Vegetation Index Comparison Using Landsat 8 and Sentinel-2

Caspian Provinces — Multi-Sensor Analysis & Visualization

This repository contains a complete Python workflow for computing, mosaicking, stretching, and visualizing vegetation indices from Landsat 8 (OLI) and Sentinel-2 (MSI) imagery across the Caspian provinces of Iran.

The project includes:

• Multi-scene mosaicking to generate unified regional composites
• Calculation of four vegetation indices: NDVI, EVI, SAVI, MSAVI
• Percentile-based stretching (2–98%) for improved visualization
• High-resolution comparison maps (2×2 panels)
• Clean, replicable Python code

---

Project Motivation

This project aims to:

• Demonstrate how different vegetation indices respond to the same landscape
• Compare results from two widely used satellite sensors
• Provide a clear educational example for students and practitioners in remote sensing, GIS, environmental science, and agriculture

---

Study Area

The analysis focuses on the Caspian provinces of northern Iran — a region with diverse forests, croplands, and coastal ecosystems. The variability in vegetation makes it an ideal area for evaluating multi-sensor vegetation indices.

---

Satellites Used

Landsat 8 — Operational Land Imager (OLI)

• 30 m resolution
• Long-term scientific continuity (Landsat program)
• Robust spectral response for vegetation monitoring

Sentinel-2 — Multispectral Instrument (MSI)

• 10–20 m resolution
• Higher spatial detail
• Dense revisit frequency (5 days with S2A/S2B)

---

Vegetation Indices Computed

1. NDVI — Normalized Difference Vegetation Index

NDVI = (NIR - RED) / (NIR + RED)

• Most widely used vegetation index
• Measures greenness, chlorophyll abundance, and canopy density
• Range typically –1 to +1

2. EVI — Enhanced Vegetation Index

EVI = 2.5 × (NIR - RED) / (NIR + 6×RED - 7.5×BLUE + 1)

• More sensitive than NDVI in dense vegetation
• Reduces atmospheric and canopy background effects
• Useful for humid and forested regions

3. SAVI — Soil-Adjusted Vegetation Index

SAVI = (NIR - RED) / (NIR + RED + 0.5) × (1.5)

• Corrects NDVI biases in sparsely vegetated or semi-arid landscapes
• Reduces soil background interference

4. MSAVI — Modified Soil-Adjusted Vegetation Index

MSAVI = (2 × NIR + 1 - sqrt((2 × NIR + 1)^2 - 8 × (NIR - RED))) / 2

• Improved version of SAVI
• Automatically adjusts for soil effects
• Ideal for heterogeneous vegetation/soil regions

---

Workflow Summary

1. Data Preparation

• Load regional boundary shapefile
• Filter Landsat 8 & Sentinel-2 collections (2024 growing season)
• Apply cloud cover masking and geometric clipping

2. Vegetation Indices

• Compute NDVI, EVI, SAVI, MSAVI
• Generate combination index (average of all indices)
• Create median composites across time series

3. Data Export

• Export median composites → GeoTIFF format
• Transfer to local storage for processing
• Mosaic multiple tiles into unified rasters

4. Visualization

• Apply 2–98% percentile stretching for optimal contrast
• Overlay on OpenStreetMap basemaps
• Create 2×2 comparison panels (Landsat vs Sentinel-2)

5. Analysis & Output

• Produce multi-sensor comparison maps
• High-resolution PNG exports (300 DPI)

---

Comparison Results