liminfoRemote Sensing Index Reference
Free reference guide: Remote Sensing Index Reference
About Remote Sensing Index Reference
This Remote Sensing Index Reference provides complete formulas, band mappings, and interpretation guidelines for vegetation indices (NDVI, EVI, SAVI, MSAVI2, GNDVI, LAI, NDRE), water indices (NDWI, MNDWI, NDMI, AWEI, NDSI), soil/geology indices (BSI, Clay Minerals Ratio, Tasseled Cap, PCA), and urban/thermal indices (NDBI, UI, LST, UTFVI, ISA).
Each index entry includes the mathematical formula, specific band mappings for Landsat 8/9, Sentinel-2, and MODIS sensors, value range interpretation with classification thresholds, advantages, limitations, and practical application domains.
The reference also covers satellite sensor specifications including Landsat 8/9 OLI, Sentinel-2 MSI, and MODIS band tables, atmospheric correction methods (DOS, FLAASH, Sen2Cor, LaSRC), cross-satellite band correspondence tables, and ready-to-use Google Earth Engine JavaScript code for computing multiple indices.
Key Features
- Seven vegetation indices (NDVI, EVI, SAVI, MSAVI2, GNDVI, LAI, NDRE) with formulas and value interpretation
- Five water/moisture indices (NDWI, MNDWI, NDMI, AWEI, NDSI) with threshold-based classification
- Soil and geology indices including BSI, Clay Minerals Ratio, Tasseled Cap transform coefficients, and PCA
- Urban and thermal indices: NDBI, Urban Index, LST step-by-step calculation, UTFVI heat island classification, ISA
- Complete band specifications for Landsat 8/9 OLI/TIRS, Sentinel-2 MSI, and MODIS sensors
- Cross-satellite band correspondence table mapping Landsat, Sentinel-2, and MODIS bands
- Atmospheric correction method comparison: DOS, FLAASH, Sen2Cor, and LaSRC
- Google Earth Engine JavaScript code examples for NDVI, EVI, NDWI, and NDBI computation
Frequently Asked Questions
What is NDVI and how is it calculated?
NDVI (Normalized Difference Vegetation Index) measures vegetation vigor using the formula (NIR - Red) / (NIR + Red). For Landsat 8/9, this is (Band 5 - Band 4) / (Band 5 + Band 4). Values range from -1 to +1: below 0 indicates water/clouds, 0-0.2 is bare soil, 0.2-0.4 is sparse vegetation, 0.4-0.6 is moderate vegetation, and 0.6-0.9 is dense healthy vegetation.
What is the difference between NDVI and EVI?
EVI (Enhanced Vegetation Index) addresses NDVI limitations by incorporating Blue band data and correction coefficients to reduce atmospheric aerosol effects and soil background noise. EVI also shows less saturation in dense canopies compared to NDVI. The formula is 2.5*(NIR-Red)/(NIR+6*Red-7.5*Blue+1). EVI is preferred for tropical forests and high-density vegetation analysis.
How do I detect water bodies using remote sensing indices?
Use NDWI = (Green - NIR) / (Green + NIR) for general water body detection (values > 0 indicate water). For urban areas, MNDWI = (Green - SWIR1) / (Green + SWIR1) better separates water from building shadows. AWEI provides the highest accuracy in complex environments by using multiple bands to eliminate shadow misclassification.
How is Land Surface Temperature (LST) calculated from Landsat imagery?
LST calculation involves four steps: (1) Convert DN to radiance using calibration coefficients, (2) Convert radiance to brightness temperature using sensor-specific K1/K2 constants, (3) Estimate surface emissivity from NDVI using the proportion of vegetation method, (4) Apply the emissivity correction to compute final LST in Celsius.
Which satellite sensor should I choose for my analysis?
Landsat 8/9 (30m, 16-day revisit) is best for long-term change analysis with its 40+ year archive. Sentinel-2 (10-20m, 5-day revisit) offers higher resolution and Red Edge bands for precision agriculture. MODIS (250m-1km, daily) is ideal for large-area daily monitoring. Your choice depends on spatial resolution, temporal frequency, and spectral band requirements.
What is atmospheric correction and why is it important?
Atmospheric correction removes scattering and absorption effects of the atmosphere from satellite imagery to derive true surface reflectance. Without it, index values are distorted. Methods include DOS (simplest, subtracts minimum DN), FLAASH (radiative transfer model), Sen2Cor (Sentinel-2 official tool), and LaSRC (Landsat Collection 2 standard). Level-2 products from USGS and ESA are already atmospherically corrected.
How do Landsat, Sentinel-2, and MODIS bands correspond to each other?
Key correspondences: Landsat B4 (Red) = Sentinel-2 B4 = MODIS B1; Landsat B5 (NIR) = Sentinel-2 B8 = MODIS B2; Landsat B6 (SWIR1) = Sentinel-2 B11 = MODIS B6. Sentinel-2 uniquely offers Red Edge bands (B5-B7) not available on Landsat or MODIS. Note that slight wavelength differences mean the same index may yield slightly different values across sensors.
How do I compute NDVI in Google Earth Engine?
In GEE, load a Sentinel-2 collection with ee.ImageCollection("COPERNICUS/S2_SR_HARMONIZED"), filter by date, bounds, and cloud percentage, create a median composite, then compute NDVI with s2.normalizedDifference(["B8","B4"]).rename("NDVI"). Visualize with Map.addLayer using a brown-yellow-green color palette with min:0 and max:0.8.