genericnumer.h

Go to the documentation of this file.

#pragma once
 
/*@+
________________________________________________________________________
 
 (C) dGB Beheer B.V.; (LICENSE) http://opendtect.org/OpendTect_license.txt
 Author:        Kristofer Tingdahl
 Date:          07-10-1999
 RCS:           $Id$
________________________________________________________________________
 
 
*/
 
#include "algomod.h"
#include "mathfunc.h"
#include <math.h>
 
template <class A, class B, class C>
inline void GenericConvolve( int lx, int ifx, const A& x,
                             int ly, int ify, const B& y,
                             int lz, int ifz, C& z)
{
    int ilx=ifx+lx-1,ily=ify+ly-1,ilz=ifz+lz-1,i,j,jlow,jhigh;
    float sum;
 
    for ( i=ifz; i<=ilz; ++i )
    {
        jlow = i-ily;
        if ( jlow < ifx ) jlow = ifx;
 
        jhigh = i-ify;
        if ( jhigh > ilx ) jhigh = ilx;
 
        for ( j=jlow,sum=0.0; j<=jhigh; ++j )
        {
            if( !mIsUdf(x[j-ifx]) && !mIsUdf(y[i-j-ify]) )
                sum += x[j-ifx]*y[i-j-ify];
        }
 
        z[i-ifz] = sum;
    }
}
 
 
template <class A, class B, class C>
inline void GenericConvolveNoUdf( int lx, int ifx, const A& x,
                             int ly, int ify, const B& y,
                             int lz, int ifz, C& z)
{
    int ilx=ifx+lx-1,ily=ify+ly-1,ilz=ifz+lz-1,i,j,jlow,jhigh;
    float sum;
 
    for ( i=ifz; i<=ilz; ++i )
    {
        jlow = i-ily;
        if ( jlow < ifx ) jlow = ifx;
 
        jhigh = i-ify;
        if ( jhigh > ilx ) jhigh = ilx;
 
        for ( j=jlow,sum=0.0; j<=jhigh; ++j )
            sum += x[j-ifx]*y[i-j-ify];
 
        z[i-ifz] = sum;
    }
}
 
 
 
template <class A, class B>
inline float similarity( const A& a, const B& b, int sz, bool normalize=false,
                         int firstposa=0, int firstposb=0 )
{
    float val1, val2;
    double sqdist = 0, sq1 = 0, sq2 = 0;
 
    double meana = mUdf(double), stddeva = mUdf(double);
    double meanb = mUdf(double), stddevb = mUdf(double);
 
    if ( normalize )
    {
        if ( sz==1 ) normalize = false;
        else
        {
            double asum=0,bsum=0;
            for ( int idx=0; idx<sz; idx++ )
            {
                asum += a[firstposa+idx];
                bsum += b[firstposb+idx];
            }
 
            meana = asum/sz;
            meanb = bsum/sz;
 
            asum = 0;
            bsum = 0;
            for ( int idx=0; idx<sz; idx++ )
            {
                const double adiff = a[firstposa+idx]-meana;
                const double bdiff = b[firstposb+idx]-meanb;
                asum += adiff*adiff;
                bsum += bdiff*bdiff;
            }
 
            stddeva = Math::Sqrt(asum/(sz-1));
            stddevb = Math::Sqrt(bsum/(sz-1));
 
            if ( mIsZero(stddeva,mDefEps) || mIsZero(stddevb,mDefEps) )
                normalize=false;
        }
    }
 
    int curposa = firstposa;
    int curposb = firstposb;
 
    for ( int idx=0; idx<sz; idx++ )
    {
        val1 = normalize ? (float) ( (a[curposa]-meana)/stddeva ) : a[curposa];
        val2 = normalize ? (float) ( (b[curposb]-meanb)/stddevb ) : b[curposb];
        if ( mIsUdf(val1) || mIsUdf(val2) )
            return mUdf(float);
 
        sq1 += val1 * val1;
        sq2 += val2 * val2;
        sqdist += (val1-val2) * (val1-val2);
 
        curposa ++;
        curposb ++;
    }
 
    if ( mIsZero(sq1,mDefEps) && mIsZero(sq2,mDefEps) )
        return 1;
 
    if ( mIsZero(sq1,mDefEps) || mIsZero(sq2,mDefEps) )
        return 0;
 
    const float rt =
           (float) ( Math::Sqrt(sqdist) / (Math::Sqrt(sq1) + Math::Sqrt(sq2)) );
    return 1 - rt;
}
 
 
mGlobal(Algo) float similarity(const FloatMathFunction&,
                 const FloatMathFunction&,
                 float x1, float x2, float dist, int sz, bool normalize );
 
 
mGlobal(Algo) float semblance( const ObjectSet<float>& signals,
                               const Interval<int>& );
 
mGlobal(Algo) float semblance( const ObjectSet<float>& signals,int signalsize,
                         const TypeSet<float>& signalstarts,
                         const Interval<int>& gate );
 
mGlobal(Algo) double LanczosKernel( int size, double x );
 
mGlobal(Algo) bool findValue(const FloatMathFunction&,float x1,float x2,
                        float& res,float targetval = 0,float tol=1e-5);
 
 
mGlobal(Algo) float findValueInAperture(const FloatMathFunction&,float startx,
                const Interval<float>& aperture,float dx,float target=0,
                float tol=1e-5);
 
mGlobal(Algo) float findExtreme(const FloatMathFunction&,bool minima,float x1,
                                float x2,float tol = 1e-5);
 
 
template <class A, class B> inline
void reSample( const FloatMathFunction& input, const A& samplevals,
               B& output, int nrsamples )
{
    for ( int idx=0; idx<nrsamples; idx++ )
    {
        const float sampleval = samplevals[idx];
        output[idx] = Values::isUdf(sampleval) ? mUdf(float) :
                      input.getValue(sampleval);
    }
}
 
 
template <class A, class B, class C>
inline void genericCrossCorrelation( int lx, int ifx, const A& x,
                                     int ly, int ify, const B& y,
                                     int lz, int ifz, C& z)
{
    mAllocLargeVarLenArr( float, xreversed, lx );
 
    for ( int i=0,j=lx-1; i<lx; ++i,--j)
        xreversed[i] = x[j];
    GenericConvolve( lx, 1-ifx-lx, xreversed, ly, ify, y, lz, ifz, z );
}
 
 
template <class A>
inline void reverseArray( A* in, int sz, A* out=0 )
{
    if ( out )
        for ( int idx=0; idx<sz; idx++ )
            out[idx] = in[sz-1-idx];
    else
    {
        mAllocVarLenArr( A, tmparr, sz/2 );
        for ( int idx=0; idx<sz/2; idx++ )
        {
            tmparr[idx] = in[idx];
            int opsamp = sz-1-idx;
            in[idx] = in[opsamp];
            in[opsamp] = tmparr[idx];
        }
    }
}