This HTML automatically generated with rman for NEMO
Table of Contents

Name

tabpeak - find peaks or valleys in a table

Synopsis

tabpeak [parameter=value]

Description

Find local maxima (and/or minima) and fit a local paraboloid, return the location and value of the peak, or valley. See also tablsqfit(1NEMO) and ccdmom(1NEMO) .

The default is to search for local 3 point extrema. Using edge= one can force it away from the edge, and using pmin= one can increase

the minimum number of points that need to be in the extreme

Parameters

The following parameters are recognized in order; they may be given in any order if the keyword is also given. Use --help to confirm this man page is up to date.
in=
Input file name [???]
xcol=
X-Column [1]
ycol=
Y-column [2]
clip=
Only consider points above this [0]
delta=
If a positive value is used, single peaks larger than this value on either side of this peak are reported. [0]
pmin=
[EXPERIMENTAL] Require the peak to contain at least pmin points. By default, only 3 are required, in which case a polynomial is fit (though this is an exact fit). By setting this higher, more points are required to be part of a peak which is then required to be monotonously decreasing. edge= Number of points near the edge to skip for a peak finder. For the simplest 3-point peak finder this should be at least 1. [1]
valley=t|f
Also fit valleys, in which case the signal has to be below -clip. [Default: f]
mean=t|f
Use intensity weighted mean ? [Default: f]
npeak=
extract the Nth peak (N>0). If npeak=0, no extraction done. Other keywords will be ignored if a peak is ectracted for further analysis. [Default: 0]
epeak=
expand around the N-th peak extracted by npeak>0 by this factor. By default the segment around a peak is defined until the first point away from the peak will increase again. [Default: 1]
nmax=
max size if a pipe [100000]

Examples

The peak in a simple paraboloid: 
  $ nemoinp 1:3 | tabmath - - ’-(2.2-%1)**2+10’ | tabpeak -
  2.200000 10.000000
Here is a pipeline extracting the first peak, fitting a gaussian to it, and plotting the data and residuals (in the PGPLOT driver the data will be in red, the residuals in green, and the fit in blue): 
  $ tabpeak spec.tab | tabnllsqfit - fit=gauss1d out=- | tee fit.log | tabcomment
- | tabplot - 1 2,3,4 color=2,3,4 line=1,1
  $ cat fit.log

Finding single peaks by setting an appropriate delta (related to a robust dispersion of the differences of the data): 


  $ tabtrend L1157-B1_109989__1.bstats.tab  2 | tabhist - robust=t
  ...
  Robust Mean Disp     : 1.52724e-05 0.00640417
  
  $ tabpeak L1157-B1_109989__1.bstats.tab  delta=0.05
  #  X       Y       Ypeak    Delta    Delta
  166.000000 1.591776  0.083979 0.087718 0.080239
  286.000000 26.544859  25.062155 25.063228 25.061082
  452.000000 1.838022  0.324911 0.324182 0.325639
  ...
  3524.000000 1.520718  0.229133 0.229248 0.229019
  3810.000000 0.427400  -0.871551 0.870384 0.872717

See Also

tablsqfit(1NEMO) , tabnllsqfit(1NEMO) , tabtrend(1NEMO) , tabhist(1NEMO) , tabplot(1NEMO)

Author

Peter Teuben

Update History


28-May-2013    V0.1 Created    PJT
30-may-2013    V0.2 added valley=    PJT
23-mar-2022    V0.4 added pmin=    PJT
25-mar-2022    V0.5 added edge=    PJT
23-jan-2023    V0.7 added npeak, epeak=    PJT
25-may-2023    0.8 added delta=    PJT

Table of Contents