D3-ID32278: 3 & 4D Topography Measurement
32278-2 Multi-baselineSAR processing for 3D/4D reconstruction
1LIESMARS ,Wuhan University, Wuhan, China; 2Collaborative Innovation Center of Geospatial Technology, Wuhan, China; 3Key Laboratory of Land Subsidence Monitoring and Prevention,Ministry of Land and Resources, Shanghai, China
InSAR techniques provide researchers a set of tools for topographic mapping, as well as for monitoring deformations on the Earth surface. In Dragon-1 and Dragon-2, our focus was on DEM generation and surface motion estimation with medium resolution InSAR. Since Dragon-3, SAR datasets of high spatial and temporal resolution (TerraSAR-X, COSMO-SkyMed) are available and with the availability of Sentinel-1 data, a global time-series coverage is now reality.
Topographic mapping and surface motion estimation with spaceborne SAR sensors are the main topics of the Dragon-4 project "Multi-baseline SAR processing for 3D/4D reconstruction (id 32278-2)” under the framework of THREE- AND FOUR-DIMENSIONAL TOPOGRAPHIC MEASUREMENT AND VALIDATION (id 32278). In Dragon-4, we work on different test sites investigating the following four topics:
1. Topographic mapping with SAR. We developed a maximum a posteriori (MAP) estimation method for multi-baseline InSAR assisted by StereoSAR. According to Bayesian theory, the combination of StereoSAR and InSAR for topographic mapping can be viewed as update of the StereoSAR DSM with InSAR phase observations. At the same time, the StereoSAR DSM is also a constraint to InSAR phase observations, which can solve the problem of elevation ambiguity and avoid phase unwrapping problems. The Mount Song has been selected as the test area, which is one of the five sacred mountains of China. The experimental result shows that there is neither systematic error nor large data voids in MAP estimated DSM and the standard deviation of height error σh of MAP estimated DSM is less than 10 m with respect to the photogrammetric DEM for the whole area, while in plain areas σh is about 5 m.
2. Urban subsidence analysis. SAR systems can measure distances and movements with high precision. Using for example PS-InSAR, deformations can be estimated with a very high precision. The long-term surveillance of urban subsidence and the infrastructure stability in Shanghai is our major research goal since Dragon-1. With data starting from ERS-1, over ENVISAT ASAR, ALOS PALSAR, up to modern systems like TerraSAR-X, COSMO SkyMed, PALSAR-2, and Sentinel-1, we continuously monitor the subsidence over Shanghai for far over a decade now. The combination of this data and the analysis of the continuous deformation is still on-going. Remarkably, the PS-InSAR precision stays stable over time even using different sensors.
3. Coseismic displacement from Sentinel-1 TOPS data. Terrain Observation by Progressive Scans (TOPS) mode from the Sentinel-1A/B satellites provides up-to-date high-quality Synthetic Aperture Radar (SAR) images over a large coverage, making it widely applied to earthquake studies. Recent work focuses on generation of co-seismic displacement of large earthquakes from Sentinel-1 TOPS images. However, many small/deep/offshore earthquakes have relatively smooth ground displacement disturbed by strong atmospheric influence. The coherence images spanning these earthquakes maybe not desirable e.g. due to the complex topography. Therefore, methods to derive such smooth co-seismic displacements from time-series is needed. We developed a new Sentinel-1 TOPS images analysis strategy with applications to earthquakes occurred recently in China.
High-precision 3D Reconstruction from Synthetic Aperture Radar and Optical Images
1LIESMARS, Wuhan University; 2ifp, University Stuttgart
The reconstruction of topographic information and their changes in the context of our dynamic Earth is one of the main applications of photogrammetry and remote sensing. With advancements in sensor technology and data processing, three-dimensional information can be retrieved in unprecedented precision. With the release of the TanDEM DEM, the DLR released the most precise world-wide DEM, an invaluable data set for numerous applications. The global availability and the high precision are proof of the extra-ordinary capability of Synthetic Aperture Radar (SAR) interferometry (InSAR) for 3D measurements.
With the SAR geodesy concept of the DLR, the radargrammetric measurements are having a comeback. Thanks to the high orbit-precision of TerraSAR-X and TanDEM-X, and by using collateral data to correct the atmospheric delay, absolute 3D positioning precision within a few centimeters is nowadays possible. However, there are several limitations for reaching the maximum precision in SAR geodesy and interferometric SAR. Therefore, a deeper understanding of the underlying error sources is necessary to fully use the potential of the aforementioned methods.
With the rapid developments in photogrammetric computer vision the multi-view photo-consistency measures for dense and accurate 3D reconstructed evolved and based on developments like semi-global matching for multi-view-stereo, photogrammetry reaches relative height precisions in the centimeter domain, comparable to or even outplaying LiDAR.
