Spectral Indices used in Remote Sensing

When processing remotely sensed data, such as satellite imagery or aerial photography, a range of spectral index equations can be applied to gain useful information which can’t be seen from the ground or using normal visible-spectrum or RGB imagery.

Satellite images and most aerial imagery contains more than just the visible spectrum (red, green and blue) – depending on the sensor, the data may contain additional measurements in parts of the spectrum which are invisible to the human eye. For example, the near-infrared part of the spectrum sits slightly “above” red, with wavelengths in the region of 750-950 nanometres, while the human eye can only respond to wavelengths between about 390nm (blue) and 700nm (red). Measurements from each part of the spectrum (e.g. blue, green, red, near-infrared) are stored in separate bands, or layers, in the image.

Spectral indices can be calculated using the values for two or more of the spectral bands, with the result being a number within a predictable range. For example, the Normalised Difference Vegetation Index (NDVI) is calculated using the red and near-infrared (NIR) bands, and always results in a value between -1 and 1.

Each spectral index has been demonstrated in peer-reviewed scientific research to correlate with a specific physical variable, such as leaf chlorophyll content or soil moisture levels. Using NDVI as an example, high values generally correlate with dense and/or healthy vegetative biomass, with low values indicating bare soil, rock, sand or snow.

The spectral index value across the mapped area can be represented using a colour scale, with NDVI commonly being mapped using red for the highest values, then orange, yellow, green, blue and finally purple for the lowest values.

Below is a brief list of spectral indices which are commonly applied to satellite imagery and other remotely sensed data. All of the below indices are used by Eden PA in satellite analysis and monitoring of crops, pastures, orchards and vineyards, along with a range of satellite environmental management and grounds maintenance applications. Some indices are only available in certain resolutions, depending on which spectral bands Eden PA can acquire from different imagery providers. Other indices are also available in addition to this list, and custom analysis services can also be provided – contact Eden PA to discuss your requirements.

IndexDescription & EquationApplications & NotesSource
ARVIAtmospherically Resistant Vegetation Index
ARVI: Atmospherically Resistant Vegetation Index
Vegetation analysis;
Resistant to atmospheric effects
Kaufman & Tanre, 1992
ARVI2Atmospherically Resistant Vegetation Index 2
ARVI2: Atmospherically Resistant Vegetation Index 2
Vegetation analysis;
Resistant to atmospheric effects;
Higher dynamic range than NDVI
Bannari et al., 1995
BAIBurn Area Index
BAI: Burn Area Index
Assessment of crop/pasture damage after fireChuvieco et al., 2002
BNDVIBlue Normalised Difference Vegetation Index
BNDVI: Blue Normalised Difference Vegetation Index
Vegetation analysis;
Strong correlation with LAI
Wang et al., 2007
BWDRVIBlue Wide Dynamic Range Vegetation Index
BWDRVI: Blue Wide Dynamic Range Vegetation Index
Vegetation analysis;
Strong correlation with LAI;
Wide dynamic range
Hancock & Dougherty, 2007
CCCICanopy Chlorophyll Content Index
CCCI: Canopy Chlorophyll Content Index
Vegetation analysis;
Correlation with chlorophyll content
Barnes et al., 2000
CVIChlorophyll vegetation index
CVI: Chlorophyll vegetation index
Vegetation analysis;
Correlation with chlorophyll content
Vincini et al., 2008
DVIDifference Vegetation Index
DVI: Difference Vegetation Index
Vegetation analysis;
Can be used to distinguish vegetation from soil;
Will produce errors if shadows are present
Tucker, 1979
EVIEnhanced Vegetation Index
EVI: Enhanced Vegetation Index
Vegetation analysis;
Useful in high LAI areas, where NDVI may saturate
Heute et al., 2002
NDVINormalised Difference Vegetation Index
NDVI: Normalised Difference Vegetation Index
Vegetation analysis;
Sensitive to chlorophyll;
Correlation with dry total green biomass
Rouse et al., 1974