Skip to content

S2-Cloudless service specifications


icon


Service Description

The Sentinel-2 Cloudless processor (S2-Cloudless) processing service provides cloud masks and probabilities computed using the s2cloudless algorithm developed at Sinergise1. They are available as Sentinel-2 bands named CLP (cloud probabilities) and CLM (cloud masks). The classification is based on a single-scene pixel-based cloud detector developed by Sentinel Hub's research team and is described in more detail in the Sentinel Hub blog2. The s2cloudless algorithm was part of an international collaborative effort aimed at intercomparing cloud detection algorithms. The S2-Cloudless algorithm3 was validated together with 9 other algorithms on 4 different test datasets and in all cases found to be on the Pareto front4. This service supports Sentinel-2 L1C products.

Warning

S2-Cloudless currently supports only L1C Datasets from the Sentinel-2 EO mission.

Inputs

The S2-Cloudless service requires as input a calibrated Dataset (L1C) from Sentinel-2 missions.

Parameters

The S2-Cloudless service requires a specified number of mandatory and optional parameters. Table 1 describes the service parameters.

Parameter Description Required Default value
Input product reference A reference to a Sentinel-2 Level-1C calibrated dataset. YES
Cloud probability threshold to mask cloudy pixels This parameter specifies the cloud probability threshold. All pixels with cloud probability above this threshold are masked as cloudy pixels. Default value is 0.4. YES 0.4
Size of the disk in pixels to perform convolution Size of the disk in pixels for performing convolution (averaging probability over pixels). Default value is 4. Value 0 means do not perform this post-processing step. YES 4
Size of the disk in pixels to perform dilation Size of the disk in pixels for performing dilation (averaging probability over pixels). Default value is 2. Value 0 means do not perform this post-processing step. YES 2

Table 1 - Service parameters for the S2-Cloudless processor.

Output

The result products of the S2-Cloudless service are two GeoTiffs (COGs):

  • Cloud Mask (CLM): 0 (no clouds), 1 (clouds), 255 (no data)

  • Cloud Probability (CLP): 0–255 (divide by 255 to get to the [0-1] range)

The CLM no data value of 255 is also returned if a tile has missing CLM and CLP bands, for example due to errors. This ensures that values of 0 and 1 can be used with confidence for each pixel. CLP will in such a case return 0.

S2-Cloudless CLM and CLP product specifications can be found in below Tables.

Attribute Value / description
Long Name S2-Cloudless Cloud Mask
Short Name s2cloudless-clm
Description CLM binary mask: 1=clouds, 0=no-clouds
Data Type UnSigned 8-bit Integer
Band Single
Format COG
Projection Native as Sentinel-2 L1C product
Units N/A
Valid Range [0 - 1]
Fill Value 255
Attribute Value / description
Long Name S2-Cloudless Cloud Probability map
Short Name s2cloudless-clp
Description Cloud probability in a [0-255] range
Data Type UnSigned 8-bit Integer
Band Single
Format COG
Projection Native as Sentinel-2 L1C product
Units N/A
Valid Range [0 - 1]
Fill Value 255

  1. Sentinel Hub Blog (2017), Improving Cloud Detection with Machine Learning, available at https://medium.com/sentinel-hub

  2. Sentinel Hub's cloud detector for Sentinel-2 imagery in python (2018), source code available in Github at https://github.com/sentinel-hub

  3. Sentinel Hub Blog (2020) Sentinel Hub Cloud Detector, available at https://medium.com/sentinel-hub

  4. Skakun S. et al, (2022) “Cloud Mask Intercomparison eXercise (CMIX): An evaluation of cloud masking algorithms for Landsat 8 and Sentinel-2”, Remote Sensing of Environment, Volume 274, 2022, 112990, ISSN 0034-4257, DOI: 10.1016/j.rse.2022.112990