Checking satellite imagery before you trust a map

Choosing imagery for a project in Asia starts with the question the image is meant to answer, then with three properties: resolution, revisit time and licensing. An image that looks convincing at screen size can be the wrong instrument, because a ten metre pixel cannot show a field boundary and a single pass cannot show a flood moving. The free archives that cover the continent are also the ones that get misread, so this note sets out what to check before an image is repeated in a report or a decision.

A satellite scene is a record, and a record earns its use the same way any other record does: when someone can say where it came from, when it was taken, and what it does not show. That habit is the same one this desk applies to a suspicious message, moved to a different kind of evidence.

How do I choose satellite imagery for a project in Asia?

Start from the phenomenon and its timescale. A monsoon flood changes over days and often under cloud, which points to radar. A rice harvest, a forest edge or a new road appears over weeks in optical data at ten to thirty metres. Urban growth along a fringe is a question about years, so a time series matters more than any single scene. Weather is the exception that reports itself continuously rather than on a revisit cycle.

Ground Truth Asia, an independent magazine on satellite remote sensing across Asia, keeps the archive side of this question in one place: its guide on choosing satellite data asia sets out where the images come from, what the scenes cannot do, and how the continent is read from orbit and checked on the ground.

Then read the licence before the pixels. An open archive can still carry conditions on redistribution, and a report that republishes a scene should be able to say which terms apply. Resolution is the last filter, not the first: a finer pixel with a poor revisit is less useful than a coarse one that returns every few days.

Where can you download free satellite images of Asia?

The two open workhorses are Landsat and Sentinel-2, which cover all of Asia every few days at no cost and are searched by place and date. For real-time weather over the continent and the Pacific, Himawari photographs the whole disk every ten minutes, which is why forecasters, sailors and flood teams read that stream rather than a daily scene. The European Space Agency documents the second of these missions on its Sentinel-2 mission page, including the instrument bands and the revisit pattern a project has to plan around.

Downloading is the easy part. Two habits make the archive useful: record the scene identifier and the acquisition date with the file, and store the original beside any processed copy, so a later reader can return to the source instead of to a screenshot.

Which Asian countries fly their own Earth observation satellites?

Japan, India, Korea, Thailand and China each operate national systems, alongside the shared European and American archives, which is why a project in the region rarely depends on a single provider. The research community around that work meets each year at the Asian Conference on Remote Sensing: its 46th edition met in Makassar in October 2025 with 450 participants from 25 countries and 244 papers, and its proceedings are archived for anyone who wants to see how a method was argued rather than only what it concluded.

National fleets change the practical question from availability to continuity. A programme that has produced a long series is more useful for change detection than a newer instrument with three scenes, and the series, not the specification sheet, is what a project should check first.

What ground truth adds to an image

An image reports reflectance, not meaning. A flooded field and a flooded road can look alike, and a classification is an argument about thresholds. Ground checks, field measurements or a second independent source are what turn a scene into evidence, and they are also what let a report say how confident it is.

The same logic applies to who gets access to the record. Public archives were argued into existence by people who wanted the data held in common, which is the older argument behind community networks, described in our note on public access computing security.

What limits the interpretation

Every scene in this note is a measurement with a footprint, a time and a licence, and none of them is a conclusion. If a decision depends on the image, keep the scene identifier, the acquisition date and the processing steps with the decision, and say plainly which claims the data supports. Where the archive cannot answer a question, a field check or a second source is the honest next step rather than a sharper-looking figure.