Sentinel-3 Level-2 Optical Processor

Overview

A lightweight Earth Observation data pipeline that processes Sentinel-3 OLCI Top-of-Atmosphere radiance with standard Python packages into Level-2 surface reflectance, aerosol-corrected composites, and Normalized Difference Vegetation Index (NDVI).

Example Outputs

Showing true color with aerosol correction (left) and the NDVI visualization (right).

Key Features & Physics Implemented

• Atmospheric Scattering (Rayleigh): Calculates real-time Rayleigh optical thickness using the Single-Scattering Approximation (SSA)
• Orbital Normalization: Dynamically calculates the Day of the Year from satellite metadata to apply Earth-Sun inverse-square solar flux corrections.
• Spatial Interpolation: Upscales coarse tie-point geometry and meteorological grids to full sensor resolution using bivariate spline interpolation
• Aerosol Correction: Implements a dynamic Near-Infrared (NIR) Dark Object Subtraction (DOS) for quick corrections.
• Ozone Absorption Correction: Uses ozone data to compute and apply gaseous transmittance corrections
• Cloud Masking: Automated masking module to detect clouds and mask pixels using configurable NDVI and near-infrared brightness thresholds.
• .nc Export: Outputs both .png composites and analysis-ready .nc (NetCDF4) datasets preserving full precision.

Installation

Clone the repository and install the required dependencies:

git clone https://github.com/drice987/sentinel3_l2_processor.git
cd sentinel3_l2_processor
pip install -e .

Usage

1. Configure the Pipeline Edit the config.yaml file to point to your unzipped Sentinel-3 L1B folder and select your processing mode.

input:
  folder_path: "path/to/SEN3"

processing:
  mode: "rgb"  # Options: 'rgb', 'aerosol_rgb', 'ndvi'
  rgb_bands: [7, 6, 4] # Corresponds to OLCI 620nm, 560nm, 490nm
  
  cloud_masking:
    enabled: true
    brightness_threshold: 0.5
    ndvi_cloud_min: -0.05
    ndvi_cloud_max: 0.15

visualization:
  generate_plot: true
  gamma: 1.5
  ndvi_cmap: "RdYlGn"

2. Run the Processor Execute the master script from the terminal:

# Uses default config.yaml by default:
s3-process

# Or pass a custom configuration file:
s3-process path/to/custom_config.yaml

Processing Modes

• rgb: Applies Rayleigh scattering correction based on local solar/viewing geometries and atmospheric pressure.
• aerosol_rgb: Applies Rayleigh correction followed by a dynamic NIR Dark Object Subtraction to strip aerosol haze over water bodies.
• ndvi: Calculates the Normalized Difference Vegetation Index using Rayleigh-corrected Red and Near-Infrared bands.

---

Developed for Earth Observation data engineering and optical physics applications.