Satellite Imagery: Data Point Intelligence Guide

Overhead imagery does not tell you what happened. It tells you what changed between two dates, and the analyst supplies the meaning.

Understanding the Satellite Imagery as an intelligence artifact

Satellite imagery is a geo-referenced raster captured by an orbiting sensor. The properties that govern analysis are ground sample distance, spectral bands, revisit interval, off-nadir viewing angle and processing level. Optical multispectral sensors provide free, regular, moderate-resolution coverage suitable for change detection. Synthetic aperture radar images through cloud and darkness and measures surface change directly. Commercial constellations offer sub-metre optical detail at cost, sometimes with the ability to task a new collection over a specific footprint.

Every scene carries acquisition time, sun angle, viewing geometry and cloud cover in its metadata, all of which govern valid interpretation. Processing level matters: top-of-atmosphere and surface reflectance products are not directly comparable. Derived products often carry more signal than the picture itself, including vegetation indices, thermal anomaly layers, radar coherence between passes, and night-time lights for activity and power availability.

Why it matters

Imagery is frequently the only independent observation available for a place where nobody will talk and no records exist. It supports dated evidence of construction, demolition, storage levels, vehicle presence, excavation and maritime activity. Because archives run back decades, it allows retrospective bracketing of when a change actually occurred rather than merely that it did. It corroborates or refutes claims made in reporting, corporate filings and social media posts.

What analysts actually look for

These are the concrete, observable signals that carry weight in this area of work:

Construction progress across dated scenes brackets when a structure appeared, to within the sensor revisit interval.
Vehicle and container counts on hardstand proxy activity level and can be compared across dates and sites.
Shadow length combined with known sun angle yields structure height without any ground access at all.
Thermal anomalies indicate flaring, furnace operation, kiln activity or fires at industrial and energy sites.
Radar coherence loss between passes flags ground disturbance, demolition or movement even under continuous cloud.
Berth occupancy and wake patterns corroborate or contradict vessel position reporting from transponder feeds.
Vegetation and soil indices reveal cultivation, clearance and excavation invisible in ordinary true-colour imagery.
Night-time light change tracks power availability and occupancy across a district, facility or camp.

Where the data comes from

Authoritative and openly available collection points. Always confirm licensing and terms before operational or commercial use:

Copernicus Data Space Ecosystem — Free Sentinel-1 radar and Sentinel-2 optical archives with browser and API access.
USGS EarthExplorer — Landsat archive back to the 1970s plus declassified historical reconnaissance imagery.
NASA Worldview and GIBS — Daily global browse imagery, thermal anomalies and fire detections for rapid situational checks.
Alaska Satellite Facility — Sentinel-1 radar products and processing tools for coherence and change analysis.
Google Earth Pro — Historical high-resolution mosaics with a date slider for retrospective visual comparison.
Maxar Open Data Program — High-resolution imagery released publicly for major disasters and crisis events.
Planet NICFI basemaps — Free high-cadence monthly basemaps covering tropical regions at useful resolution.
Umbra Open Data — Free sample high-resolution synthetic aperture radar scenes over selected sites.

A working method

A repeatable sequence beats ad-hoc searching. This is a practical starting workflow:

Define footprint and window — Fix the polygon and date range before searching, so scene selection is driven by the question rather than availability.
Search all archives — Query free optical and radar sources first, noting cloud cover, off-nadir angle and acquisition time for each candidate.
Establish the baseline — Select a clean pre-event scene under comparable sun and seasonal conditions to make any change interpretable.
Compare like with like — Co-register scenes, match processing level and band combination, and avoid cross-season comparison wherever possible.
Apply derived products — Use indices, thermal layers or radar coherence where the change of interest is invisible in true-colour imagery.
Corroborate independently — Test the interpretation against vessel and flight tracking, permits, ground photography or local reporting before asserting it.
Preserve the evidence — Archive scene identifier, acquisition timestamp, sensor and processing chain alongside every annotated export you produce.

How this connects across the intelligence taxonomy

Intelligence work does not respect neat boundaries. The mission domain you are working, the disciplines you practise, and the data points you pivot on are one connected system. These are the direct relationships for this entry — every link is also a tag, so you can follow any thread across the whole library.

Collected by these disciplines

Geospatial Intelligence — Intelligence Derived from Place
Imagery Intelligence — Interpretation of Visual Imagery
Environmental Intelligence — Environmental Conditions, Damage, and Crime
Economic Intelligence — Economic Conditions, Trade, and Market Signals
Risk Intelligence — Structured Assessment of Threat and Consequence
Meteorological Intelligence — Weather, Ocean, and Atmospheric Conditions
Technical Intelligence — Technology Capability, Design, and Exploitation
Measurement & Signature Intel — Signatures, Measurements, and Physical Phenomena
Legal Intelligence — Law, Litigation, and Regulatory Intelligence
Sanctions Intelligence — Screening, Designations, and Evasion Detection

Investigated in these domains

Pivots to these data points

Location / Coordinates — A geographic point, place, or region — the basis of GEOINT analysis.
GPS Coordinates — Precise latitude/longitude coordinates identifying an exact point on Earth — the atomic unit of GEOINT analysi
Facility / Site — A physical installation — plant, base, port, data centre — with a fixed footprint and function.

Inside the platform: where Satellite Imagery lives

The Quantus platform is 204 pages behind a 147-item sidebar organised into six working groups: Command (24 items), Dashboards (15), Threat Theaters (14), Intelligence Domains (15), Investigate (34), and Administration (45). This entry is not a page in isolation — it is a thread running through several of them.

The modules that matter most here:

datapoint.php?dp=dp_satellite — Data point hub
domain.php?d=mining — Mining & Resource Crime dashboard
domain.php?d=wmd — WMD / Proliferation dashboard
domain.php?d=mil — Military & Defense dashboard
domain.php?d=conflict — Conflict & Humanitarian dashboard
search.php — Advanced search, filter and pivot
correlate.php — Correlation graph
cases.php — Case management

Each dashboard is local-first: it renders from the platform’s own database rather than depending on a live third-party call, so it still works when an upstream API is unreachable or rate-limited. Heavy aggregates are cached with a hard query time cap and degrade to the last good value instead of hanging the page.

Automation, playbooks and AI skills

Analysis that only happens when someone remembers to run it is not a capability. The platform ships a 30-step automation pipeline (cron.php) that collects, ingests, resolves, enriches, correlates and scores on a schedule — 25 seeders, 11 resolvers and 7 enrichment runners, all idempotent and cursor-based so a run can be interrupted and resumed without duplicating or losing work.

AI skills that apply

The 16 one-click operations in ai-skills.php are deterministic jobs, not free-text generation. The ones that matter here:

Enrichment Runner
Enrichment → Local
Correlate Infrastructure
Summarise (Copilot)
Generate Report

Alerting closes the loop: rules in alerts.php fire on new indicators matching a saved query, so a first sighting in this area raises a notification rather than waiting to be noticed at the next review.

Feeds, data sources and the API

The collection layer runs a feed registry of free, machine-readable sources — bulk blocklists and trackers (Maltrail, IPsum, FireHOL, the full abuse.ch corpora, phishing databases, Emerging Threats, Spamhaus, DigitalSide, ThreatView), authoritative government feeds (CISA KEV, OFAC, UN and EU sanctions lists), and reference datasets (RIR allocations, ip-to-ASN and geolocation tables, MITRE ATT&CK, EPSS). collect.php pulls them server-side on a schedule; feeds.php and source-catalog.php show what is registered, what it covers and when it last ran.

Anything the platform holds is reachable programmatically. The REST API in api.php exposes 11 endpoints — status, stats, search, lookup, recent, export, bulk_check, top_threats, by_category, categories, check — and export.php streams 18 formats in bounded chunks, so a million-row export neither exhausts memory nor times out:

STIX 2.1, MISP, OpenIOC 1.1, CEF (ArcSight), LEEF 2.0 (QRadar), Zeek/Bro intel, Snort/Suricata rules, Palo Alto EDL, BIND RPZ, hosts blackhole, iptables, CSV, JSON, NDJSON/JSONL, XML.

That covers the CTI standards (STIX 2.1, MISP, OpenIOC), SIEM ingestion (CEF, LEEF, Zeek), detection engines (Snort/Suricata), and direct enforcement (Palo Alto EDL, BIND RPZ, hosts, iptables) — so intelligence developed here can be actioned in the tools you already run, without a manual reformatting step. A TAXII 2.1 server and a MISP/RSS feed are also served for pull-based sharing.

