Statistical API & multipart POLYGON geometry


#1

Hi !
I would like to know if Statistical API is supposed to use multipart POLYGON geometry. I tried with a real POLYGON, it works but what about multipart POLYGON (not MULTIPOLYGON)?

Because if I have to use real POLYGON for a big agricole area, I have to make a loop through all blocks and it will explode daily request limit (10 000).


#2

This is not possible for the moment, sorry, so you have to make a loop.
Your limit is 1000 per minute, so it should work OK.


#3

I tried a double request, one for ndvi and one for clouds, for 2828 blocks and from first L2A available date to today. The result json only contains 2 or 3 last dates…


#4

My first guess would be that you have a lon/lat order issue as described here:
https://www.sentinel-hub.com/faq/why-result-different-when-i-am-using-wms-or-wcs-when-coordinate-system-epsg4326
Can you check if this is the case. If not, please post here some of the requests (mask part of the instance ID).


#5

I’m using WGS84 Pseudo Mercator (EPSG 3857), not WGS84 (EPSG 4326). Is the problem the same?

This one of my two requests:

url_ndvi = 'http://services.sentinel-hub.com/ogc/fis/{0}?LAYER={1}&STYLE=INDEX&CRS=EPSG:3857&TIME={2}/{3}&RESOLUTION=10&MAXCC=100&GEOMETRY={4}'.format(instance, layer_ndvi, date1, date2, block_geom)

where:

  • layer_ndvi = NDVI (name of my layer in instance)
  • date1 = 2017-03-01
  • date2 = 2018-12-18
  • block_geom = for example: POLYGON ((-13451714.0402926 4250703.932622137, -13451681.32572064 4250775.961107044, -13451672.74855388 4250802.149914345, -13451601.7601146 4250758.394643527, -13451601.19683797 4250745.787898315, -13451606.06149972 4250738.275084979, -13451616.1225553 4250690.446595318, -13451615.14294378 4250664.590890378, -13451611.6196819 4250649.129818885, -13451605.46037447 4250622.105868792, -13451624.0373711 4250591.246873056, -13451633.55964034 4250584.518993479, -13451655.37937373 4250582.621984207, -13451674.73560679 4250590.255222793, -13451699.26485658 4250592.472739182, -13451711.64581035 4250664.034795471, -13451714.0402926 4250703.932622137))

I tried again, for a shape with 326 blocks, I get only 1150 results or 3,5 dates for a block. I checked and it’s only last dates.

Example of request answer:

{'C0': [{'date': '2018-12-18', 'basicStats': {'min': 0.189161017537117, 'max': 0.5656565427780151, 'mean': 0.3844637523861067, 'stDev': 0.05799917667988943}}, {'date': '2018-12-16', 'basicStats': {'min': 0.17397882044315338, 'max': 0.5803357362747192, 'mean': 0.3549524728940121, 'stDev': 0.07208269679348431}}, {'date': '2018-12-13', 'basicStats': {'min': 0.1798017919063568, 'max': 0.5343618392944336, 'mean': 0.35772913323243055, 'stDev': 0.057320467032393733}}]}


#6

These coordinates seem to be in United States, is this correct? If so, the response makes sense. L2A data are available globally only since 2018-12-13. Before they were only available in Europe (since March 2017).


#7

Oh yes that’s true, it’s in US. I have misunderstood the availability of L2A data. :confounded:
L2A is so cool for use Scene classification map, and have more uniform NDVI data over time as you said to me in the other post.

I give my vote for a petition in support of worldwide and over time calculation of L2A product.:raising_hand_man: