LSSTApplications  17.0+11,17.0+34,17.0+56,17.0+57,17.0+59,17.0+7,17.0-1-g377950a+33,17.0.1-1-g114240f+2,17.0.1-1-g4d4fbc4+28,17.0.1-1-g55520dc+49,17.0.1-1-g5f4ed7e+52,17.0.1-1-g6dd7d69+17,17.0.1-1-g8de6c91+11,17.0.1-1-gb9095d2+7,17.0.1-1-ge9fec5e+5,17.0.1-1-gf4e0155+55,17.0.1-1-gfc65f5f+50,17.0.1-1-gfc6fb1f+20,17.0.1-10-g87f9f3f+1,17.0.1-11-ge9de802+16,17.0.1-16-ga14f7d5c+4,17.0.1-17-gc79d625+1,17.0.1-17-gdae4c4a+8,17.0.1-2-g26618f5+29,17.0.1-2-g54f2ebc+9,17.0.1-2-gf403422+1,17.0.1-20-g2ca2f74+6,17.0.1-23-gf3eadeb7+1,17.0.1-3-g7e86b59+39,17.0.1-3-gb5ca14a,17.0.1-3-gd08d533+40,17.0.1-30-g596af8797,17.0.1-4-g59d126d+4,17.0.1-4-gc69c472+5,17.0.1-6-g5afd9b9+4,17.0.1-7-g35889ee+1,17.0.1-7-gc7c8782+18,17.0.1-9-gc4bbfb2+3,w.2019.22
LSSTDataManagementBasePackage
imageStatistics.cc
// -*- LSST-C++ -*-
/*
* LSST Data Management System
* Copyright 2008, 2009, 2010 LSST Corporation.
*
* This product includes software developed by the
* LSST Project (http://www.lsst.org/).
*
* 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 3 of the License, or
* (at your option) any later version.
*
* This program 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 LSST License Statement and
* the GNU General Public License along with this program. If not,
* see <http://www.lsstcorp.org/LegalNotices/>.
*/
#include <cmath>
#include <iostream>
#include <limits>
#include <memory>
#include "lsst/geom.h"
namespace image = lsst::afw::image;
namespace math = lsst::afw::math;
typedef image::Image<float> ImageF;
typedef image::MaskedImage<float> MaskedImageF;
typedef math::Statistics ImgStat;
typedef math::MaskedVector<float> MaskedVectorF;
/*
* An example of how to use the Statistics class
*/
template <typename Image>
void printStats(Image &img, math::StatisticsControl const &sctrl) {
// initialize a Statistics object with any stats we might want
ImgStat stats = math::makeStatistics(
img,
sctrl);
// get various stats with getValue() and their errors with getError()
double const npoint = stats.getValue(math::NPOINT);
double const mean = stats.getValue(math::MEAN);
double const var = stats.getValue(math::VARIANCE);
double const dmean = stats.getError(math::MEAN);
double const sd = stats.getValue(math::STDEV);
double const min = stats.getValue(math::MIN);
double const max = stats.getValue(math::MAX);
double const meanclip = stats.getValue(math::MEANCLIP);
double const varclip = stats.getValue(math::VARIANCECLIP);
double const stdevclip = stats.getValue(math::STDEVCLIP);
double const median = stats.getValue(math::MEDIAN);
double const iqrange = stats.getValue(math::IQRANGE);
// output
std::cout << "N " << npoint << std::endl;
std::cout << "dmean " << dmean << std::endl;
std::cout << "mean: " << mean << std::endl;
std::cout << "meanclip: " << meanclip << std::endl;
std::cout << "var: " << var << std::endl;
std::cout << "varclip: " << varclip << std::endl;
std::cout << "stdev: " << sd << std::endl;
std::cout << "stdevclip: " << stdevclip << std::endl;
std::cout << "min: " << min << std::endl;
std::cout << "max: " << max << std::endl;
std::cout << "median: " << median << std::endl;
std::cout << "iqrange: " << iqrange << std::endl;
}
int main() {
// declare an image and a masked image
int const wid = 1024;
ImageF img(lsst::geom::Extent2I(wid, wid));
MaskedImageF mimg(img.getDimensions());
MaskedVectorF mv(wid * wid);
// fill it with some noise (Cauchy noise in this case)
for (int j = 0; j != img.getHeight(); ++j) {
int k = 0;
MaskedImageF::x_iterator mip = mimg.row_begin(j);
for (ImageF::x_iterator ip = img.row_begin(j); ip != img.row_end(j); ++ip) {
double const xUniform = M_PI * static_cast<ImageF::Pixel>(std::rand()) / RAND_MAX;
double xLorentz = xUniform; // tan(xUniform - M_PI/2.0);
// throw in the occassional nan ... 1% of the time
if (static_cast<double>(std::rand()) / RAND_MAX < 0.01) {
xLorentz = NAN;
}
*ip = xLorentz;
// mask the odd rows
// variance actually diverges for Cauchy noise ... but stats doesn't access this.
*mip = MaskedImageF::Pixel(xLorentz, (k % 2) ? 0x1 : 0x0, (k % 2) ? 1.0e99 : 1.0);
v.push_back(xLorentz);
++k;
++mip;
}
}
int j = 0;
for (MaskedVectorF::iterator mvp = mv.begin(); mvp != mv.end(); ++mvp) {
*mvp = MaskedVectorF::Pixel(v[j], (j % 2) ? 0x1 : 0x0, 10.0);
++j;
}
std::shared_ptr<std::vector<float> > vF = mv.getVector();
// make a statistics control object and override some of the default properties
math::StatisticsControl sctrl;
sctrl.setNumIter(3);
sctrl.setNumSigmaClip(5.0);
sctrl.setAndMask(0x1); // pixels with this mask bit set will be ignored.
sctrl.setNanSafe(true);
// ==================================================================
// Get stats for the Image, MaskedImage, and vector
std::cout << "image::Image" << std::endl;
printStats(img, sctrl);
std::cout << "image::MaskedImage" << std::endl;
printStats(mimg, sctrl);
std::cout << "std::vector" << std::endl;
printStats(v, sctrl);
std::cout << "image::MaskedVector" << std::endl;
printStats(mv, sctrl);
std::cout << "image::MaskedVector::getVector()" << std::endl;
printStats(*vF, sctrl);
// Now try the weighted statistics
sctrl.setWeighted(true);
sctrl.setAndMask(0x0);
std::cout << "image::MaskedImage (weighted)" << std::endl;
printStats(mimg, sctrl);
// Now try the specialization to get NPOINT and SUM (bitwise OR) for an image::Mask
math::Statistics mskstat = makeStatistics(*mimg.getMask(), (math::NPOINT | math::SUM), sctrl);
std::cout << "image::Mask" << std::endl;
std::cout << mskstat.getValue(math::NPOINT) << " " << mskstat.getValue(math::SUM) << std::endl;
return 0;
}