[GIS] Sentinel-2 cloud mask with fmask

cloud coverremote sensingsentinel-2

I'm trying to run fmask on a Sentinel-2 granule.
I downloaded the 3.2 version of Windows stand alone software for Sentinel 2 from https://github.com/prs021/fmask
When I try to run fmask by calling the FmaskMSI.exe through cmd from the folder containing the Sentinel-2 Granule (or from within the granule folder) i get the following error:

No L*MTL.txt header in the current folder!
Output argument "clr_pct" (and maybe others) not assigned during call to "autoFmask".
MATLAB:unassignedOutputs

The L*MTL.txt suggests that its's looking for a Landsat type product but i specifically point to the Sentinel-2 installation of fmask

How do i get fmask to run for Sentinel-2 data?

Best Answer

You can use the cloud mask .gml file provided by ESA with the archive you download from the hub or other vendors. You only need to convert the .gml file into .shp. I use GDAL's ogr2ogr to do this, a link for this conversion is here: Using ogr2ogr to convert GML to shapefile in Python?

Related Solutions

[GIS] Removing cloud from Sentinel 2A imagery

In optical remote sensing in the visible spectrum you cannot see through clouds. So there is nothing you can do, except to wait for images without clouds.

Cloud masks are (as far as i know) used to exclude clouded areas from (for example) landcover classification, because results there would be incorrect anyways.

edit

As Aaron mentioned, you can sometimes use a cloud mask to insert cloud-free imagery into a cloudy image, but that only works if you a) do have cloud free imagery and b) you don't care that the resulting image is now a mosaic of images from different acquisition dates, and thus you cannot compute change detection etc. on it.

[GIS] Add a ‘cloud’ band to Sentinel 2 images (GEE)

If you want the value of 'cloud' to be 1 (true) do:

// Sentinel 2 cloud function 
var cloudfunction_ST2 = function(image) {
  // If band 4 is higher than 2500, the pixel is considered as cloudy.
  var b4 = image.select("B4");
  // Get pixels above the threshold.
  var cloud = b4.gt(2500).select([0], ["cloud"]);
  // Create a mask from high likelihood pixels.
  var cloudmask = image.mask().and(cloud.not())
  image = image.updateMask(cloudmask);
  // Mask those pixels from the image.
  return image.addBands(cloud)
};

Otherwise:

// Sentinel 2 cloud function 
var cloudfunction_ST2 = function(image) {
  // If band 4 is higher than 2500, the pixel is considered as cloudy.
  var b4 = image.select("B4");
  // Get pixels above the threshold.
  var cloud = b4.gt(2500);
  // Create a mask from high likelihood pixels.
  var cloudmask = image.mask().and(cloud.not()).select([0], ["cloud"])
  image = image.updateMask(cloudmask);
  // Mask those pixels from the image.
  return image.addBands(cloudmask)
};

Then, when you go ImageCollection.map(cloudfunction_ST2) you'll get an extra band called 'cloud' in every image.

Remember you have a band called 'QA60' which is the bit mask band with cloud mask information.

Best Answer

Related Solutions

[GIS] Removing cloud from Sentinel 2A imagery

[GIS] Add a ‘cloud’ band to Sentinel 2 images (GEE)

Related Question