Use cases

Three ways this entry earns its keep in day-to-day work:

Triage under time pressure. An artifact or report lands and you need a defensible read in minutes, not days. Define footprint and window is the first move; the platform pre-computes the enrichment so the analyst spends the time on judgement rather than lookups.
Building the picture. A single indicator is rarely the story. Establish the baseline turns one artifact into a network — shared infrastructure, repeated selectors, the same operator behind different names — via the correlation graph and the cross-entity link engine.
Producing something actionable. Analysis that ends in a document nobody can use is wasted. Preserve the evidence feeds the case file, the detection rule, the block list or the referral — with sourcing attached so the recipient can verify it.

Case management (cases.php), watchlists, saved searches and scheduled reports mean the work persists between sessions and survives an analyst leaving the team.

How each sector uses Satellite Imagery

The same entry is worked very differently depending on who you are, what authority you hold, and what you are ultimately producing. A military analyst is supporting a commander’s decision; a journalist is meeting a publication standard; an NGO caseworker is protecting a person. The underlying artifacts are shared — the constraints, outputs and thresholds are not.

🎖 Military and defence

Commercial and civil imagery now supplies a large share of unclassified battlespace awareness: order of battle changes, damage assessment, airfield and port activity, and construction of defensive works. The analyst's product is a dated change assessment feeding intelligence preparation of the battlespace and battle damage assessment, with explicit collection times because a scene is a moment, not a state. Constraints matter operationally and legally: a moderate resolution scene cannot support positive identification, imagery alone does not satisfy the verification required before an attack, and the obligation to take feasible precautions means an ambiguous scene should generate more collection rather than a decision. Licence terms on commercial imagery still bind military users.

🕵 National intelligence

Imagery answers requirements where no other collection reaches, and its principal analytic virtue is that it is repeatable: the same footprint can be revisited on a schedule and compared. All-source fusion typically uses imagery to corroborate or refute reporting from other disciplines, and the classification of the product is driven by whether the imagery is commercial or national. Building the assessment on open imagery wherever possible keeps it releasable to partners, industry and the public, which is frequently the point. Every conclusion should be traceable to scene identifiers, acquisition timestamps and processing level so a partner service can reproduce the observation independently.

👮 Law enforcement

Imagery supports search planning, environmental and wildlife crime cases, illegal construction and extraction enforcement, and reconstruction of scenes that no longer exist. Evidentially it is documentary evidence requiring provenance: the scene identifier, sensor, acquisition time, processing level and supplier, with an unmodified original retained and annotation kept as a separate layer. An analyst may need to explain in evidence why a feature is interpreted as it is, so interpretation keys and comparison scenes should be preserved. Commercial licences may restrict courtroom use or redistribution, so check before relying on a scene, and obtain tasked collection through a properly documented procurement.

🔍 Private investigation and corporate security

Corporate investigators use imagery for site verification, insurance claim validation, construction dispute timelines, supply chain and plantation monitoring, and asset confirmation in jurisdictions where a visit is impractical. It is lawful, non-intrusive and leaves no footprint on the subject. The limits are licence terms, which frequently restrict redistribution and derivative works, and resolution, which for free sources cannot support claims about individuals or vehicles. A private actor should not present a moderate resolution interpretation as fact, should not task collection to observe an individual's home, and should preserve the acquisition metadata so a client's counsel can rely on the finding.

📰 Journalism and OSINT media

Imagery is now a standard evidentiary pillar of investigative journalism, and it is also where credibility is most easily lost. The standards are: cite the scene identifier and acquisition date, state the sensor and resolution, show the before and after with matched season and sun angle, and have the interpretation reviewed by someone with imagery experience who is not invested in the story. Avoid over-reading, since absence in one scene means not visible then, not never present. Check licence terms before publishing crops, seek comment from the operator or state concerned, and weigh whether publishing a precise site creates risk for people on the ground.

🌍 NGO, humanitarian and human rights

For human-rights and humanitarian organisations, imagery documents village destruction, camp growth, mass graves, detention construction and disaster impact without exposing local monitors. It is often the only evidence obtainable where access is denied. Preserve to accountability standards with hashes, scene identifiers and an unedited original, because these products are increasingly submitted to courts and commissions. Do no harm applies to publication: imagery that identifies a shelter, a clandestine facility or a displaced community can direct violence toward it. Analysts working repeatedly on destruction and casualty imagery need genuine psychosocial support, and this should be planned rather than improvised.

