QGIS Plugin for Sentinel-2

Overview

The Fabric QGIS Plugin brings satellite data processing directly into your GIS workflow. Search for locations, download Sentinel-2 imagery, and calculate forest indices—all without leaving QGIS.

What you can do:

• Download Sentinel-2 bands for any location and date range
• Calculate NDVI, NDMI, NBR, and EVI indices
• Auto-load styled results into your QGIS project
• Organize analyses in reproducible workspaces

Prerequisites

Before installing the plugin:

• QGIS 3.0+ (3.28 or later recommended)
• Docker Desktop installed and running
• 8GB RAM minimum (16GB recommended)
• 2GB disk space for Docker images

Install Docker

If you don't have Docker:

1. Download from docker.com
2. Install and launch Docker Desktop
3. Verify it's running: docker info

Installation

Step 1: Locate Your Plugins Folder

Find your QGIS plugins directory:

Step 2: Copy the Plugin

Copy the fabric-qgis-plugin folder to your plugins directory.

Step 3: Enable in QGIS

1. Restart QGIS
2. Go to Plugins → Manage and Install Plugins
3. Search for "Fabric"
4. Check the box to enable Fabric Sentinel-2

A Fabric icon appears in your toolbar.

First Run Setup

On first use, the plugin automatically:

1. Checks Docker — Prompts you to start Docker Desktop if needed
2. Pulls the Fabric image (~2GB download, one-time)
3. Builds the forest-indices image locally

First run takes 2-5 minutes. Subsequent runs start in seconds.

Using the Plugin

1. Open the Plugin

Click the Fabric icon in the toolbar, or go to Raster → Fabric Sentinel-2.

2. Set Your Location

Three ways to define your area of interest:

Search by name:

Yosemite National Park

The plugin geocodes place names automatically.

Enter coordinates:

Use current map extent: Click "Get from Current Map Extent" to use your current view.

3. Set Date Range

Select start and end dates for imagery search. The plugin finds cloud-free scenes within this window.

Tip: Wider date ranges find more imagery but may span different conditions.

4. Configure Cloud Cover

Set maximum cloud cover percentage (default: 50%).

5. Select Bands

Choose which Sentinel-2 bands to download:

Default selection (B04, B08, B11, B12) covers all common indices.

6. Select Indices

Enable "Calculate indices after download" and choose:

The plugin validates that required bands are selected.

7. Run

Click Run Fabric. Progress updates show each phase:

Downloading Sentinel-2 imagery...
Processing bands: B04, B08, B11, B12
Calculating indices: NDVI, NDMI
Loading results into QGIS...
Complete! Total time: 2m 34s

Results load automatically with a red-yellow-green color ramp.

Understanding Results

Index Interpretation

All indices range from -1 to 1:

Color Ramp

The default RdYlGn (Red-Yellow-Green) styling:

• Red: Low values (stress, damage, bare ground)
• Yellow: Moderate values
• Green: High values (healthy vegetation)

To customize: Right-click layer → Properties → Symbology.

Workspace Management

Each analysis creates a workspace folder:

~/fabric-workspaces/
└── 37p8_-119p6_20260601_20260630/
    ├── params.json          # Your settings (reproducible)
    ├── sentinel2/           # Downloaded bands
    │   ├── merged_B04.tif
    │   ├── merged_B08.tif
    │   └── report.json
    └── indices/             # Calculated results
        ├── ndvi.tif
        ├── ndmi.tif
        └── nbr.tif

Workspace naming: Auto-generated from coordinates and dates, or specify a custom name.

Reproducibility: The params.json file records exactly what you ran, enabling exact re-runs later.

Example Workflows

Forest Health Monitoring

Monitor vegetation stress in a forest area:

1. Location: Search "Black Forest, Germany"
2. Date range: June 1-30 (peak growing season)
3. Bands: B04, B08, B11
4. Indices: NDVI, NDMI
5. Analysis: Compare NDVI (greenness) with NDMI (moisture)—high NDVI but low NDMI indicates drought stress

Wildfire Damage Assessment

Assess burn severity after a fire:

1. Location: Enter coordinates of fire area
2. Date range: 1-2 weeks post-fire
3. Bands: B08, B12
4. Indices: NBR
5. Analysis: Low NBR values indicate severe burn; compare with pre-fire imagery for dNBR

Agricultural Monitoring

Track crop health through the season:

1. Location: Farm boundary coordinates
2. Date range: Single month windows through growing season
3. Bands: B04, B08
4. Indices: NDVI
5. Analysis: Create time series of NDVI to track crop development

Troubleshooting

Docker Issues

No Imagery Found

Index Calculation Failed

Performance

Typical processing times:

Disk space per workspace: 100-500MB depending on area size.

Next Steps

• Understanding Vegetation Indices — Learn the science behind NDVI, NDMI, NBR, EVI
• Forest Health Indices Pattern — Use indices in Fabric pipelines
• Writing Pipelines — Automate multi-step workflows