Sentinel-2 L1C and L2A pixel samples for band regression

This dataset contains pixels sampled from Sentinel-2 images. 198 scenes on the period going from early 2016 to the end of 2020 from 128 different MGRS tiles were used.

For each acquisition, the data was obtained at 2 processing levels: 1C
(from PEPS, CNES' mirror of Sentinel data) and 2A (from Theia's
catalogue), the latter having been produced by the MAJA processor.

For each acquisition, 100,000 pixels where sampled. Only non-saturated
pixels were selected, regardless of their cloud or shadow status. Pixel
positions were selected on the 20m resolution grid. For each 20m pixel
position, the following information was recorded:
- whether the pixel was detected as a cloud or a shadow (without
  distinction of these 2 states),
- the reflectance in the 20m bands for levels 1C and 2A,
- the reflectance of the 4 corresponding pixels of each of the 10m
  resolution bands for levels 1C and 2A,
- the reflectance at the 20m pixel position of the 60m resolution bands
  after bicubic resampling for level 1C.
- the solar and viewing angles for each pixel.

For each sampled scene, a CSV file with the name TILE_DATE_samples.csv (for example T05KRA_20171124_samples.csv) is provided.

Each row in the file corresponds to a pixel. The columns provide the following variables:
- An integer used as unique identifier of the pixel in the file.
- The tile name in TIJXYZ format.
- The date in YYYMMDD format.
- The coverage: the percentage of the tile covered by the relative orbit of the acquisition.
- The x and y integer coordinates of the 20m. resolution pixel in the array.
- The reflectances of the 4 10 m resolution pixels corresponding to the 20 m. resolution pixels. The columns are named using the format Level_Band_i with Level being L1C or L2A, and i in {1, 2, 3, 4}. For the band, we keep the ESA (L1C) and Theia (L2A) respective nomenclatures, so we have L1C_B02_1 but L2A_B2_1.
- The reflectances of the 20 m resolution bands with columns named Level_Band and the same band name conventions, so we have L1C_B05 and L2A_B5.
- The reflectances of the L1C 60m resolution bands resampled to the 20m grid. The 3 columns are named L1C_B01, L1C_B09, L1C_B10.
- The reflectances of all the bands (10m and 20m resolution bands for L1C and L2A and the 60 m resolution bands for L1C) resampled to a 60 m resolution grid. The columns are named for instance L1C_B02_60 or L2A_B7_60.
- The solar Zenith and Azimuth angles: sun_zen, sun_az:
- The sensor Zenith and Azimuth angles split into even and odd detectors (inc_even_zen, inc_odd_zen, inc_even_az, inc_odd_az). The angles values are not recorded (empty value) for the detector to which the pixel does not belong to.
- A binary value (CLM) for the cloud and cloud shadow mask (0 if the pixel is clear, 1 if it is a cloud or a cloud shadow). This information is retrieved from the L2A masks.

The list of column names is the following:
- tile
- date
- coverage
- x
- y
- L1C_B02_0
- L1C_B02_1
- L1C_B02_2
- L1C_B02_3
- L1C_B03_0
- L1C_B03_1
- L1C_B03_2
- L1C_B03_3
- L1C_B04_0
- L1C_B04_1
- L1C_B04_2
- L1C_B04_3
- L1C_B08_0
- L1C_B08_1
- L1C_B08_2
- L1C_B08_3
- L1C_B05
- L1C_B06
- L1C_B07
- L1C_B8A
- L1C_B11
- L1C_B12
- L1C_B01
- L1C_B09
- L1C_B10
- L2A_B2_0
- L2A_B2_1
- L2A_B2_2
- L2A_B2_3
- L2A_B3_0
- L2A_B3_1
- L2A_B3_2
- L2A_B3_3
- L2A_B4_0
- L2A_B4_1
- L2A_B4_2
- L2A_B4_3
- L2A_B8_0
- L2A_B8_1
- L2A_B8_2
- L2A_B8_3
- L2A_B5
- L2A_B6
- L2A_B7
- L2A_B8A
- L2A_B11
- L2A_B12
- sun_zen
- sun_az
- inc_even_zen
- inc_odd_zen
- inc_even_az
- inc_odd_az
- L1C_B02_60
- L1C_B03_60
- L1C_B04_60
- L1C_B08_60
- L1C_B05_60
- L1C_B06_60
- L1C_B07_60
- L1C_B8A_60
- L1C_B11_60
- L1C_B12_60
- L2A_B2_60
- L2A_B3_60
- L2A_B4_60
- L2A_B8_60
- L2A_B5_60
- L2A_B6_60
- L2A_B7_60
- L2A_B8A_60
- L2A_B11_60
- L2A_B12_60
- CLM