🎓 University and research

Researchers use imagery for land use change, conflict damage, environmental compliance and development studies, where reproducibility depends on documenting scene selection, atmospheric correction, processing level, classification method and validation data. Report accuracy with a confusion matrix against ground truth rather than presenting a classified map as fact. Free archives such as Sentinel and Landsat are preferred for reproducibility because others can retrieve the identical scenes; commercial imagery frequently cannot be redistributed, which limits replication and should be disclosed. Deposit code, scene identifier lists and validation points, and cite the mission and data provider as required by their attribution terms.

Playbook: working Satellite Imagery end to end

A repeatable sequence, from the moment the requirement lands to the moment a product is delivered and the case is closed out. Each phase states what you are trying to establish, not merely what to click — the point is a defensible chain of reasoning, not a checklist.

Phase 1 — Define footprint, window and question

Fix the polygon, the date range and the specific observable before searching, so scene selection is driven by the question rather than by what is convenient. State what would count as confirmation and what would count as refutation. A good output is a tasking note naming the area, the period and the feature of interest. Stop when someone else could run the search and know whether they had succeeded.

Phase 2 — Survey all available archives

Query free optical and radar archives first, then commercial catalogues, recording for each candidate the acquisition time, cloud cover, off-nadir angle, sensor and processing level. Note coverage gaps as findings in their own right. A good output is a candidate scene table sorted by date with quality attributes. Stop when you know what coverage exists across the whole window, including the periods with none.

Phase 3 — Select a valid baseline

Choose a pre-event scene with comparable season, sun angle and viewing geometry. A baseline from a different season or a sharply different off-nadir angle will manufacture change that did not occur. Where nothing comparable exists, say so and adjust the confidence accordingly. A good output names the baseline scene and justifies the match. Stop when the baseline and the comparison scene differ mainly in what happened on the ground.

Phase 4 — Co-register and standardise

Align the scenes geometrically, match processing level and band combination, and apply consistent stretch so visual differences reflect the surface rather than the display. Mixing top-of-atmosphere and surface reflectance products produces spurious spectral change. A good output is a comparison pair that a reviewer can flip between without noticing anything except the subject. Stop when the only visible difference is the one you are analysing.

Phase 5 — Observe before interpreting

Write down what is physically present or absent, in measurable terms: structures, their footprint dimensions, vehicle counts, stockpile extent, berth occupancy, disturbed ground. Keep interpretation of meaning in a separate column. This discipline is what allows another analyst to disagree with your conclusion while accepting your observation. A good output is an observation record with measurements. Stop when the description would be identical regardless of your hypothesis.

Phase 6 — Apply derived products

Where the change of interest is not visible in true colour, use vegetation, moisture or burn indices, thermal anomaly layers, radar backscatter and coherence, or night-time lights. Each has known failure modes that should be stated, such as thermal detections from unrelated flaring or coherence loss from vegetation growth. A good output records the product, the parameters and the limitation. Stop when the derived layer either shows the change or is shown to be unsuited to it.

Phase 7 — Bracket the event in time

Establish the last scene showing the prior state and the first showing the new state, and report the change as occurring between those two acquisition times. Resist narrowing that bracket beyond what the coverage supports. Radar can often tighten a bracket that cloud has widened. A good output is an explicit date range with the two anchoring scene identifiers. Stop when the bracket is as tight as the coverage genuinely allows.

Phase 8 — Measure what can be measured

Derive structure height from shadow length and sun geometry, footprint area from digitised outlines, stockpile volume from area and repose angle, and vessel length from imagery scale. Express results as ranges with the method stated. Measurements survive scrutiny far better than impressions. A good output is a measurement table with assumptions. Stop when the quantities that constrain the claim have been estimated.

Phase 9 — Corroborate independently

Test the interpretation against vessel and flight tracking, permits, corporate filings, ground photography, local reporting or social media content. Imagery corroborated by an independent line of evidence is a finding; imagery alone is an interpretation. A good output pairs each significant conclusion with its corroboration or an explicit note that none was found. Stop when the core claim is either supported by a second source or flagged as single-source.

Phase 10 — State confidence and alternatives

Write the assessment with an explicit confidence level and name the plausible alternative explanations you considered and why you rejected them. Seasonal variation, construction unrelated to the hypothesis and imagery artefacts should be addressed by name. A good output pre-empts the obvious counter-argument rather than waiting for it. Stop when the strongest alternative explanation has been dealt with in the text.

