findpeaks: Peak Detection

In kuwisdelu/matter: Out-of-core statistical computing and signal processing

Peak Detection

Description

Find peaks in a signal based on its local maxima, as determined by a sliding window.

Usage

# Find peaks
findpeaks(x, width = 5L, prominence = NULL,
    snr = NULL, noise = "quant", bounds = TRUE,
    relheight = 0.005, ...)

# Local maxima
locmax(x, width = 5L)

# Local minima
locmin(x, width = 5L)

Arguments

x	A numeric vector.
width	The number of signal elements to consider when determining if the center of the sliding window is a local extremum.
prominence	The minimum peak prominence used for filtering the peaks. The prominence of a peak is the height that the peak rises above the higher of its bases (i.e., its lowest contour line). A peak's bases are found as the local minima between the peak and the next higher peaks on either side.
snr	The minimum signal-to-noise ratio used for filtering the peaks.
noise	The method used to estimate the noise. See Details.
bounds	Whether the boundaries of each peak should be calculated and returned. A peak's boundaries are found as the nearest local minima on either side.
relheight	The minimum relative height (proportion of the maximum peak value) used for filtering the peaks.
...	Arguments passed to the noise estimation function.

Details

For locmax() and locmin(), a local extremum is defined as an element greater (or less) than all of the elements within width / 2 elements to the left of it, and greater (or less) than or equal to all of the elements within width / 2 elements to the right of it. That is, ties are resolved such that only the leftmost sample is considered the local extremum.

For findpeaks(), the peaks are simply the local maxima of the signal. The peak boundaries are found by descending a local maximum until a local minimum is found on either side, using the same criteria as above. The peaks are optionally filtered based on their prominences.

Optionally, the signal-to-noise ratio (SNR) can be estimated and used for filtering the peaks. These use the functions estnoise_quant, estnoise_diff, estnoise_filt, etc., to estimate the noise in the signal.

Value

For locmax() and locmin(), an logical vector indicating whether each element is a local extremum.

For findpeaks(), an integer vector giving the indices of the peaks, with attributes 'left_bounds' and 'right_bounds' giving the left and right boundaries of the peak as determined using the rule above.

Author(s)

Kylie A. Bemis

See Also

findpeaks_cwt, findpeaks_knn, estnoise_quant, estnoise_sd, estnoise_mad, estnoise_diff, estnoise_filt, peakwidths, peakareas, peakheights, binpeaks, mergepeaks

Examples

# simple signal
x <- c(0, 1, 1, 2, 3, 2, 1, 4, 5, 1, 1, 0)
locmax(x)
findpeaks(x)

# simulated spectrum
set.seed(1)
x <- simspec(size=5000)

# find peaks with snr >= 3
p <- findpeaks(x, snr=3, noise="quant")
plot(x, type="l")
points(p, x[p], col="red")

# find peaks with derivative-based noise
p <- findpeaks(x, snr=3, noise="diff")
plot(x, type="l")
points(p, x[p], col="red")

kuwisdelu/matter documentation built on Dec. 8, 2024, 8:09 p.m.