Example usage:

./MarkovClassification1Example Input/QB_Suburb.png Output/MarkovRandomField1.png 1.0 20 1.0 1

Example source code (MarkovClassification1Example.cxx):

// This example illustrates the details of the \doxygen{otb}{MarkovRandomFieldFilter}.
// This filter is an application of the Markov Random Fields for classification,
// segmentation or restoration.
// This example applies the \doxygen{otb}{MarkovRandomFieldFilter} to
// classify an image into four classes defined by their mean and variance. The
// optimization is done using an Metropolis algorithm with a random sampler. The
// regularization energy is defined by a Potts model and the fidelity by a
// Gaussian model.

#include "otbImageFileReader.h"
#include "otbImageFileWriter.h"
#include "otbImage.h"
#include "otbMarkovRandomFieldFilter.h"
#include "itkUnaryFunctorImageFilter.h"
#include "itkRescaleIntensityImageFilter.h"

// The first step toward the use of this filter is the inclusion of the proper
// header files.

#include "otbMRFEnergyPotts.h"
#include "otbMRFEnergyGaussianClassification.h"
#include "otbMRFOptimizerMetropolis.h"
#include "otbMRFSamplerRandom.h"

int main(int argc, char* argv[])

  if (argc != 7)
    std::cerr << "Missing Parameters " << argc << std::endl;
    std::cerr << "Usage: " << argv[0];
    std::cerr << " inputImage output lambda iterations optimizerTemperature" << std::endl;
    std::cerr << " useRandomValue" << std::endl;
    return 1;

  //  Then we must decide what pixel type to use for the image. We
  //  choose to make all computations with double precision.
  //  The labelled image is of type unsigned char which allows up to 256 different
  //  classes.

  const unsigned int Dimension = 2;

  using InternalPixelType = double;
  using LabelledPixelType = unsigned char;
  using InputImageType    = otb::Image<InternalPixelType, Dimension>;
  using LabelledImageType = otb::Image<LabelledPixelType, Dimension>;

  //  We define a reader for the image to be classified, an initialization for the
  //  classification (which could be random) and a writer for the final
  //  classification.

  using ReaderType = otb::ImageFileReader<InputImageType>;
  using WriterType = otb::ImageFileWriter<LabelledImageType>;

  ReaderType::Pointer reader = ReaderType::New();
  WriterType::Pointer writer = WriterType::New();

  const char* inputFilename  = argv[1];
  const char* outputFilename = argv[2];


  //  Finally, we define the different classes necessary for the Markov classification.
  //  A \doxygen{otb}{MarkovRandomFieldFilter} is instantiated, this is the
  // main class which connect the other to do the Markov classification.

  using MarkovRandomFieldFilterType = otb::MarkovRandomFieldFilter<InputImageType, LabelledImageType>;

  //  An \doxygen{otb}{MRFSamplerRandomMAP}, which derives from the
  // \doxygen{otb}{MRFSampler}, is instantiated. The sampler is in charge of
  // proposing a modification for a given site. The
  // \doxygen{otb}{MRFSamplerRandomMAP}, randomly pick one possible value
  // according to the MAP probability.

  using SamplerType = otb::MRFSamplerRandom<InputImageType, LabelledImageType>;

  //  An \doxygen{otb}{MRFOptimizerMetropoli}, which derives from the
  // \doxygen{otb}{MRFOptimizer}, is instantiated. The optimizer is in charge
  // of accepting or rejecting the value proposed by the sampler. The
  // \doxygen{otb}{MRFSamplerRandomMAP}, accept the proposal according to the
  // variation of energy it causes and a temperature parameter.

  using OptimizerType = otb::MRFOptimizerMetropolis;

  // Two energy, deriving from the \doxygen{otb}{MRFEnergy} class need to be instantiated. One energy
  // is required for the regularization, taking into account the relationship between neighboring pixels
  // in the classified image. