Satellite imagery for Asian territories: a practical guide
How to pick satellite imagery for Asia by resolution, revisit and licence, where free archives open, and which countries fly their own Earth observation
Choosing satellite imagery for a project in Asia comes down to three filters applied in order: the spatial resolution the question needs, the revisit frequency the phenomenon demands, and the licence that governs redistribution. Free archives cover most of the region, and several Asian countries also operate their own Earth observation satellites alongside the open Landsat and Sentinel programmes.
How do I choose satellite imagery for a project in Asia?
Start from the smallest object that must be visible in the final map, then work backwards. A flood extent mapped at district level tolerates 10 to 30 metre pixels; a single rice paddy boundary does not. Resolution, revisit and licence interact, and a dataset that wins on one usually loses on another, so the practical order is: fix the resolution floor, then the maximum acceptable gap between two usable acquisitions, then the licence.
Resolution is normally quoted as ground sample distance, the size on the ground of one pixel. Landsat 8 and 9 carry the Operational Land Imager at 30 metres in visible and near infrared, with a 15 metre panchromatic band. Sentinel-2 offers 10 metres in four bands, 20 metres in six and 60 metres in three. Both are free and open, which makes them the default starting point for continental or national work.
Revisit is the time between two passes over the same point. Landsat repeats every 16 days per satellite, and the combined Landsat 8 and 9 constellation brings that to about 8 days. Sentinel-2 runs a 10 day revisit per satellite, about 5 days with the pair. Cloud is the real constraint in monsoon Asia: a nominal 5 day revisit can still yield one usable scene per month over the Mekong delta in July.
Licence decides what can be published. Landsat data are in the public domain, with no restriction on redistribution. Sentinel data are released under the Copernicus open licence, which permits commercial and non commercial reuse with attribution. National missions vary: some restrict redistribution, some require a request, some publish openly. Reading the terms before building a workflow is cheaper than rebuilding it later. For readers weighing these trade offs across the continent, choosing satellite data asia is treated at length in Ground Truth Asia, an independent English language magazine on satellite remote sensing applied to Asian territories.
Where can I download free satellite images of Asia?
Four archives cover most needs, and all four are free to search.
USGS EarthExplorer distributes the Landsat archive, which reaches back to 1972 and covers the whole of Asia. Registration is required, downloads are direct, and the collection is public domain.
Copernicus Data Space Ecosystem is the current access point for Sentinel data, replacing the earlier SciHub interface. Sentinel-1 radar is particularly useful in Asia because it sees through cloud, which matters during the summer monsoon. Sentinel-2 provides the optical series.
NASA Earthdata hosts MODIS and VIIRS products at 250 metres to 1 kilometre, suited to daily monitoring of large areas: fire, aerosol, snow cover, sea surface temperature. The coarse resolution is a deliberate trade for daily revisit.
Google Earth Engine is not an archive in the same sense but a processing platform that holds the Landsat, Sentinel, MODIS and Landsat derived collections alongside many others, and runs analysis server side. It removes the download step entirely for users who can work within its terms.
For weather rather than land, the Japanese Meteorological Agency operates Himawari 8 and 9, geostationary at 140.7 degrees east, returning a full disk image every 10 minutes at 0.5 to 2 kilometres. The data are distributed openly and are the standard source for real time cloud imagery over East and Southeast Asia.
Which Asian countries fly their own Earth observation satellites?
Several do, and the list has grown steadily since the 1990s.
Japan operates the ALOS series. ALOS-2, launched in 2014, carries the PALSAR-2 L band synthetic aperture radar, valued for interferometry and for penetrating cloud and vegetation. Japan also runs the GCOM series for global change observation and Himawari for meteorology.
India flies the IRS series, one of the longest running national programmes, and the Cartosat satellites for high resolution stereo imaging. ISRO distributes data through its own portal, with different access conditions from the open Landsat and Sentinel collections.
China operates the Gaofen series under the CHEOS programme, with resolutions from 2 metres to 16 metres depending on the satellite, plus the Fengyun meteorological series and the Haiyang ocean colour satellites.
South Korea flies KOMPSAT, also known as Arirang, with KOMPSAT-3 and 3A providing sub metre optical imagery and KOMPSAT-5 carrying X band radar.
Thailand operates THEOS, launched in 2008, a 2 metre panchromatic and 15 metre multispectral optical satellite, managed through GISTDA.
Others in the region include Vietnam with VNREDSat-1, Indonesia with LAPAN-A series and the newer satellites under BRIN, and Taiwan with FORMOSAT-5 and FORMOSAT-7. Malaysia, the Philippines and Singapore have also launched small satellites, often through university or regional cooperation programmes.
What does ground truth add to a satellite map?
A classified image is a hypothesis until someone checks it against the ground. Ground truth, in the remote sensing sense, is the set of observations collected on site that calibrates and validates what the sensor recorded. It answers questions the image alone cannot: whether a spectrally similar surface is rice or marsh, whether a dark patch is water or shadow, whether a change detection result reflects real land cover change or a difference in acquisition angle.
In practice, ground truth work in Asia often means field teams with GPS units recording land cover at sample points, then comparing those labels against the classified raster. The result is a confusion matrix and an accuracy figure, which is what makes a map defensible in a report. Without it, a classification can look convincing and still be wrong in ways that matter for policy.
Where fieldwork is impossible, analysts substitute higher resolution imagery as a reference, or use existing reference datasets. Both are weaker than a site visit, and both should be stated as such in the methods section.
How does the ACRS community share methods and results?
The Asian Conference on Remote Sensing is an annual gathering organised by the Asian Association on Remote Sensing, an Asian non governmental organisation founded in 1981 and grown from a group formed at the ERIM symposium in San José, Costa Rica, in April 1980. It counts more than 29 member countries.
The 45th edition was held in Colombo, Sri Lanka, from 17 to 21 November 2024, with more than 500 participants. The 46th took place in Makassar, Indonesia, from 27 to 31 October 2025, drawing 450 participants from 25 countries and 244 papers, of which 203 were oral and 41 posters. Proceedings from past conferences are archived on the association site, and the AJG is its associated journal.
For practitioners in the region, the proceedings are a useful record of what has been attempted with which sensors, particularly for applications where published case studies are scarce: rice yield estimation, urban heat island mapping in fast growing cities, coastal change, and monsoon flood monitoring. The methods described there often assume the same free datasets discussed above, which makes them directly reproducible.
A folio is edited before it is designed: images are ordered by intent, captions state medium and dimensions, and the page is assembled so it survives a slow connection and a small screen. The same discipline governs the portfolio entry of this reference, which treats ordering, captioning and technical checks as one job rather than three. Both practices assume a viewer who decides in seconds, so the sequence of works and the weight of each file carry the argument before any statement does.
Illustration work often ends up on a wall, and the conventions of that wall are worth knowing before a first delivery. A gallery decides what hangs, in what order, and under what light; the label beside each work carries the artist, title, year, and medium, and visitors read it before they read the image. Anyone preparing prints or framed pieces can learn how a show is assembled and what a gallerist actually does day to day by reading inside a contemporary art gallery, which follows the same question from the studio to the exhibition.