Phase 11 — Preserve the evidential record

Archive the original scenes with unmodified metadata, record scene identifiers, sensor, acquisition timestamps, processing chain and supplier, and keep annotations in separate files. Hash the originals. A good output is a package another analyst could re-open in five years and verify. Stop when the archive is complete and the licence terms for each scene are recorded alongside it. Store the comparison scenes as well as the exhibit, since the baseline is part of the reasoning.

Phase 12 — Review licence and harm before release

Check redistribution and derivative-work terms for every scene used, check jurisdictional restrictions on publishing imagery of defence sites, and assess whether the published product exposes individuals or vulnerable locations. A good output is a release record naming the licence position and the harm decision. Stop when the published crop is one you can defend both legally and ethically. Where terms are unclear, obtain written confirmation from the supplier before the product goes out.

The platform ships this as a step-checked workflow in playbooks.php, so progress is recorded against a case rather than held in someone’s head.

Source register: what to collect from, and how

Sources are listed with their access model so you can plan around cost and licensing before you build a dependency on them. Open means no account required; registration means a free account or API key; licensed means paid or institutional access. Always confirm current terms — licensing changes, and a source that was free for research may not be free for commercial or evidential use.

Source	Access	What it gives you	How it is used here
Copernicus Data Space Ecosystem	Registration	Free Sentinel-1 radar, Sentinel-2 optical, Sentinel-3 and Sentinel-5P archives with browser, API and cloud processing services.	The default free baseline for change detection: ten metre optical every few days and radar that sees through cloud and darkness.
USGS EarthExplorer	Registration	Landsat archive from the 1970s onward, plus declassified Corona and other historical reconnaissance imagery and aerial photography.	Establishes long-run history of a site, including what preceded current construction decades before commercial imagery existed.
NASA Worldview and GIBS	Open	Daily global browse imagery from MODIS and VIIRS with thermal anomaly, fire detection, aerosol and night-lights layers.	Rapid situational checks, fire and flaring detection, and night-time light change at coarse resolution across large areas.
Alaska Satellite Facility	Registration	Distributed active archive for synthetic aperture radar including Sentinel-1 products, with processing services and tutorials.	Source and processing environment for radar coherence and change products used where optical coverage fails.
Maxar Open Data Program	Open	Sub-metre optical imagery released publicly under open licence for major disasters and selected crisis events.	Provides the high resolution needed for structure-level damage assessment when the event qualifies for release.
Planet and NICFI basemaps	Registration	High-cadence commercial imagery, with monthly tropical basemaps made freely available for forest and land use monitoring.	Fills temporal gaps with near-daily coverage and supports deforestation and land clearance monitoring at useful resolution.
Umbra Open Data	Open	Freely released high-resolution synthetic aperture radar sample scenes over selected sites with permissive licensing.	Demonstrates and supplies high-resolution radar for sites where optical collection is persistently obstructed.
Google Earth and Earth Pro	Open	Historical high-resolution mosaics assembled from multiple commercial providers with a date slider and measurement tools.	Fast retrospective visual comparison and shadow measurement, useful for triage before formal archive work.
Sentinel Hub EO Browser	Registration	Web interface for visualising Sentinel and Landsat data with band combinations, indices and time-lapse generation.	Quick index and false-colour comparison without downloading scenes, suitable for triage and for reproducible screenshots.
ESA Earth Online	Open	Mission documentation, product specifications and processing baselines for European Earth observation missions.	Establishes what a product actually contains and how it was processed, which is required for defensible interpretation.
NOAA and NASA Earthdata	Registration	Broad catalogue of Earth science datasets including thermal, atmospheric, oceanographic and land surface products.	Supplies environmental context such as sea surface conditions, smoke and weather that constrain optical interpretation.
Copernicus Emergency Management Service	Open	Rapid mapping and risk products generated for disasters and crises, including damage grading maps and delineation products.	Provides validated third-party damage assessments that can corroborate or challenge your own interpretation of an event.

Prefer sources that publish a methodology and a revision history. A dataset that changes silently is a liability in any product that has to survive challenge.

Tooling

Tools commonly used against Satellite Imagery. None of these replace judgement, and each carries its own failure modes — know what a tool infers versus what it observes.

