Level 1b Processing

Level-1B processing consists of measurement calibration, time domain averaging of SWIR channels, re-gridding into co-located along track and nadir images, cosmetic filling of all missing pixels, classification (surface, clouds) of all pixels and computation of meteorological annotations on the tie-points grid.

It is divided into eight main steps:

  1. Radiance scaling aims to calibrate numerical counts associated with each pixel using calibration coefficients derived during Level-1A processing for the thermal IR channels, and those provided as an Auxiliary Data File (ADF) (derived from VISCAL data of the previous orbit) for visible, NIR and SWIR channels.
  2. Re-gridding derives calibrated SLSTR pixels into co-located nadir and oblique images, onto a 1 km grid for thermal IR channels or a 0.5 km grid for visible, NIR and SWIR channels, using the pixel positions derived previously.
  3. Cosmetic filling takes care of missing image pixels, as well as nadir scan and solar and viewing angles necessary for cloud flagging.
  4. Surface classification uses the image pixel latitude and longitude to derive the surface type for each image pixel.
  5. Cloud identification identifies image pixels as cloudy or cloud-free.
  6. Reflectance to radiance conversion Since solar channels are calibrated with respect to the reflectance of the VISCAL diffuser, and cloud flagging algorithms use thresholds defined in reflectance, the conversion from reflectance to radiance should be carried out at the end of the processing.
  7. Meteorological annotations computationinterpolates meteorological parameters provided in ECMWF auxiliary files onto a sub-grid of tie-points. It should be noted that before this step, an interpolation is done on the tie-point grid to go from a 16 km resolution to a 1 km resolution in the along-track direction (to comply with the continuity requirements). It should also be noted that the same kind of algorithm is used in OLCI Level-1B processing.
  8. Further cloud identification is performed using Bayesian and Probabilistic techniques.