hsdar (version 0.5.1)

smgm: SMGM

Description

Calculate Gaussian model on soil spectra

Usage

smgm(x, percentage = TRUE, gridsize = 50)

Arguments

x

Object of class Speclib.

percentage

Flag if spectra in x are in range [0, 100]. If FALSE, the spectra are scaled to [0,100].

gridsize

Size of the grid used to perform least squares approximation.

Value

Object of class Speclib containing the fitted Gaussian spectra and the parameters derived from the Gaussian curve. The three parameters (absorption depth, R0; distance to the inflection point, sigma; area between the curve and 100 % reflectance, area) are stored in the attributes of the new Speclib. Additionally, the function returns the final root mean square error of the Gaussian fit.

Details

The algorithm fits a Gaussian function to the continuum points of the spectra in the spectral region between approx. 1500 to 2500 nm. The continuum points are derived constructing the convex hull of the spectra (see transformSpeclib). The Gaussian function requires three parameter: (1) the mean values which is set to the water fundamental of 2800 nm, (2) the absorption depth at 2800 nm, and (3) the distance to the inflection point of the function. The latter two parameters are iteratively chosen using a grid search. The mesh size of the grid can be adjusted with the gridsize parameter. Note that the function requires the spectral reflectance values to be in interval [0, 100].

References

Whiting, M. L., Li, L. and Ustin, S. L. (2004): Predicting water content using Gaussian model on soil spectra. Remote Sensing of Environment, 89, 535-552.

See Also

soilindex, Speclib

Examples

## Use PROSAIL to simulate spectra with different soil moisture content
Spektr.lib <- smoothSpeclib(PROSAIL(parameterList = data.frame(psoil = seq(0,1,0.1), LAI = 0)))

smgm_val <- smgm(Spektr.lib)

for (i in 1:nspectra(smgm_val))
  plot(smgm_val, FUN = i, new = i==1, col = i)

attribute(smgm_val)