denoise: Noise filtering through principal components

In environmentalinformatics-marburg/remote: Empirical Orthogonal Teleconnections in R

Noise filtering through principal components

Description

Filter noise from a RasterStack by decomposing into principal components and subsequent reconstruction using only a subset of components

Usage

denoise(
  x,
  k = NULL,
  expl.var = NULL,
  weighted = TRUE,
  use.cpp = TRUE,
  verbose = TRUE,
  ...
)

Arguments

x	RasterStack to be filtered
k	number of components to be kept for reconstruction (ignored if expl.var is supplied)
expl.var	minimum amount of variance to be kept after reconstruction (should be set to NULL or omitted if k is supplied)
weighted	logical. If TRUE the covariance matrix will be geographically weighted using the cosine of latitude during decomposition (only important for lat/lon data)
use.cpp	logical. Determines whether to use Rcpp functionality, defaults to TRUE.
verbose	logical. If TRUE some details about the calculation process will be output to the console
...	additional arguments passed to stats::princomp()

Value

a denoised RasterStack

See Also

anomalize(), deseason()

Examples

data("vdendool")
vdd_dns <- denoise(vdendool, expl.var = 0.8)

opar <- par(mfrow = c(1,2))
plot(vdendool[[1]], main = "original")
plot(vdd_dns[[1]], main = "denoised")
par(opar)

environmentalinformatics-marburg/remote documentation built on June 12, 2025, 1:10 p.m.