Remote Sensing: Fractional Cover
Fractional cover (FC) is the percentage of vegetative cover on
the ground. In the context of a remotely-sensed image, the FC values
represent the percentage of vegetative cover present in each pixel.
Fractional cover is derived from the Normalized Difference Vegetation
Index (NDVI). The equation is
where VIsoil is the NDVI
value of a pixel containing 0% vegetative cover, and VI
veg is the NDVI value of a pixel containing 100%
vegetative cover.
Once a value has been determined for each of the two variables,
the FC formula is applied to the NDVI image as an algorithm which
outputs a preliminary FC image (FCprelim
). This preliminary image is then modified so that it ranges from
0 to 1. Due to the well-documented nature of NDVI saturating at the
upper end in tropical environments, it is quite common to end up with
FCprelim values greater than 1. It
is also common to have FCprelim values
below 0, as the NDVI of water can be less than bare soil. Therefore,
we are working under the assumption that our VI
soil and VIveg
values are accurate. So rather than stretching the values to
fit between 0 and 1, we simply change all negative FCprelim
values to 0 and all FCprelim values
greater than 1 to 1.
Fractional cover has been generated for every TM image that has
been classified. Up to this point, it's primarily been used to add a
bit of additional detail to the generalized classifications in the
Path 8/Row 60 and Path 9/Row 61 classifications. Through the use of
FC, we have been able to expand the Agriculture/Pasture
class to Agriculture, Sparse Agriculture,
Pasture, and Agriculture/Pasture
(for those areas where we are unsure). We have also been able to split
the Forest class into as many as 20 different classes with differing
levels of canopy density. This provides us with more information than
previously available.
