. Bron, Generating orthometric heights from the ERS-1 altimeter geodetic mission dataset: results from an expert systems approach. Geodesy on the move: gravity, geoid, geodynamics and Antarctica: IAG Scientific Assembly, Rio de Janeiro, Brazil, September 3-9, : pp Berry, P. A. M., J. D. Garlick and R. G. Smith, Near-global validation of the SRTM DEM using satellite radar altimetry. Remote Sensing of Environment 106(1): pp

9 Fu, L. L. and P.-Y. Le Traon, Satellite altimetry and ocean dynamics. Comptes Rendus Geosciences 338(14-15): pp Hastings, D. and P. Dunbar, Development & assessment of the Global Land One-km Base Elevation digital elevation model (GLOBE). International Society of Photogrammetry and Remote Sensing (ISPRS), Commission IV Symposium on GIS - Between Visions and Applications, September 1998, Stuttgart, Germany.6 Hirano, A., R. Welch and H. Lang, Mapping from ASTER stereo image data: DEM validation and accuracy assessment. ISPRS Journal of Photogrammetry and Remote Sensing 57(5-6): pp Nolan, M. and D. R. Fatland, New DEMs may stimulate significant advancements in remote sensing of soil moisture. EOS, Transactions, AGU 84(25): pp Rabus, B., M. Eineder, A. Roth and R. Bamler, The shuttle radar topography mission--a new class of digital elevation models acquired by spaceborne radar. ISPRS Journal of Photogrammetry and Remote Sensing 57(4): pp Rosen, P. A., S. Hensley, I. R. Joughin, F. K. Li, S. N. Madsen, E. Rodriguez and R. M. Goldstein, Synthetic aperture radar interferometry. Proceedings of the IEEE 88(3): pp Schiewe, J., Status and future perspectives of the application potential of digital airborne sensor systems. International Journal of Applied Earth Observation and Geoinformation 6(3-4): pp Skaloud, J., P. Schaer, Y. Stebler and P. Tomé, Real-time registration of airborne laser data with sub-decimeter accuracy. ISPRS Journal of Photogrammetry and Remote Sensing 65(2): pp USGS GTOPO30. [Last access: 1 Aug 2010], at nfo. Zebker, H. A. and R. M. Goldstein, Topographic mapping from interferometric synthetic aperture radar observations. J. Geophys. Res. 91(B5): pp