RGB - Bassa risoluzione

Ecosystem Servicesv02.00 credits/per_km2

Responsible:Niccolò Calandri

RGB — Sentinel-2 True Color

Overview

The Sentinel-2 RGB layer provides a true-color visual representation of the Earth's surface using the red, green, and blue spectral bands of the Sentinel-2 satellite constellation. This is the most intuitive visualization of satellite data, showing the landscape as it would appear to the human eye from space.

Spectral bands used:

B4 (Red): 665 nm, 10 m resolution
B3 (Green): 560 nm, 10 m resolution
B2 (Blue): 490 nm, 10 m resolution

The RGB composite is generated as a median composite over the requested time period, which reduces cloud contamination and provides a representative view of the surface. Contrast stretching (2nd–98th percentile) is applied to optimize visual quality across different land cover types.

RGB imagery is used for: visual site inspection, land cover validation, stakeholder communication, change detection (comparing composites from different periods), and context interpretation of other spectral indices (NDVI, LST, etc.).

Calculation Methodology

Composite generation:

Image collection — All Sentinel-2 L2A (Surface Reflectance) images within the requested time window are retrieved
Cloud masking — Pixels classified as cloud, cloud shadow, cirrus, or defective by the Scene Classification Layer (SCL) are excluded
Temporal compositing — Per-pixel median across all valid observations
Band extraction — B4, B3, B2 bands are selected and stacked as Red, Green, Blue channels
Contrast stretching — Linear stretch between the 2nd and 98th percentiles of the combined histogram
Output — 8-bit RGB GeoTIFF at 10 m resolution

Data Sources

Code	Name	Provider	Type	Coverage	Resolution	Availability
`COPERNICUS/S2_SR_HARMONIZED`	Sentinel-2 L2A Harmonized	ESA/Copernicus	Multispectral	Global	10 m	2017-03-28 — present

Documentation: Copernicus Sentinel-2

Technical Notes

Median compositing effectively removes clouds but may produce visual artifacts in areas with rapidly changing land cover
Persistent cloud cover (tropical regions, winter at high latitudes) may result in incomplete composites
Color representation depends on atmospheric correction quality; Sen2Cor L2A processing provides surface reflectance
Shadows from tall buildings or mountains may appear in composite images if shadow angles are consistent across observations
The 10 m resolution allows identification of individual buildings, roads, and field boundaries

References

Main-Knorn, M., Pflug, B., Louis, J., Debaecker, V., Müller-Wilm, U., Gascon, F. (2017). "Sen2Cor for Sentinel-2". Proc. SPIE 10427, Image and Signal Processing for Remote Sensing XXIII.
Drusch, M., Del Bello, U., Carlier, S. et al. (2012). "Sentinel-2: ESA's Optical High-Resolution Mission for GMES Operational Services". Remote Sensing of Environment, 120, pp. 25–36.

Data Sources & References

Sources

WRD_S2XXX_995d

Methodology

Le immagini RGB vengono generate a partire dalle immagini Sentinel-2 Surface Reflectance Harmonized, combinando le bande B4 (rosso, 665nm), B3 (verde, 560nm) e B2 (blu, 490nm) alla risoluzione nativa di 10 metri. Le immagini vengono filtrate per periodo e area di interesse, quindi viene applicato lo stesso mascheramento delle nuvole utilizzato per l'NDVI, basato sulla banda SCL. Per ottenere immagini cloud-free anche in aree con frequente copertura nuvolosa, viene calcolata la mediana pixel-wise delle tre bande RGB su tutto il periodo temporale. Per ottimizzare la visualizzazione, viene applicato uno stretching lineare del contrasto: per ogni banda vengono calcolati i percentili al 2% e 98%, quindi i valori vengono normalizzati linearmente nell'intervallo 0-255. L'immagine finale viene riproiettata in Web Mercator (EPSG:3857). La risoluzione viene adattata automaticamente per aree estese (12m se l'area supera 9000 m²).