[GIS] Finding Sentinel tile for specific Long/Lat coordinate

amazon-web-serviceslatitude longitudemgrsremote sensingsentinel-2

I'm building a tool that downloads images from "Sentinel-2 on AWS"

I have a trouble translating Long/Lat coordinates to a specific Sentinel tile. I know that Sentinel 2 uses MGRS and I tried to simply convert Long/Lat to MGRS and with low precision I get the S2A tile ID. And it works, but not always.

For example, Long/Lat 34.665,31.625 resolves to "36R XA" in MGRS, but the Sentinel tile 36RXA doesn't exist.

What would be the right way of determining Sentinel tile using Long/Lat coordinates?

Best Answer

Such a tool already exists. It is called Sentinelsat and the source is available on GitHub. It offers a command line interface and a Python API. It works with Sentinel 1 and 2. The spatial query is based on a polygon and not a point, but otherwise this is exactly what you need.

EDIT: 1) you can return the product ID (or product ID list) using the query function. And from the command line, you can omit the -d option in order not to launch the download (and use the -f option to create the footprint of each image) 2) Concerning the geojson, you can see below that a wkt could be used directly instead. As mentionned by @ThingumaBob, you can create a tiny polygon based on your lat/long coordinates. But in fact, I've tested it with a point GeoJason and it also worked fine. So there is no need to convert your point into polygon. 

# search by polygon, time, and Hub query keywords
footprint = geojson_to_wkt(read_geojson('map.geojson'))
products = api.query(footprint,
                     date = ('20151219', date(2015, 12, 29)),
                     platformname = 'Sentinel-2',
                     cloudcoverpercentage = (0, 30))

which could be adjusted into 

# search by point, time, and Hub query keywords

products = api.query('POINT({0} {1})'.format(longitude,latitude),
                     date = ('20151219', date(2015, 12, 29)),
                     platformname = 'Sentinel-2',
                     cloudcoverpercentage = (0, 30))
Related Question