QGIS — Loads, co-registers, measures and annotates imagery with extensive plugins; visual stretching choices can create apparent change, so parameters must be recorded.
SNAP — European toolbox for Sentinel processing including radar interferometry and coherence; steep learning curve and easy to produce artefacts read as ground change.
Google Earth Engine — Cloud processing over petabyte archives for time series and index computation; free tier terms restrict commercial use and results depend on undisclosed preprocessing.
GDAL — Converts, mosaics, reprojects and subsets raster data reliably at scale; assumes file metadata is correct about projection and processing level.
SNAP or ASF HyP3 radar services — Generates ready-made interferometric and change products from Sentinel-1 without local processing; queue times and fixed parameter sets limit flexibility.
Shadow height calculators — Derive structure height from shadow length with known sun elevation; results degrade sharply on sloping terrain and high off-nadir scenes.
Hashing and archive tooling — Preserves original scenes with checksums and retrieval records for evidential integrity; does not by itself establish that the supplier delivered an unaltered product.

AI skills and automation in detail

These are deterministic jobs with defined inputs and outputs, not open-ended prompting. Each is idempotent and cursor-based: interrupt one and it resumes where it stopped rather than duplicating work or losing progress.

Enrichment Runner — Walks the indicator set through a chosen provider in time-boxed, cursor-based batches that resume rather than restart.
Enrichment → Local — Materialises enrichment into the local store so dashboards render from your own database instead of a live third-party call.
Correlate Infrastructure — Builds the cross-entity link graph: shared hosting, reused certificates, overlapping registrants, repeated selectors.
Summarise (Copilot) — Produces a narrative summary beside the underlying records. It explains; it never creates indicators or assigns attribution.
Generate Report — Assembles a sourced product from the current case or query, with provenance attached to each element.

A note on the boundary: the only skill that involves a language model is Summarise (Copilot), and it writes prose about records that already exist. Nothing else on this list involves generation of any kind. No indicator, relationship or attribution in the platform originates from a model. See the full skill list.

Tradecraft notes

The distinctions that separate a competent analyst from a fast one:

Match the sun, not just the season. Two scenes months apart at different solar elevations produce different shadow lengths, different apparent vegetation and different visibility into courtyards, and inexperienced analysts read all of it as change.
Radar is the answer to cloud, not a harder optical image. Backscatter responds to roughness, moisture and geometry, so a bright return is not a building and coherence loss is not necessarily demolition. Learn what the sensor measures before asserting what it shows.
Absence in one scene is weak evidence. A vehicle, a stockpile or a ship may simply have been elsewhere at the acquisition time, which is a single instant every few days. Only a consistent absence across a properly sampled series supports a claim of absence.
Measure rather than describe. Shadow-derived height, digitised footprint area and counted objects can be checked by a hostile reviewer and survive; impressions of increased activity cannot and do not.
Off-nadir angle changes what you can see into and behind. A steeply oblique scene hides the far side of structures and lengthens apparent shadows, and comparing an oblique with a near-nadir scene generates false change routinely.
The archive gap is part of the finding. State when coverage exists and when it does not, because a reader who is not told about a six-week cloud gap will assume your bracket reflects observation rather than absence of it.
Have the interpretation reviewed by someone who does not want your hypothesis to be true. Imagery analysis is unusually vulnerable to confirmation bias because the human eye finds whatever pattern it is asked to look for.

Measuring whether it is working

Capability claims should be falsifiable. These are the measures that show whether work on Satellite Imagery is producing anything, and they are worth baselining before you change process or tooling.

Proportion of change assessments issued with an explicit date bracket anchored to two named scene identifiers.
Rate at which imagery interpretations are subsequently confirmed by independent evidence, tracked to calibrate analyst confidence over time.
Share of products where a season and sun-angle matched baseline was used rather than the nearest available prior scene.
Median time from tasking question to first usable coverage, separated by optical and radar to expose weather-driven delay.
Number of assessments revised or withdrawn after peer review, treated as a health indicator rather than a failure count.
Proportion of published imagery products for which licence terms were checked and recorded before release rather than after a query arrived.

Beware of measuring volume alone. Indicator counts and report counts rise easily and say little; time-to-attribution, proportion of findings that survive review, and how often a product changed a decision say a great deal.

Common pitfalls

Absence of a feature in one scene means it was not visible then, not that it never existed at that site.
Seasonal, illumination and off-nadir differences produce apparent change where nothing on the ground actually moved.
Free moderate-resolution imagery cannot resolve individual vehicles or people, so claims at that scale need higher resolution.
Cloud, haze and smoke systematically bias optical coverage away from exactly the events most worth observing.
Publication delay and archive gaps mean the nearest available scene may be weeks away from the event of interest.
Confident interpretation without ground reference remains the most common source of high-profile imagery analysis errors.

