source: trunk/src/ATrous/baselineSubtract.cc @ 1441

// -----------------------------------------------------------------------
// baselineSubtract.cc: Obtaining and subtracting spectral baselines.
// -----------------------------------------------------------------------
// Copyright (C) 2006, Matthew Whiting, ATNF
//
// This program is free software; you can redistribute it and/or modify it
// under the terms of the GNU General Public License as published by the
// Free Software Foundation; either version 2 of the License, or (at your
// option) any later version.
//
// Duchamp is distributed in the hope that it will be useful, but WITHOUT
// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
// for more details.
//
// You should have received a copy of the GNU General Public License
// along with Duchamp; if not, write to the Free Software Foundation,
// Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA
//
// Correspondence concerning Duchamp may be directed to:
//    Internet email: Matthew.Whiting [at] atnf.csiro.au
//    Postal address: Dr. Matthew Whiting
//                    Australia Telescope National Facility, CSIRO
//                    PO Box 76
//                    Epping NSW 1710
//                    AUSTRALIA
// -----------------------------------------------------------------------
#include <iostream>
#include <iomanip>
#include <math.h>
#include <duchamp/param.hh>
#include <duchamp/ATrous/filter.hh>
#include <duchamp/ATrous/atrous.hh>
#include <duchamp/Utils/utils.hh>
#include <duchamp/Utils/feedback.hh>

namespace duchamp
{

  void baselineSubtract(size_t numSpec, size_t specLength, float *originalCube, 
                        float *baselineValues, Param &par)
  {
    ///  A routine to find the baseline of each spectrum in a cube (the spectral 
    ///    direction is assumed to be the third dimension) and subtract it off 
    ///    the original.
    /// \param numSpec Number of spatial pixels in original cube.
    /// \param specLength Size of spectral dimension of original cube.
    /// \param originalCube The cube that is to have its baseline removed. 
    /// \param baselineValues The cube of baseline values -- the same
    ///                       size as the original.
    /// \param par The Param set: information on BLANK values and on how
    ///            the subtraction is done.

    float *spec     = new float[specLength];
    float *thisBaseline = new float[specLength];
    int minscale = par.getMinScale();
    par.setMinScale(par.filter().getNumScales(specLength));
    float atrouscut = par.getAtrousCut();
    par.setAtrousCut(1);
    bool flagVerb = par.isVerbose();
    par.setVerbosity(false);

    ProgressBar bar;
    if(flagVerb) bar.init(numSpec);
    for(size_t pix=0; pix<numSpec; pix++){ // for each spatial pixel...

      if(flagVerb) bar.update(pix+1);

      for(size_t z=0; z<specLength; z++) 
        spec[z] = originalCube[z*numSpec + pix];

      atrous1DReconstruct(specLength,spec,thisBaseline,par);

      for(size_t z=0; z<specLength; z++) {
        baselineValues[z*numSpec+pix] = thisBaseline[z];
        if(!par.isBlank(originalCube[z*numSpec+pix])) 
          originalCube[z*numSpec+pix] = originalCube[z*numSpec+pix] - 
            baselineValues[z*numSpec+pix];
      }

    }    
    if(flagVerb) bar.rewind();

    par.setMinScale(minscale);
    par.setAtrousCut(atrouscut);
    par.setVerbosity(flagVerb);
  
    delete [] spec;
    delete [] thisBaseline;
    
  }

  void findAtrousBaseline(size_t size, float *input, float *baseline, Param &par)
  {
    ///   Uses the a trous reconstruction, keeping only the highest two scales, 
    ///     to reconstruct the baseline.
    ///    To avoid contamination by very strong signals, the input spectrum is 
    ///     trimmed at 5*MADFM above the median before reconstruction. 
    ///     This reduces the strong dips created by the presence of very strong 
    ///     signals.
    ///  \param size Length of the spectrum.
    ///  \param input The input array : this is not affected.
    ///  \param baseline The returned baseline array. This needs to be allocated
    ///   before the function is called.
    ///  \param par The Param set, needed for the atrous reconstruction.

    int minscale = par.getMinScale();
    par.setMinScale(par.filter().getNumScales(size));
    float atrouscut = par.getAtrousCut();
    par.setAtrousCut(1);
    bool flagVerb = par.isVerbose();
    par.setVerbosity(false);

    float *spec = new float[size];
    float med,sig;
    findMedianStats(input,size,med,sig);
    float threshold = 5. * sig;
    for(size_t i=0;i<size;i++) {
      if(fabs(input[i]-med)>threshold){
        if(input[i]>med) spec[i] = med + threshold;
        else spec[i] = med - threshold;
      }
      else spec[i] = input[i];
    }

    //    atrous1DReconstruct(size, input, baseline, par);
    atrous1DReconstruct(size, spec, baseline, par);

    par.setMinScale(minscale);
    par.setAtrousCut(atrouscut);
    par.setVerbosity(flagVerb);

    delete [] spec;

  }

 
  void findAtrousBaseline(size_t size, float *input, float *baseline)
  {
    ///  This version is designed for programs not using Param classes -- it 
    ///   keeps that side of things hidden from the user.
    ///  Uses the a trous reconstruction, keeping only the highest two scales, 
    ///   to reconstruct the baseline.
    ///  To avoid contamination by very strong signals, the input spectrum is 
    ///   trimmed at 8*MADFM above the median before reconstruction. This 
    ///   reduces the strong dips created by the presence of very strong 
    ///   signals.
    ///  \param size Length of the spectrum.
    ///  \param input The input array : this is not affected.
    ///  \param baseline The returned baseline array. This needs to be allocated
    ///   before the function is called.

    Param par;
    par.setMinScale(par.filter().getNumScales(size));
    par.setAtrousCut(1);
    par.setVerbosity(false);

    float *spec = new float[size];
    float med,sig;
    findMedianStats(input,size,med,sig);
    float threshold = 8. * sig;
    for(size_t i=0;i<size;i++) {
      if(fabs(input[i]-med)>threshold){
        if(input[i]>med) spec[i] = med + threshold;
        else spec[i] = med - threshold;
      }
      else spec[i] = input[i];
    }

    atrous1DReconstruct(size, spec, baseline, par);

    delete [] spec;

  }

}