138
B.G.H. Gorte
line. The difference index shown above is a perpendicular index. In flat terrain with
uniform illumination the advantage might be that the results are less dependent on
soil conditions than those of ratio indices. After normalizing the data into the the
range of reflectances between 0 and 1, DVI values are between -1 and 1, although
negative values (below the soil line) are not supposed to occur.
7.3.2 Normalized Difference Vegetation Index (NDVI)
A well-known index is the Normalized Difference Vegetation Index (NDVI), defined
as
NDVI= NIR-R
NIR+R
(7.3)
Despite its name, NDVI is a ratio index, with isolines through the origin. NDVI
results are independent of illumination. The range is theoretically between -1 and 1,
but in practice the value of R is hardly ever larger that the one for NIR (Fig. 7.2),
which reduces the range to [0 .. 1].
Figure 7.3 shows a near infrared band of a multi-spectral SPOT image, followed
by three vegetation indices: DVI, RVI and NDVI. The near infrared band shows
illumination differences due to the topography. These differences are still visible in
the DVI, whereas they do not affect RVI and NDVI.
Many studies have shown a good correlation between NDVI and amount of vegetation. The results have proven to be useful in a large range of mapping scales,
based on imagery with very different spatial resolutions, from SPOT-XS (20 m) to
NOAA-AVHRR (1-4 km). The latter allows for the creation of global NDVI coverages on a regular basis, which enables scientists to study change in the biosphere in
detail. whereas the former allow detailed studies on catchment scales.
It should be noted, that the information obtained from NDVI data is mostly used
qualitatively, in relation to other NDVI data, e.g. to monitor vegetation over time,
or to cQmpare different regions with similar characteristics. Documentation of the
FEWS (Famine Early Warning System) project ofthe U.S. Agency for International
Development (US AID) (FEWS Bulletin, 1996), which is based on NOAA satellite
imagery, states:
The magnitude of NDVI is indicative of the level of photosynthetic
activity in the vegetation being monitored. ( ... ) FEWS receives an NDVI
image for the entire continent of Africa three times a month. ( ... ) The basic
NDVI image gives an indication of where vegetation is present and active,
but has limited value beyond that. More information is regularly gained by
comparing a ten-day period (dekad) against 'normal' conditions for that
time of year. In general terms, persistent worse-than-average conditions
would be one piece of information in the FEWS "convergence of evidence"
approach. However, it must be stressed that a satellite image alone does not
give a definitive answer on food-security conditions on the ground.
B.G.H. Gorte
line. The difference index shown above is a perpendicular index. In flat terrain with
uniform illumination the advantage might be that the results are less dependent on
soil conditions than those of ratio indices. After normalizing the data into the the
range of reflectances between 0 and 1, DVI values are between -1 and 1, although
negative values (below the soil line) are not supposed to occur.
7.3.2 Normalized Difference Vegetation Index (NDVI)
A well-known index is the Normalized Difference Vegetation Index (NDVI), defined
as
NDVI= NIR-R
NIR+R
(7.3)
Despite its name, NDVI is a ratio index, with isolines through the origin. NDVI
results are independent of illumination. The range is theoretically between -1 and 1,
but in practice the value of R is hardly ever larger that the one for NIR (Fig. 7.2),
which reduces the range to [0 .. 1].
Figure 7.3 shows a near infrared band of a multi-spectral SPOT image, followed
by three vegetation indices: DVI, RVI and NDVI. The near infrared band shows
illumination differences due to the topography. These differences are still visible in
the DVI, whereas they do not affect RVI and NDVI.
Many studies have shown a good correlation between NDVI and amount of vegetation. The results have proven to be useful in a large range of mapping scales,
based on imagery with very different spatial resolutions, from SPOT-XS (20 m) to
NOAA-AVHRR (1-4 km). The latter allows for the creation of global NDVI coverages on a regular basis, which enables scientists to study change in the biosphere in
detail. whereas the former allow detailed studies on catchment scales.
It should be noted, that the information obtained from NDVI data is mostly used
qualitatively, in relation to other NDVI data, e.g. to monitor vegetation over time,
or to cQmpare different regions with similar characteristics. Documentation of the
FEWS (Famine Early Warning System) project ofthe U.S. Agency for International
Development (US AID) (FEWS Bulletin, 1996), which is based on NOAA satellite
imagery, states:
The magnitude of NDVI is indicative of the level of photosynthetic
activity in the vegetation being monitored. ( ... ) FEWS receives an NDVI
image for the entire continent of Africa three times a month. ( ... ) The basic
NDVI image gives an indication of where vegetation is present and active,
but has limited value beyond that. More information is regularly gained by
comparing a ten-day period (dekad) against 'normal' conditions for that
time of year. In general terms, persistent worse-than-average conditions
would be one piece of information in the FEWS "convergence of evidence"
approach. However, it must be stressed that a satellite image alone does not
give a definitive answer on food-security conditions on the ground.