Satellite imagery for open-source change detection

Block

01 · Geolocation & Chronolocation

Topic

Satellite Change Detection

Tools

Copernicus Browser · USGS Earth Explorer · QGIS

Difficulty

Intermediate

Prerequisites

Pinpointing location through visual clues: shadows, architecture and terrain.

01	Satellite archives give you a time machine no witness can match Multispectral satellite archives let you establish what was at a location, when it changed, and what replaced it, without physical access or a witness.

When ground access is denied and witnesses have fled, multispectral satellite imagery lets you establish what was there, when it changed, and what replaced it. ESA's Sentinel-2 and USGS Landsat programmes provide free, repeating coverage of every point on Earth at resolutions sufficient for documenting construction, demolition, agricultural clearance, flood extent and military activity. The core investigative value is revisit frequency and historical depth, not resolution: Sentinel-2 images any given location every five days under cloud-free conditions, and the Landsat archive extends back to 1972.

Change detection does not require specialist remote-sensing training. It requires a repeatable workflow: identify a baseline image before the event, select a comparison image after it, and produce a documented sequence that any reader can verify by replicating the same query on the same public platform.

In the field

In 2021, InfoAmazonia and Earthrise Media used free Sentinel-2 imagery to document the expansion of illegal gold mining along Brazil's Tapajós River in Pará state, comparing archive imagery from 2016 against 2021 to quantify growth in mining sites and trace the resulting sediment pollution downstream.

Temporal sequencing. A comparison between 2016 and 2021 Sentinel-2 imagery showed active mining points increasing from roughly 2,700 to 4,700, a growth of nearly 75%.
Algorithmic detection at scale. A neural network trained to recognise mining-pit signatures in Sentinel-2 imagery flagged candidate sites across the basin, turning free 10 m optical data into a scalable investigative tool without commercial-grade resolution.
Downstream corroboration. Sediment plumes traced in the same imagery extended hundreds of kilometres downstream, corroborating the mining-site findings with an independent visual signature.

InfoAmazonia / Earthrise Media · Tapajós River gold mining pollution · 2021

Learning outcomes

By the end of this tutorial you will be able to:

Access and filter Sentinel-2 and Landsat imagery for a defined area of interest and time window
Produce a documented before-and-after image pair that establishes a change date range
Apply basic spectral indices (NDVI, NDWI) to distinguish change type from visual inspection alone
Export image composites with full provenance metadata for inclusion in an evidence package

The rest of this tutorial is for Signal subscribers.

What remains: the decision framework, the tool configuration, the failure modes, and the evidentiary standard required to use the finding defensibly. Signal is €90 a year, or €9 a month. Students, €49 a year.

Join Signal

A Signal subscription gives you:

Full OSINT Reference Card library, 21 domains
Methods, every tradecraft tutorial in full
AI in OSINT, every prompt and field report
Shadow Analysis, every forensic dossier

Satellite imagery change detection: Sentinel and Landsat

Satellite archives give you a time machine no witness can match

Learning outcomes

The rest of this tutorial is for Signal subscribers.

A Signal subscription gives you:

Keep Reading