1-D median filtering
Apply a running median of odd span to the input x
medfilt1(x, n = 3, MARGIN = 2, na.omit = FALSE, ...)
x |
Input signal, specified as a numeric vector, matrix or array. |
n |
positive integer width of the median window; must be odd. Default: 3 |
MARGIN |
Vector giving the subscripts which the function will be applied over. E.g., for a matrix 1 indicates rows, 2 indicates columns, c(1, 2) indicates rows and columns. Where X has named dimnames, it can be a character vector selecting dimension names. Default: 2 (columns). |
na.omit |
logical indicating whether to omit missing values,
or interpolate then using a cubic spline function
( |
... |
other arguments passed to |
This function computes a running median over the input x, using the
runmed function. Because of that, it works a little
differently than the 'Matlab' or 'Octave' versions (i.e., it does not produce
exactly the same values).
The 'Mablab' and 'Octave' functions have a
'nanflag' option that allows to include or remove missing values. If
inclusion is specifies, then the function returns a signal so that the median
of any segment containing NAs is also NA. Because the 'runmed' function
does not include an na.omit option, implementing this functionality
would lead to a considerable speed loss. Instead, a na.omit parameter
was implemented that allows either omitting NAs or interpolating them with a
spline function.
Instead of the 'zeropad' and
'truncate' options to the 'padding' argument in the 'Matlab'
and 'Octave' functions, the present version uses the standard
endrule parameter of the 'runmed' function, with options
keep, constant, or median.
Filtered signal, returned as a numeric vector, matrix, or array, of
the same size as x.
Geert van Boxtel, G.J.M.vanBoxtel@gmail.com.
## noise suppression
fs <- 100
t <- seq(0, 1, 1/fs)
x <- sin(2 * pi * t * 3) + 0.25 * sin(2 * pi * t * 40)
plot(t, x, type = "l", xlab = "", ylab = "")
y <- medfilt1(x, 11)
lines (t, y, col = "red")
legend("topright", c("Original", "Filtered"), lty = 1, col = 1:2)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.