Parallel InSAR Pipeline with Docker & ISCE2

An automated, containerized, and parallel processing pipeline for Interferometric Synthetic Aperture Radar (InSAR) processing based on ISCE2 (topsApp).

Designed to handle large-scale Sentinel-1 datasets efficiently by leveraging Python multiprocessing and Docker isolation.

🚀 Key Features

• Parallel Processing: Built-in process pool scheduler (multiprocessing) to handle multiple Interferogram pairs simultaneously.
• Dockerized Environment: Zero-dependency deployment. Runs anywhere (Linux, Windows/WSL2, Cloud) with Docker.
• Auto-Provisioning:
  • Automated DEM download from OpenTopography (with on-the-fly XML/VRT generation).
  • Automated Orbit files download from ASF.
• Robustness:
  • Phase-Split Execution: Smartly handles ISCE step dependencies to enable checkpoint restart (e.g., skipping ion step in memory).
  • Fault Tolerance: Automatic retry logic for unstable steps and task isolation (one task failure doesn't stop the pipeline).
• Storage Optimized: "Nuclear" cleaning scripts to minimize disk usage by 90% after processing.

📂 Project Structure

InSAR_Pipeline/
├── Dockerfile              # Docker build configuration
├── run_docker.sh           # One-click startup script
├── .env.example            # set your API_KEY，delete`.example`
├── environment.yml         # python site-packages
├── code/
│   ├── main_parallel.py    # Main scheduler & processor
│   ├── modules/            # Helper modules (DEM, Orbit, XML)
│   └── entrypoint.sh       # support utilities
└── data/                   # Data directory (Mounted volume)
    ├── raw/                # Sentinel-1 SLC data (.SAFE)
    ├── orbit_data/         # Sentinel-1 Orbit data (.EOF)
    ├── dem/                # DEM storage
    └── runs/               # Processing outputs
    pairs.txt               # target pairs

🛠️ Getting Started

Prerequisites

• Docker Desktop (Windows/Mac) or Docker Engine (Linux)
• OpenTopography API Key

1. Setup Data

Place your Sentinel-1 SLC data into data/raw.

2. Configure Pipeline

Edit pairs.txt to set your target pairs:

("S1A_...20230203...SAFE", "S1A_...20230215...SAFE")
("S1A_...20230215...SAFE", "S1A_...20230227...SAFE")
("S1A_...20230203...SAFE", "S1A_...20230227...SAFE")

Put a pair of pairs on one line, and then jump

Edit .env to set your target pairs:

OPENTOPO_API_KEY=your API
ROI=38.0,41.0,33.0,38.0  # Set the range of downloaded DEM
WORKERS = 2  # Set based on your RAM (rec: 16GB per worker)

3. Run with Docker

Execute the startup script. It will mount your local data and code into the container.

chmod +x run_docker.sh
./run_docker.sh

📊 Results

Interferogram (Wrapped Phase)

Interferogram

see the results_sample/Interferogram

Unwrapped Phase

Unwrapped

see the results_sample/Unwrapped Phase

📝 License

This project is open-sourced under the MIT License.