Legal and ethical considerations

Commercial and public imagery carries licence terms governing redistribution and derivative works, so check them before publishing crops or annotations. Some jurisdictions restrict imagery or publication concerning defence sites. Imagery of small sites can reveal information about identifiable individuals, so apply proportionality when the subject is a home or place of worship. Preserve original scenes with unmodified metadata and keep annotation in a separate layer so the evidential record stays verifiable and reproducible.

Data integrity: no fabrication, no drift, no hallucination

Intelligence that cannot be traced back to a source is not intelligence, it is assertion. Everything in this entry — and everything in the platform behind it — is built on a small number of non-negotiable rules.

Provenance on every record

Every indicator carries the source that supplied it, a first-seen and last-seen timestamp, and a sighting count. Where several feeds report the same artifact, each contribution is recorded separately rather than collapsed, so you can see whether a finding rests on one source or twelve. Source attribution travels with the data into every export, so a recipient can audit a claim without asking you for the working.

Nothing is invented to fill a gap

If the platform has no data for Satellite Imagery, it says so. Empty is displayed as empty — never padded with plausible-looking placeholder values, sample records or illustrative examples that a reader might mistake for observations. A dashboard with no rows is a true statement about collection coverage, and it is treated as a gap to close, not a blemish to hide.

Scoring is deterministic and reproducible

Threat scores, reputation grades and risk tiers are computed from stated inputs with fixed weights, not estimated. The same inputs always produce the same output, and the formula is visible rather than a black box. Aggregates are cached with an explicit time-to-live so a figure on screen is never silently stale — and when a heavy query exceeds its time budget the platform serves the last known-good value and labels it, rather than inventing a fresh number or hanging.

Where AI is used, and where it is not

Language models summarise and explain. They do not create indicators, assign attribution or manufacture relationships. No IP address, wallet, hash or identity in the platform originates from a model — every one is ingested from a named feed, resolved from a reference dataset, or entered by an analyst with a source recorded. Copilot output is presented as narrative alongside the underlying records, never in place of them, so a reader can always check the summary against the evidence.

Guarding against drift

Enrichment is additive and timestamped rather than overwriting. Reference data — sanctions lists, allocations, taxonomies — is re-synchronised from the authority on a schedule instead of being edited in place, so local copies cannot quietly diverge from the source of truth. Attribution is recorded with a confidence level and the reporting it rests on, and inferred relationships are labelled as inferred. When a source retracts or corrects, the correction propagates rather than leaving a stale assertion behind.

What this means for you

You can put a finding from this platform in front of a regulator, a court, a board or a partner agency and show where each element came from. That is the standard the tooling is built to — because in this work, being confidently wrong is more damaging than being usefully uncertain.

By the numbers

The taxonomy this entry belongs to is not a marketing list — it is the actual structure of the platform: 52 mission domains, 52 intelligence disciplines and 65 data points, each with a live dashboard behind it. Supporting that: 18 indicator types, 14 playbooks, 16 AI skills, 18 export formats and a 30-step automated pipeline.

This particular entry connects directly to 10 intelligence disciplines, 9 mission domains, 3 closely related entries — every one of them a tag you can follow, and a dashboard you can open.

Questions analysts actually ask

Free imagery is ten metres. What claims can I actually make with it?

Plenty, provided they are about things larger than the pixel. Ten metre optical supports claims about buildings appearing and disappearing, roads and berms being constructed, land clearance, flooding, large stockpiles, ship presence at berths and agricultural change. It does not support claims about individual vehicles, people, weapons or the identification of specific equipment, and any product asserting those from Sentinel data is wrong. The honest approach is to use free imagery to establish the pattern and the dates, then use higher resolution selectively for the specific observations that require it, stating the resolution alongside each claim.

Nothing is visible because of persistent cloud. What are my options?

Move to radar. Sentinel-1 images through cloud, smoke and darkness, and while it does not look like a photograph it directly answers questions about whether ground has been disturbed, whether structures have appeared or been removed, and whether ships are present. Coherence between passes is especially sensitive to change. Beyond that, consider thermal and night-lights layers for activity indicators, look for oblique commercial collections that may have caught a gap in the cloud, and check whether a mapping service or an emergency response product covers the period. If none of that works, report the coverage gap as a finding.

How do I know whether an apparent change is real?

