eigenFilter: Eigenimage filter

In emanuelhuber/RGPR: Ground-penetrating radar (GPR) data visualisation, processing and delineation

Eigenimage filter

Description

Decompose the GPR data (radargram) using singular value decomposition (SVD) and return the reconstructed data for the selected singular values (eigenvalues). This method is sometimes refered to in the litterature as the Karhunen-Loeve (KL) transformation or the eigenimage decomposition.

Usage

## S4 method for signature 'GPR'
eigenFilter(x, eigenvalue = NA, center = TRUE, scale = FALSE, track = TRUE)

Arguments

x	An object of the class GPR
eigenvalue	An integer vector specifying the eigenvalues selected. eigenvalue = 1:5 returns the image reconstructed using the first five eigenvalues while c(2,4,6) will return the image reconstructed for these three specific eigenvalues. If NA, a plot of the eigenvalues spectrum will be displayed and the user will be prompted to enter the pair of selected eigenvalues separated by a comma. That is for example 1,2 will return the image reconstructred using tthe first two eigenvalues. If NULL, the image is reconstructed using all eigenvalues. The number of eigenvalue is equal to min(dim(x)).
center	Logical or numeric. If TRUE, centering is done by subtracting the layer means (omitting NA's), and if FALSE, no centering is done. If center is a numeric vector with length equal to the nlayers(x), then each layer of x has the corresponding value from center subtracted from it.
scale	Logical or numeric. If TRUE, scaling is done by dividing the (centered) layers of x by their standard deviations if center is TRUE, and the root mean square otherwise. If scale is FALSE, no scaling is done. If scale is a numeric vector with length equal to nlayers(x), each layer of x is divided by the corresponding value. Scaling is done after centering.

Value

An object of the class GPR.

References

• Textbook: Sacchi (2002) Statistical and Transform Methods in Geophysical Signal Processing

Examples

 
data(frenkeLine00)
x <- frenkeLine00
x1 <- eigenFilter(x, eigenvalue = c(1,3))
plot(x)
plot(x1)

emanuelhuber/RGPR documentation built on May 13, 2024, 9:31 p.m.