L-Band TomoSAR Deep Imaging: From Glacier Internal Structure Mapping to Anomaly Detection
The Earth’s climate has changed through its history. A potential outcome of climate change is glacier melting, which is what we observe over the last decades. It has recently become of concern, as it releases captured gases accumulated over millions of years, and makes the Earth’s surface absorb more incoming solar energy. These accelerate the glacier melting process.
Synthetic Aperture Radars (SAR) and Interferometric SARs (InSAR) are well-understood satellite Earth observation methodologies, which work day and night, and under various weather conditions. They are powerful tools for Earth surface monitoring, displacement characterisation, and deformation mapping. Tomographic Synthetic Aperture Radar (TomoSAR) is a relatively new radar imaging methodology, which adds another dimension to the conventional SAR imagery, i.e. height, using multiple measurements and coherent integration of the information. The idea is similar to computational tomography (CT), intensively used for medical imaging and diagnosis. TomoSAR is suitable for Earth observations, either in the airborne or spaceborne format, for glacier and forest monitoring. L-band RF measurements enable us to further see-through the scene to generate true volumetric mapping. It provides an extra layer of details on top of InSAR imagery. Such glacier images are suitable for quantifying firm bodies and detecting crevasses and underlying bedrock, if applicable, up to 50 meters from the surface .
This project will use L-band radar datasets from ALOS PalSAR-2, SAOCOM-CS for the internal structure volumetric imaging of glaciers using modern machine learning approaches to enable low-cost and accurate solutions for the next generation TomoSAR data processing chains, see for example  and  for X-band TomoSARs, where the height super-resolution is sought. This study enables future developments for ESA Copernicus ROSE-L data processing, when it is launched. On the other hand, the volumetric maps will be used for automatic glacier monitoring tasks, including crevasse and anomaly detection.
 Stefano Tebaldini, et al. “Technical Assistance to Fieldwork in the Austrian Alps During AlpTomoExp”, ESA ESTEC Contract No. 4000111146/14/NL/FF/lf, 2015. (https://earth.esa.int/eogateway/documents/20142/37627/AlpTomoExp-Final-Report.pdf)
 K. Qian, Y. Wang, Y. Shi and X. X. Zhu, “γ-Net: Superresolving SAR Tomographic Inversion via Deep Learning,” in IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1-16, 2022, Art no. 4706116, doi: 10.1109/TGRS.2022.3164193.
 Wang, Muhan, et al. “TomoSAR-ALISTA: Efficient TomoSAR Imaging via Deep Unfolded Network.” arXiv preprint arXiv:2205.02445 (2022).
Please be aware that due to funding requirements this project is only available to applicants from the UK or who have settled or pre-settled status in the UK (i.e. a home fees student)