Advanced geospatial intelligence (GEOINT) with satellite imagery
From illegal mining to conflict zones, learn how to turn satellite imagery into an empirical audit trail
1. Introduction and context
Satellite imagery has become one of the most powerful tools in any investigative journalist’s arsenal. Whether documenting the expansion of illegal mining in the Amazon, tracking the growth of detention camps, or verifying the destruction of civilian infrastructure in conflict zones, geospatial intelligence (GEOINT) provides an empirical eye in the sky that is difficult for governments or corporations to refute.
🔍 1.1. The investigative need
Traditional reporting often relies on witness testimony or leaked documents, both of which can be challenged. Satellite imagery provides objective, time-stamped physical evidence. It allows journalists to monitor remote or hostile locations where ground access is impossible, providing a longitudinal view of change that reveals patterns of intent and impact over months or years.
1.2. Learning outcomes
Master multi-platform integration: Combine Google Earth Pro’s historical archives with Sentinel Hub’s multispectral data.
Conduct temporal analysis: Identify and document physical changes over specific timeframes.
Verify ground truth: Cross-reference satellite visual data with OpenStreetMap (OSM) vector data for naming and functional identification.
Maintain evidentiary standards: Create a verifiable audit trail of geospatial findings for publication.
1.3. Case study hook
Imagine investigating an industrial “buffer zone” where residents claim a factory is encroaching on protected wetlands. By layering 2015 imagery over 2025 data, a journalist can calculate the exact square footage of lost vegetation and pinpoint the date construction began, even if official records are missing or altered.
💡 2. Foundational theory and ethical-legal framework
2.1. Key terminology
Optical vs. multispectral imagery: Optical imagery is what the human eye sees (RGB); multispectral imagery captures wavelengths such as infrared, useful for detecting healthy vegetation or heat signatures.
Spatial resolution: The detail level (e.g., 30cm/pixel in high-res vs. 10m/pixel in Sentinel).
Temporal resolution: How often a satellite captures an image of the same spot.
Orthorectification: The process of removing geometric distortions caused by terrain and camera tilt to ensure the map is “flat” and measurable.
⚖️ 2.2. Ethical and legal boundaries
2.2.1. Consent & privacy
While satellite imagery captures public spaces, journalists should exercise caution when zooming into private residences. The “stop at the login” rule in GEOINT applies to attempts to access private, non-commercial surveillance feeds or other hacked footage. Stick to commercially available or public-domain orbital data.
2.2.2. Legal considerations
Publicly disseminating high-resolution imagery of sensitive military installations may violate national security laws in certain jurisdictions.
🛑 Disclaimer: This tutorial is for educational purposes. Always consult with your newsroom’s legal department before publishing imagery that could be deemed “classified” by a nation or when navigating copyright licenses of commercial imagery providers like Vantor (formerly Maxar) or Planet.
🛠️ 3. Applied methodology: step-by-step practical implementation
3.1. Required tools & setup
Google Earth Pro (Desktop): Essential for its “Historical Imagery” slider.
Sentinel Hub/EO Browser: For free, weekly updates at lower resolution but high spectral variety.
OpenStreetMap (OSM): To identify buildings, roads, and land use designations.
QGIS (Optional): For advanced layering and professional cartography.
👷♀️ 3.2. Practical execution (The “How”)
Scenario 1: Auditing illegal land clearing
Locate the target: Enter coordinates into Google Earth Pro.
Activate history: Click the “Clock” icon in the top toolbar to reveal the timeline slider.
Compare eras: Move the slider to the earliest available high-res date (e.g., 2010) and the most recent (2024).
Spectral verification: Open Sentinel Hub EO Browser for the same coordinates to check the NDVI (Normalized Difference Vegetation Index). A drop in NDVI confirms the “green” seen in old photos is gone, ruling out seasonal changes.
Investigative query table
💾 3.3. Data preservation and chain of custody
To ensure your evidence holds up in court or a formal rebuttal:
Capture metadata: Always take a screenshot of the entire interface, including the coordinates, capture date, and satellite provider (e.g., “Image © 2024 Airbus”).
Log the source: Record the exact URL and sensor ID (e.g., Sentinel-2 L2A).
Hash the files: Once you save your final comparison image, generate a SHA-256 hash immediately.
Command (Mac/Linux): shasum -a 256 image.pngWindows: CertUtil -hashfile image.png SHA256
Archive: Save the original files in a “Write-Once” environment or a secured cloud drive with version history.
🧠 4. Verification and analysis for reporting
4.1. Corroboration strategy
Never rely on a single image. Cross-reference your satellite findings with:
Ground-level imagery: Use Mapillary or Google Street View to see if the “white box” on the satellite is indeed a warehouse.
Social media geolocated posts: Search for Instagram or TikTok posts from that coordinate to find video evidence from eyewitnesses.
4.2. Linking data to narrative
🤖 4.3. AI assistance in analysis
Journalists can use AI/LLMs to parse large amounts of metadata or caption text associated with imagery:
Entity extraction: Use AI to cluster and identify key dates and infrastructure types across dozens of imagery logs.
Multimodal analysis: For more advanced workflows, refer to the framework in The OSINT prompt for AI-powered image geolocation by Spotlight, which details how to structure prompts to assist in place identification while maintaining investigative diligence.
Translation: Use LLMs to translate local land-use permits or planning documents obtained from foreign government portals.
⚠️ Warning: AI can “hallucinate” geographical facts. Never ask an LLM to “tell you what is in this satellite image” and use the result as your primary fact. AI may misidentify a shadow as a building or a cloud as a smoke plume. Human verification of every pixel is vital.
🚀 5. Practice and resources
5.1. Practice exercise
The challenge: Locate the “Tesla Gigafactory Berlin-Brandenburg” in Google Earth Pro. Use the historical slider to find the exact month that major forest clearing began. Use OpenStreetMap to identify which water bodies may have been affected by the nearby construction.
5.2. Advanced resources
Global Forest Watch: An essential tool for monitoring real-time deforestation alerts and historical tree cover loss using Landsat and Sentinel data.
SkyTruth: Specializes in using satellite imagery to track environmental incidents, such as oil spills and illegal mining, often providing expert-verified datasets.
Google Earth Engine: For large-scale planetary data analysis and processing of multi-petabyte geospatial datasets.
Sentinel Hub EO Browser: Free access to multispectral satellite data with pre-built visualization scripts (NDVI, False Color, etc.).
Vantor: A professional geospatial platform for advanced asset monitoring and risk assessment, useful for enterprise-level investigative tracking.
Spotlight tutorials: Repository for the latest AI-assisted OSINT techniques and prompt engineering for journalists.
✅ 6. Key takeaways and investigative principles
Context is king: A change in the landscape is only an investigation if you can link it to a person, company, or policy.
Trust but verify: Use multispectral data (such as infrared) to confirm what your eyes see in RGB.
Historical depth: Always look back at least 5-10 years to understand the land's natural state.
Chain of custody: An image without a timestamp, coordinate, and hash is just a picture, not evidence.
Resolution awareness: Understand the limits of your data; don’t over-interpret a blurry pixel.
👁️ Coming next week…
Tracking aviation and maritime assets
Using open-source trackers (e.g., FlightRadar24, MarineTraffic) to track aircraft and ships. Cross-referencing live transponder data with ownership registries to expose the movement and identity of global aviation and maritime assets.
It's interesting how GEOINT strengthens reporting. Objective data, especially with AI, offers undeniable proof.