Test the alternatives explicitly. Seasonal vegetation, soil moisture after rain, different sun elevation, different off-nadir angle, different processing level and display stretch all create apparent change. Check a third scene between your pair, check the same feature at a control site nearby that should not have changed, and check whether the change persists across subsequent scenes rather than appearing once. Then measure it: if the footprint dimension changed, it is real; if only the tone changed, be very cautious. Write the alternatives you eliminated into the assessment, because the reviewer will ask.

Can I publish a cropped commercial satellite image in a report?

Only if the licence permits it, and terms vary substantially. Open programmes such as the Sentinel missions and Landsat allow redistribution with attribution. Commercial providers usually license imagery for internal use, with separate terms for publication, and some prohibit derivative works or require a specific credit line and resolution limit. Open data releases for disasters are typically permissive. Check the exact terms for each scene, record them with the archive, and where the imagery is central to a public product, obtain written confirmation. A licence dispute after publication distracts entirely from the substance of the finding.

How far back can I look at a site?

Further than most analysts expect. Landsat provides moderate resolution from the 1970s, which is enough to see whether a site existed at all and when major construction began. Declassified reconnaissance imagery covers much of the world from the 1960s at surprisingly high resolution and is free through the United States Geological Survey. Commercial high resolution generally starts around 2000, and historical mosaics in Google Earth often go back to that period. National aerial photography archives frequently predate all of it. The practical limit is usually not the imagery but the effort of georeferencing scanned historical material.

How should imagery evidence be preserved for a court or commission?

Keep the original product exactly as delivered, with its metadata intact, and hash it on receipt. Record the scene identifier, sensor, acquisition timestamp in UTC, processing level, supplier and the date and method of acquisition. Keep every annotation, crop and enhancement as a separate derived file with its own provenance note describing what was done, so the chain from original to exhibit is explicit and reversible. Retain the comparison scenes and the interpretation reasoning, not only the final image. The Berkeley Protocol sets out this practice in detail and is increasingly the standard that accountability bodies expect.

Standards, frameworks and further reading

Work that references a recognised framework is easier to defend, easier to hand over, and easier for a partner to consume:

The Berkeley Protocol on Digital Open Source Investigations defines preservation, verification and documentation practice for imagery used in accountability proceedings.
ISO 19115 geographic metadata governs the lineage, resolution and quality statements that should accompany imagery products.
The CEOS analysis ready data specifications define processing levels that make scenes comparable across sensors and dates.
The Copernicus open data policy governs free use and redistribution of Sentinel imagery, subject to attribution.
The Landsat open data policy places United States civil imagery in the public domain, enabling reproducible research and publication.
The Open Geospatial Consortium web map and coverage service specifications define how imagery services are accessed programmatically.
The International Charter Space and Major Disasters governs the activation and release of imagery for disaster response.
Additional Protocol I to the Geneva Conventions requires feasible precautions in attack, a standard that imagery-only identification does not satisfy.

References

Primary sources and authoritative references for this entry. Publishers revise and retire material, so treat the retrieval date as part of the citation and re-check before relying on any of it in a formal product.

Copernicus Data Space Ecosystem — European Space Agency and European Commission. Free access to Sentinel optical and radar archives with processing services.
EarthExplorer and the Landsat archive — United States Geological Survey. Decades of moderate resolution imagery plus declassified historical reconnaissance.
Worldview and Global Imagery Browse Services — United States National Aeronautics and Space Administration. Daily global browse imagery with thermal, fire and night-lights layers.
Alaska Satellite Facility SAR data and services — University of Alaska Fairbanks and NASA. Radar archives and processing services for coherence and change analysis.
Berkeley Protocol on Digital Open Source Investigations — United Nations Office of the High Commissioner for Human Rights. The reference standard for preserving and documenting open source imagery evidence.
Copernicus Emergency Management Service rapid mapping — European Commission. Validated damage and delineation products for disasters and crises.
Open Data Program — Maxar Technologies. Publicly released high resolution imagery for major crisis events.
Earth Online mission and product documentation — European Space Agency. Authoritative specification of what each imagery product contains and how it was processed.

Link integrity: every reference above was verified with a live request when this page was generated. Where a publisher had moved or withdrawn a document, the link was repointed at a preserved copy in the Internet Archive and marked as archived. Anything with no reachable copy anywhere had its link removed rather than left to rot — the source is still credited, it simply cannot be linked.

Put it into practice

The Quantus Intel threat intelligence platform operationalises this entry: pairs site footprints with dated imagery baselines and flags change against prior observations. Explore the platform, or browse the rest of the library by following any tag above.