Spectral Analysis

Spectral Index

A mathematical combination of two or more spectral bands designed to highlight a specific surface property. Common indices include NDVI (vegetation), NDWI (water), and NDBI (built-up areas). Indices normalize band ratios to reduce illumination and atmospheric effects.

Overview

A spectral index is a mathematical combination of two or more spectral bands designed to enhance a specific feature of the Earth's surface while suppressing confounding factors. By ratioing, differencing, or normalizing band values, spectral indices amplify subtle spectral signals that would be difficult to detect in individual bands. Over 200 published spectral indices exist, covering vegetation health, water detection, soil exposure, urban mapping, fire damage, and mineral identification.

How It Works

Most spectral indices use the normalized difference formula: (Band1 − Band2) / (Band1 + Band2). This normalization produces values between -1 and +1, canceling out variations in overall brightness caused by illumination differences, topography, and atmospheric effects. This makes indices more reliable than raw band values for comparing across scenes, dates, and sensors.

Beyond normalized differences, indices may incorporate soil adjustment factors (SAVI), atmospheric correction coefficients (EVI), or multiple bands simultaneously (BSI). The choice of bands determines what the index measures — red and NIR for vegetation (NDVI), green and NIR for water (NDWI), SWIR and NIR for moisture (NDMI) or built-up areas (NDBI).

Key Facts

  • Over 200 spectral indices have been published in the remote sensing literature.
  • The normalized difference formula — (B1-B2)/(B1+B2) — is used by NDVI, NDWI, NDBI, NDMI, NBR, and many others.
  • Indices are more reliable than raw band values because normalization reduces illumination and atmospheric effects.
  • The Index DataBase (IDB) at indexdatabase.de catalogs hundreds of published spectral indices with formulas and references.
  • Choosing the right index depends on the target feature, sensor bands available, and environmental conditions.

Applications

Vegetation Monitoring

NDVI, EVI, SAVI, GNDVI, and other vegetation indices quantify plant health, density, and phenology from space.

Water Resource Management

NDWI and MNDWI detect and map surface water bodies, floods, and drought conditions.

Urban and Land Use Analysis

NDBI and UI (Urban Index) highlight built-up areas for urban sprawl monitoring and planning.

Fire and Disaster Assessment

NBR (Normalized Burn Ratio) and dNBR quantify fire severity and burn scar extent for post-fire recovery planning.

Limitations & Considerations

No single index captures all surface properties — each is designed for a specific purpose and can be misleading when applied outside its intended domain. Indices that use the same bands in different environments may require different interpretation thresholds. Atmospheric correction is still important despite normalization. Indices compress multi-dimensional spectral information into a single value, inevitably losing information. Some indices are sensitive to sensor-specific band positions and widths, complicating cross-sensor analysis.

History & Background

The first widely used spectral index was the Simple Ratio (NIR/Red), followed by NDVI in 1973 (Rouse et al.). The 1980s-90s saw proliferation of soil-adjusted (SAVI, 1988), enhanced (EVI, 1990s), and water-specific (NDWI, 1996) indices. The Index DataBase now catalogs hundreds of indices. Despite the enormous variety, NDVI remains the most widely used index in the history of remote sensing due to its simplicity and universality.

Analyze Spectral Index data with LYRASENSE

Use our agentic notebook environment to work with satellite data and compute indices like Spectral Index — no setup required.