24
A few months ago I wrote a technical blog post (Planespotting) on intra-detector parallax effects in Sentinel-2 imagery, which can cause aircraft contrails to appear as rainbow stripes. The post also discusses inter-detector parallax effect and motion effects, which also can cause color shifts.
Here is a summary of intra-detector spectral band parallax:
...
14
For images of the same location but different dates, I would rather talk about compositing than mosaicing (which combines images from different extents into a larger image). You will find a lot of details if you search "compositing" keyword, but here is a short summary:
There are two main approaches for the compositing of time series:
Best available pixel ...
13
Sentinel-2 Level 1C data are expressed in reflectance with a scaling factor, not in radiance.
You have to divide by 10000 to get the reflectance. In the preliminary products shown this autumn, you had to divide by 1000.
The scaling factor is given in the xml file at the root of the product directory.
<QUANTIFICATION_VALUE unit="none">10000</...
13
ESA's Sen2cor Toolbox removes haze, cirrus and cloud shadows, not dense clouds. It's performed over the same scene, not by mosaicking.
The AC flow is the following from L2A algorithms document:
So, Sen2cor performs an image classification and use this layer to select where to correct pixel values. You can check documentation to know which equation is used ...
13
You are using the variable NDVI to compute the chart, but this variable is a single image, because you applied median(). If you want a time series plot you have to compute the chart using the NDVI collection before calculating median.
Therefore, I just change the name of the median
//Import GEE Feature Collection (Somaliland kml)
var geometry = ee....
12
The Sentinel 2 L1C images started out as 12-bit, but that has been changed in early 2016 when ESA changed QUANTIFICATION_VALUE from 1000 to 10000. Now L1C is encoded as an UINT16 jp2 file with (at least?) 15 significant bits. This is from gdalinfo on a recent B02.jp2 file:
Band 1 Block=1024x1024 Type=UInt16, ColorInterp=Gray
Overviews: 5490x5490, ...
12
10m resolution for building footprints is optimistic at best. Most buildings aren't that all that big (20m by 20m is a fairly big house), and they are also very inconsistently shaped and coloured. All in all, you chose a fairly difficult target and a dataset not entirely suited for the application.
The whole problem that you are looking at is much better ...
11
First of all, you should take note that the simulated Sentinel-2 images come from the SPOT 4 (and maybe SPOT 5) satellite. So you'll have less bands than with the actual Sentinel-2. For SPOT 4 and 5, the bands are
B1 = green, B2=red, B3=NIR, B4=SWIR
Second, NDVI is more often (NIR-RED)/(NIR+RED), and you should make sure that you have reflectances in ...
11
If you have a look at Sentinel 2 bands you'll find that all bands are type unsigned int 16 except for QA20 that is unsigned int 32, that band is causing the issue.
You can solve it in different ways, but I post here what I think is what you need:
var image = ee.ImageCollection('COPERNICUS/S2')
.filterBounds(PuntoQuellaveco1)
....
10
I received help from an application developer at JNCC. I will post their answer here to help others.
My problem was that I needed to escape the $ character before value. so the wget command should read (using the apihub, which you could replace with dhus):
wget --no-check-certificate --user=username --password=usrpass "https://scihub.copernicus.eu/apihub/...
10
There are known issues with Sentinel-2 Level 1-c data being misaligned. The best way to correct this misalignment is with the Shift tool, if the shift is in the whole granule.
Sometimes though, the pre-processing error exist in one or some of the strips which are stitched together.
I'd recommend that you inspect all areas of you image (and all other ...
10
I believe you are using the wrong bucket name.
It should be s3://sentinel-s2-l1c, not s3://sentinel-pds
Try:
url = 's3://sentinel-s2-l1c/tiles/10/S/DG/2015/12/7/0/B01.jp2'
See more info here:
http://sentinel-pds.s3-website.eu-central-1.amazonaws.com/
and
https://aws.amazon.com/public-datasets/sentinel-2/
10
Each Sentinel-2 datatake is cut in pieces to be smaller in size and easier to handle. The data is cut to match the Military Grid Reference System. The pieces are called granules or tiles.
For Level-1C and Level-2A, the granules, also called tiles, are 100x100km2 ortho-images in UTM/WGS84 projection. [...]
Products are a compilation of elementary ...
9
An alternate source is the Sentinel-2 Amazon S3 archive, where you can search fast and preview images (without need for login). This one is preferred if you're interested in individual granules and not the whole tile file (which is few GBs)
http://sentinel-pds.s3-website.eu-central-1.amazonaws.com
A Tip: the date slider has a minimum of 25/11/2015 but if ...
9
You can download Sentinel 2 Imagery by going to the Sentinel Data Access Portal and selecting the Sentinel Data Hub
Once you have arrived here select Scientific Hub
In the top right hand corner select Sign Up (Enter all your info and verify the email they send you)
Login to your new account. Begin your search by using the search bar and type in "Nigeria". ...
9
The latest version of GDAL (2.1) has the ability to read Sentinel 2 data (see http://gdal.org/frmt_sentinel2.html) in the SAFE file structure. It reads as sub-datasets so you need to run gdalinfo first on the main XML file to show all subdatasets (one for each resolution and UTM zone) then to select the one you want. You can then create a GeoTiff from the ...
9
I can provide three elements of the design of Sentinel-2 spectral band resolutions :
there is a data volume/ Data transmission rate issue : the data volume gathered by the Sentinel-2 is huge : 1 TB per day for each Sentinel-2. I guess the mission designers decided to provide the most frequently used bands at 10m, and the other bands at larger resolution. ...
9
The closest you can get to that are EOX's cloudless map tiles:
https://s2maps.eu/
These are not latest, but they are global, cloudless and they are free (CC BY 4.0). WMS and WMTS options available.
The problem with "the latest" imagery is that it will often be useless due to clouds. This makes it difficult to justify the processing and storage costs as ...
8
Two issues with your script:
1) A single image works best when you use ee.Image instead of ee.ImageCollection.
2) You switched up the granule name. The first part is the sensing time and the second the ingestion time. An acquisition taken in 2016 cannot be ingested in 2015. It should be 20150828T110656_20160412T015159_T30SVG not the other way round.
As a ...
8
There are 2 parts of the problem. The first is that you want to convert from 16 bits to 8bit, and the -scale option of gdal_translate does it, as mentioned in the previous answer.
-scale minOriginal maxOriginal minOutput maxOutput
The second problem is a contrast enhancement problem: when you rescale, you want to have a high contrast for the pixels ...
8
Here is an example of cloud masking for Sentinel-2:
var s2 = ee.ImageCollection("COPERNICUS/S2");
var pt = ee.Geometry.Point([-122.41653442382812, 37.77180027337861]);
var cloudyImage = ee.Image('COPERNICUS/S2/20151207T190417_20151207T221905_T10SEG');
Map.centerObject(cloudyImage, 10);
Map.addLayer(cloudyImage, {bands: ['B4', 'B3', 'B2'], max: 2000}, '...
7
The original answer is out of date. Don't use this for installing gdal now. The latest stable version is 2.2 and is easily installed from gisinternals.com or via conda.
Original out of date answer
Rather than building as a plugin for an older release, you could download and use a GDAL 2.1 development snapshot build from GISinternals that contains the new ...
7
A good alternative to the official Scihub is the mirrored Sentinel-2 data on Amazon Web Services.
Sentinel-2 on AWS
This has the advantage of better uptime and the products are already saved in their MGRS tiles, which makes downloading a lot faster.
The data is stored in a public bucket with the scheme tiles/[UTM code]/latitude band/square/[year]/[month]/[...
7
You'll need to run ESA's Sen2Cor algorithm to process Sentinel-2 Level-1C data to Level-2A.
This will give you atmospherically corrected BOA data. If you then divide the values by the quantification value (10000) you'll end up with BOA reflectance data.
Download and install Sen2Cor: http://step.esa.int/main/third-party-plugins-2/sen2cor/
STEP help forum (...
7
Sentinel 2 L1C reflectances are multiplied by 10000. Therefore you should adapt your equation in order to take the into account (or you can divide all values by 10000, but I wouldn't do this). On the other hand, it is worth noting that, while most values will ly between -1 and 1, EVI could mathematically be out of this range.
(2.5 * float("Band 8" - "Band ...
7
That's because these bands come as unsigned integer 16 so the numpy division returns only positive integers.
You can replace ndvi = (NIR-RED)/(NIR+RED) by ndvi = (NIR.astype(float) - RED.astype(float)) / (NIR+RED)
7
In the Earth Engine public data catalog, Sentinel-2 data is stored in the "COPERNICUS/S2" image collection. The naming convention for the images is described on the Sentinel-2: MultiSpectral Instrument (MSI), Level-1C dataset description page:
Each Sentinel 2 product (zip archive) contains multiple granules. Each
granule becomes a separate Earth Engine ...
7
Here's an example that hopefully does something like what you want, but you'll need to modify it to meet your requirements for "manually selected marker points" and Sentinel 2.
In general, do NOT use for-loops or getInfo(). To be clear, DO NOT USE FOR LOOPS unless you have a very good reason to be doing so. If you don't know, use map() for all the ...
7
You can simply use the TCI.jp2 file that its included in the SAFE.zip files.
Note that these files are not available in S2 files before October 2016
Alternatively you can convert the bands using GDAL:
# Merge bands
gdalbuildvrt -separate TCI.vrt B04.jp2 B03.jp2 B02.jp2
# Convert to uncompressed GeoTiff
gdal_translate -ot Byte -co TILED=YES -scale 0 4096 0 ...
7
As far as I know, Sentinel 2 has no computed cloud-shadow mask, so I ommit that part. But for the rest, I think you should at least mask out clouds, and filter a little.
I've made public a repo I have to perform this: users/fitoprincipe/geetools
First the core:
var computeQAbits = function(image, start, end, newName) {
var pattern = 0;
for (var i=...
Only top voted, non community-wiki answers of a minimum length are eligible