GEOINT — Geospatial Intelligence Course
Goal: Acquire, process, and analyze satellite and aerial imagery for intelligence — from finding free data to producing change detection reports and target identification.
Prerequisites
- Python 3.10+ (comfortable with numpy, matplotlib)
- Basic image processing concepts (pixels, channels, histograms)
- Helpful: ML Computer Vision course (object detection, CNNs)
- Helpful: basic Linux CLI (you’ll run GDAL commands)
Tools You Need
| Tool | Purpose |
|---|---|
| Python 3.10+ | Everything |
| rasterio | Read/write geospatial rasters |
| numpy | Band math, array operations |
| matplotlib | Visualization |
| GDAL | Raster/vector processing CLI |
| shapely + geopandas | Vector geometry, clipping |
| Optional: QGIS | Visual inspection, quick checks |
| Optional: Google Earth Engine | Cloud-scale processing |
pip install rasterio numpy matplotlib geopandas shapely pyproj
# GDAL: sudo apt install gdal-bin libgdal-devCourse Levels
L0 — Foundation (~6 hours)
Build the mental model. What satellites exist, how they see, what limits them.
- Satellite Fundamentals — orbits, resolution, revisit, major systems
- Sensor Types and Imagery — optical, SAR, thermal, LiDAR
- Gate: explain 4 resolution types, choose correct sensor for 3 different intelligence questions
See Also
- End to End Learning for Self-Driving Cars — Pomerleau (1989). Early end-to-end learning paper demonstrating that neural networks can learn to drive from camera input alone — foundational for modern autonomous systems and remote sensing applications.
L1 — Basic Skills (~10 hours)
Hands on data. Download, load, visualize, compute indices.
- Tutorial - Acquiring Free Satellite Imagery — find and download Sentinel-2, Sentinel-1, Landsat
- Tutorial - Working with Raster Data — rasterio, CRS, band math, visualization
- Multispectral Analysis — vegetation/water/urban indices, spectral signatures
- Gate: load Sentinel-2, compute NDVI, classify land cover into 4 categories, export GeoTIFF
L2 — Applied Techniques (~12 hours)
Core GEOINT workflows: change detection, SAR, object detection, terrain.
- Change Detection — image differencing, NDVI change, CVA, time series
- SAR Fundamentals and Analysis — Sentinel-1, speckle, backscatter, InSAR
- Tutorial - Object Detection in Satellite Imagery — tiling, YOLO, ship detection, vehicle counting
- Terrain Analysis and Geolocation — DEM, viewshed, shadow analysis, OCOKA
- Gate: produce change detection report for a facility over 3 months with before/after imagery
L3 — Judgment (~10 hours)
Analyst thinking. Fuse sources, assess confidence, work real scenarios.
- Multi-Source Intelligence Fusion — ACH, structured analysis, bias
- SATCOM and Space Segment — link budgets, ASAT, space EW
- Case Study - Monitoring Military Installations — 6-month facility monitoring
- Case Study - Maritime Domain Awareness — SAR ship detection, AIS correlation
- Gate: produce multi-source assessment of a scenario with confidence levels and alternative hypotheses
L4 — Integration (~7 hours)
Build a working pipeline that ties it all together.
- GEOINT Capstone — automated change monitoring pipeline
- Gate: working end-to-end pipeline + analysis report + explanation of confidence and limitations
Total estimated time: ~45 hours
How to Use This Vault
- Go in order. L0 → L1 → L2 → L3 → L4. Each level assumes prior levels.
- Run the code. Every note has runnable Python. If you don’t run it, you didn’t learn it.
- Pass the gates. Each level has assessment criteria. Don’t skip to the next until you can do them.
- Use the flashcards. GEOINT Flashcards covers every key concept in spaced repetition format.
- Cross-reference. This course connects to your EW-Recon (SAR, signals), Security OSINT (open sources), AI/ML CV (object detection), and Forensics (image analysis) courses. Follow the links.
The Big Idea
Read The GEOINT Mind Map first. Five patterns connect everything in this course. Every technique, every tool, every analysis method traces back to one of them. Understanding these patterns is more important than memorizing any single technique.
Quick Reference
| Data Source | Type | Resolution | Revisit | Cost |
|---|---|---|---|---|
| Sentinel-2 | Optical | 10m | 5 days | Free |
| Sentinel-1 | SAR | 5x20m | 6-12 days | Free |
| Landsat-9 | Optical | 30m | 16 days | Free |
| SRTM DEM | Elevation | 30m | Static | Free |
| Planet | Optical | 3m | Daily | Commercial |
| MAXAR | Optical | 30cm | Variable | Commercial |
Next: The GEOINT Mind Map