BUG: Check and fix the gain/offset to compute the reflectance of optical products

[remi-braun]

Check for other optical products if the gain and offset for reflectance computation is correctly handled:

• Sentinel-2 L2A & L1C
  • PB < 04.00: ✅ divide by 10000
  • PB > 04.00: ✅ Use coefficients in the metadata (should be bias = -1000 and gain =1/10000)
• Sentinel-2 Theia
  -> ✅ divide by 10000
• Landsats
  -> ✅ Use the coefficients in metadata
• Sentinel-3 OLCI and SLSTR
  -> ✅ Use the coefficients in metadata
• DIMAP (Pleiades, Pleiades-Neo, SPOT-6 and 7):
  -> ✅ divide by 10000 (gain and offset are used to compute the radiance)
  -> ⚠️ Check if the product is delivered in Reflectance Radiometric Processing (should me, most of the time)
• Maxar (GeoEYE, WV)
  • ARD (not handled by EOReader): divide by 10000
  • Others: ⚠️Convert to radiance then to reflectance (formula and last coefficients)
• PlanetScope
  • ⚠️ Multiply DN by reflectanceCoefficient from metadata
• Vision-1
  -> ⚠️ provided in TOA radiance (uint16) !
  Divide by 100 to get scaled radiance, then apply the formula to convert to reflectance

✅ : was OK
⚠️: to fix !

[remi-braun]

• Add an attribute to the spectral arrays saying they are in reflectance.

[remi-braun]

• Create a function scaling raw data and maybe a keyword allowing the user to keep raw values.