In this context, the Dragon-4 project “Topographic Mapping - Validation (32278-1)” is working on the three- and four-dimensional high-precision measurement using SAR and photogrammetry. Currently, we are working on three main objectives:
With the recent progress in the three-dimensional measurement precision of photogrammetry and SAR, four-dimensional time-series measurements for surface motion estimation can become possible offering alternatives to differential interferometry based methods. This is especially beneficial, because these methods can be used as supplement to differential interferometry and derived methods, because they are especially applicable for fast surface motions that are otherwise especially difficult to be measured.
Temporal Decorrelation analysis of TropiScat
1Politecnico di Milano, Italy; 2Wuhan University, China
In this paper, we provide a better understanding of temporal decorrelation of tropic forest from TropiScat in various timescales. TropiScat is a ground-based campaign operated at the Paracou field station, French Guiana since October 2011, which allows gathering the tomogram of the forest in all polarizations at P band with a temporal sampling of 15 minutes . To analyze the temporal decorrelation we evaluate the intensity and coherence of the signal array of 15 transmitting and receiving antenna pairs. The vertical distribution of temporal coherence is obtained by comparing two tomograms acquired at different times. To further understand the temporal decorrelation, we evaluate the coherence between tomogram obtained from antenna pairs of different times and that of the same time. The procedure to produce the tomogram and different tomogram processors will be discussed in the final paper .
Deriving coseismicdisplacement from time-series Sentinel-1 TOPS images spanning earthquakes
1LIESMARS, Wuhan University, Wuhan, China; 2Earth Observatory of Singapore, Nanyang Technological University, Singapore; 3Collaborative Innovation Center for Geospatial Technology, Wuhan, China
Terrain Observation by Progressive Scans (TOPS) mode from the Sentinel-1A/B satellites provides up-to-date high-quality Synthetic Aperture Radar (SAR) images over a large coverage, making it widely applied to earthquake studies. Recent work focuses on driving coseismic displacement of large earthquakes from a pair of TOPS images. However, many small/deep/offshore earthquakes have relatively smooth ground displacement disturbed by strong atmosphere influence. The coherence images spanning these earthquakes maybe not desirable either duo to the complex topography. Therefore, methods to derive such smooth coseismic displacement from time-series analysis is needed. Here we present a novel Sentinel-1 TOPS images analysis strategy with applications to a few earthquakes occurred recently in China.
First is the 5 February 2016Mw 6.4 MeiNong earthquake that occurred in Taiwan. For this case, we take a modified spectral diversity method for coregistration of TOPS images to get a smooth interferogram in which the obvious phase jumps are well corrected. Then we take time-series analysis using TOPS images acquired before and after the earthquake to improve the derived coseismic displacement. The results are validated with GPS data near the epicenter area.
Another case includes a series of M<7 earthquakes occurred on the Qinghai-Tibet plateau, To study these earthquake, we collect time-series Sentinel-1 TOPS images acquired before and after earthquakes. We improve the time-series TOPS data processing chain to estimate the coseismic displacement on detected persistent scatters. Our results show that the temporally uncorrelated atmospheric signal can be largely reduced and the subtle coseismic displacement signal can be derived more precisely than single interferogram.
High Precision DSM Generation in Mountainous Areas with Multi-Baseline InSAR
1LIESMARS, Wuhan University, Wuhan, China; 2Collaborative Innovation Center of Geospatial Technology, Wuhan, China
Synthetic Aperture Radar Interferometry (InSAR) is a powerful tool for large-area topographic mapping due to its capability of all-time all-weather imaging and high sensitivity to terrain relief. However, there is an inherent contradiction between geometric decorrelation and sensitivity of height measurement for topographic mapping with a single InSAR pair. A normal baseline of proper length is required to keep a balance between the two issues. A promising solution to this problem is the so-called multi-baseline InSAR analysis. The basic principle of multi-baseline InSAR is to derive an optimal height estimate by joint analysis of multiple phase measurements from several interferograms with different normal baselines. Compared with single-baseline InSAR, the major benefit of using multi-baseline observations is the possibility of exploiting redundant topographic phase observations with different height of ambiguities to improve the accuracy of phase unwrapping, or even avoid phase unwrapping.
In this study, after the analysis and discussion of the probability distribution of interferometric phase, we propose a maximum likelihood (ML) estimation method assisted by SRTM DEM data for multi-baseline InSAR topographic mapping. Based on our method, a spaceborne data multibaseline InSAR processing flow is established. In order to well adapt it to the repeat-pass interferometric pairs, the processing flow integrates the atmospheric effect correction method to improve the reliability of multi baseline estimation. The simulation experiments were designed to test the effectiveness of the maximum likelihood height estimation method and atmospheric effect correction method. The proposed multibaseline InSAR processing flow was applied in ALOS/PALSAR dataset covering Mount Tai area, China. The accuracy of resultant DEMs at spatial resolution of 20 m is evaluated by using a standard Chinese national DEM of scale 1:10,000 as reference. The results show that multi-baseline InSAR can improve DEM accuracy compared with single-baseline case. The quality of multi-baseline InSAR DEM can meet the American DTED-2 standard.