The list of available CSV files is the following:

- T05KRA_20171124_samples.csv
- T05KRA_20180329_samples.csv
- T05KRA_20180523_samples.csv
- T05KRA_20190408_samples.csv
- T05KRA_20191129_samples.csv
- T05KRA_20200402_samples.csv
- T05KRA_20200601_samples.csv
- T05KRA_20200805_samples.csv
- T05KRA_20201203_samples.csv
- T06KTF_20160324_samples.csv
- T06KTF_20171010_samples.csv
- T06KTF_20180627_samples.csv
- T06KTF_20181224_samples.csv
- T06KTF_20200616_samples.csv
- T06KTF_20200924_samples.csv
- T11SPC_20180918_samples.csv
- T11SPC_20180928_samples.csv
- T11SPC_20181013_samples.csv
- T11SPC_20181112_samples.csv
- T14SPF_20170321_samples.csv
- T14SQE_20170308_samples.csv
- T14SQE_20190402_samples.csv
- T14SQF_20180405_samples.csv
- T14SQF_20190420_samples.csv
- T18TUR_20180509_samples.csv
- T18TVS_20191026_samples.csv
- T18TXS_20171018_samples.csv
- T18UVU_20190703_samples.csv
- T18UVU_20190827_samples.csv
- T18UWU_20190327_samples.csv
- T18UXU_20200210_samples.csv
- T18UXV_20191013_samples.csv
- T18UYV_20191217_samples.csv
- T19LHH_20200925_samples.csv
- T19LHJ_20190723_samples.csv
- T19LHJ_20200218_samples.csv
- T19TCL_20170811_samples.csv
- T19TCL_20201103_samples.csv
- T20LKP_20171001_samples.csv
- T20LKP_20191125_samples.csv
- T20LLQ_20180901_samples.csv
- T20NNP_20190807_samples.csv
- T20PPC_20191103_samples.csv
- T21NYG_20191002_samples.csv
- T21NZG_20180714_samples.csv
- T22KHA_20191027_samples.csv
- T22MGB_20201216_samples.csv
- T22MHB_20200803_samples.csv
- T22NCH_20181009_samples.csv
- T23KKR_20160808_samples.csv
- T23MKS_20200830_samples.csv
- T23MLS_20191219_samples.csv
- T23MLT_20190319_samples.csv
- T23MLT_20191114_samples.csv
- T23MLT_20200313_samples.csv
- T24MWU_20190916_samples.csv
- T24MWU_20201214_samples.csv
- T24MYT_20160918_samples.csv
- T24MYT_20200425_samples.csv
- T25LBL_20161214_samples.csv
- T25LBL_20180801_samples.csv
- T25LBL_20181209_samples.csv
- T25MBM_20171214_samples.csv
- T25MBN_20200113_samples.csv
- T25MBP_20180607_samples.csv
- T28PCB_20181129_samples.csv
- T28PDU_20181101_samples.csv
- T28PEV_20170405_samples.csv
- T28PGA_20170909_samples.csv
- T28PHC_20170323_samples.csv
- T28QCD_20160513_samples.csv
- T28QFD_20171002_samples.csv
- T28QFD_20180729_samples.csv
- T29SNC_20180927_samples.csv
- T29SPR_20201222_samples.csv
- T30PUT_20200522_samples.csv
- T30QZE_20181213_samples.csv
- T30QZE_20190924_samples.csv
- T30TTM_20181128_samples.csv
- T30TYM_20200126_samples.csv
- T31PGS_20191023_samples.csv
- T31QBU_20170107_samples.csv
- T31QCU_20170308_samples.csv
- T31QEU_20180330_samples.csv
- T31SDV_20171126_samples.csv
- T31SFA_20180907_samples.csv
- T31TDL_20190903_samples.csv
- T31UFS_20170814_samples.csv
- T31UGT_20171229_samples.csv
- T32PLS_20200919_samples.csv
- T32ULD_20161224_samples.csv
- T32ULE_20181124_samples.csv
- T32UMD_20190224_samples.csv
- T33KWP_20190729_samples.csv
- T33TUJ_20201007_samples.csv
- T36SXB_20191127_samples.csv
- T36SYB_20180317_samples.csv
- T36SYB_20200902_samples.csv
- T36SYB_20200907_samples.csv
- T36SYC_20180307_samples.csv
- T36SYC_20180327_samples.csv
- T36SYD_20170804_samples.csv
- T36SYD_20190111_samples.csv
- T37SBT_20190126_samples.csv
- T37SBU_20161122_samples.csv
- T37SBU_20201022_samples.csv
- T38KMB_20170817_samples.csv
- T38KNU_20200228_samples.csv
- T38KPE_20191026_samples.csv
- T38TNL_20180328_samples.csv
- T39KTU_20201126_samples.csv
- T40KCB_20160805_samples.csv
- T40KCB_20170323_samples.csv
- T40KCB_20170820_samples.csv
- T40KCB_20171123_samples.csv
- T40KCB_20200317_samples.csv
- T40KCB_20200829_samples.csv
- T42FVL_20180319_samples.csv
- T42FVL_20180729_samples.csv
- T42FVL_20201208_samples.csv
- T42FWL_20170405_samples.csv
- T42FWL_20200812_samples.csv
- T43QHA_20170315_samples.csv
- T43QHA_20180213_samples.csv
- T43QHV_20181110_samples.csv
- T43TCG_20200925_samples.csv
- T43TEG_20181227_samples.csv
- T43TEG_20201017_samples.csv
- T43TFH_20190516_samples.csv
- T44QLG_20181016_samples.csv
- T44TKM_20180408_samples.csv
- T44TLM_20181214_samples.csv
- T44TLN_20171224_samples.csv
- T44TLN_20190123_samples.csv
- T44TLN_20190128_samples.csv
- T45QWE_20200305_samples.csv
- T45QXG_20170316_samples.csv
- T45QXG_20190510_samples.csv
- T45QYF_20190721_samples.csv
- T45RXH_20190110_samples.csv
- T45RYH_20181002_samples.csv
- T45RYH_20201105_samples.csv
- T45RYJ_20190316_samples.csv
- T45RYK_20190105_samples.csv
- T45RYK_20190301_samples.csv
- T46QBM_20190117_samples.csv
- T46QBM_20190201_samples.csv
- T46QCL_20161213_samples.csv
- T46QCL_20191014_samples.csv
- T46RBN_20181118_samples.csv
- T46RCN_20181118_samples.csv
- T46RCN_20200127_samples.csv
- T46RCQ_20200117_samples.csv
- T47QRB_20200221_samples.csv
- T47QRB_20201117_samples.csv
- T47QRC_20170820_samples.csv
- T47QRC_20180318_samples.csv
- T47QRC_20200302_samples.csv
- T47QRC_20201217_samples.csv
- T48PVQ_20170814_samples.csv
- T48PVU_20170521_samples.csv
- T48PWA_20181222_samples.csv
- T48PWB_20180928_samples.csv
- T48PWR_20200525_samples.csv
- T48QTE_20201227_samples.csv
- T49MFM_20181031_samples.csv
- T49MFN_20200109_samples.csv
- T49REQ_20190418_samples.csv
- T49RFP_20200318_samples.csv
- T49RGN_20170525_samples.csv
- T49TCF_20171004_samples.csv
- T49TCF_20190502_samples.csv
- T49TCF_20191228_samples.csv
- T49TCF_20201013_samples.csv
- T50RKT_20200429_samples.csv
- T50RMS_20170706_samples.csv
- T50RMT_20200819_samples.csv
- T50RMU_20200804_samples.csv
- T50RNV_20190606_samples.csv
- T51RTQ_20161011_samples.csv
- T51RTQ_20161230_samples.csv
- T51RTQ_20170429_samples.csv
- T51RTQ_20171205_samples.csv
- T51RTQ_20180728_samples.csv
- T51RTQ_20200831_samples.csv
- T51STR_20160504_samples.csv
- T51STR_20171021_samples.csv
- T51STR_20180728_samples.csv
- T51STR_20200801_samples.csv
- T51STR_20201020_samples.csv
- T55HCB_20191003_samples.csv
- T55HCB_20200311_samples.csv
- T55HDA_20180513_samples.csv
- T58KCC_20180527_samples.csv
- T58KEB_20170906_samples.csv
- T58KGA_20180531_samples.csv
- T58KGC_20170529_samples.csv
- T58KHB_20171018_samples.csv