InSAR & Digital Terrain Specialist (Contractor)

The specialist will be responsible for transforming raw InSAR data (Single Look Complex - SLC pairs) into high-fidelity, bare-earth Digital Terrain Models. This terrain model will be the primary input for hydrodynamic modeling. 

1. Primary InSAR Processing (SLC to DSM)

• Data Selection & Ingest: Select suitable SLC image pairs (e.g., Sentinel-1, TanDEM-X) based on optimal perpendicular and temporal baselines.
• Coregistration: Precisely register slave images to the grid of master images with sub-pixel accuracy to maintain signal coherence.
• Interferogram Generation: Calculate the complex product of image pairs and perform flattening to remove the flat-earth phase.
• Adaptive Filtering: Apply filters (e.g., Goldstein) to reduce speckle and phase noise, facilitating more reliable phase unwrapping.
• Phase Unwrapping: Resolve the 2π ambiguities of the wrapped phase using industry-standard algorithms such as SNAPHU (Statistical-cost, Network-flow Algorithm for Phase Unwrapping).
• Geocoding & Terrain Correction: Convert absolute phase into elevation relative to a reference ellipsoid and orthorectify the data into a standard map projection (UTM/WGS 84), resulting in a Digital Surface Model (DSM).

2. Terrain Extraction (DSM to DTM)

• Ground Filtering: Identify and remove above-ground objects (buildings, trees, bridges, highways, man-made infrastructure) from the DSM using morphological ground filtering or probabilistic voting frameworks.
• Surface Reconstruction: Fill empty spaces left by removed objects using advanced interpolation techniques like Kriging, Linear Tinning, or Least Squares Surface Reconstruction. 
• Deep Learning Integration (Optional but Preferred): Implement end-to-end deep learning architectures (e.g., EfficientUNet or PGENet) for automated phase unwrapping and ground filtering to reduce manual labor and improve accuracy in complex terrain.

3. Validation & Quality Control

• Error Analysis: Mitigate atmospheric delay artifacts through stack averaging or numerical weather models.
• Accuracy Assessment: Validate InSAR-derived DTMs against high-precision reference datasets, such as LiDAR-derived models or GPS independent check points (ICPs). 
• We will mainly work in urban areas in Vietnam, with the road network as our primary concern.

• Experience with hydrodynamics modeling.
• Experience with urban Southeast Asia terrain model. 

• We will first do a paid test processing open-source InSAR data into a DTM. 
• After the test, the contract will be negotiated based on the project scope.