extrap {nnDiag} | R Documentation |
Comparison of the range of observations in the feature space of both the reference set and target set. Any target set pixel whose value is beyond the range of the reference set's feature space requires an extrapolation during kNN classification. This function prepares an object to plot the comparison and find if and where an extrapolation is necessary.
extrap(image, refSet.spectral, mask = NULL, verbose = TRUE)
image |
the target set image, either an object of class
GDALReadOnlyDataset
or an object of class "nnDaoi" |
refSet.spectral |
matrix of the feature space values of the reference set |
mask |
optional mask, as an object of class
GDALReadOnlyDataset |
verbose |
logical indicating whether to display a progress bar |
Both the image
and mask
have to be in the class
"GDALReadOnlyDataset"
. This can be achieved by using the
command GDAL.open
from the package rgdal.
Returns an object of class
"nnDext"
, which is a
list containing the following components:
image.range |
a matrix of the range of spectral values found on the image or in the AOIs for each spectral layer |
refSet.range |
a matrix of the range of spectral values for the reference set of data elements |
Brian Walters walte137@msu.edu
McRoberts, R.E. (2009) Diagnostic tools for nearest neighbors techniques when used with satellite imagery, Remote Sensing of Environment. 113, 489–499.
data(LuceVolume_subset)
LuceTasscap <- GDAL.open(system.file("data/LuceSubset_Tcap.tif", package = "nnDiag"))
LuceMask <- GDAL.open(system.file("data/LuceSubset_Mask.tif", package = "nnDiag"))
spect <- as.matrix(LuceVolume_subset[12:14])
## Not run:
x <- extrap(LuceTasscap, spect, LuceMask)
plot(x)
## Using an "nnDaoi"
object
coords <- as.matrix(LuceVolume_subset[,10:11])
tcap.aoi <- aoiMaker(coords, 3, LuceTasscap, LuceMask, aoi.size = 20, min.points = 4)
x <- extrap(tcap.aoi, spect)
plot(x)## End(Not run)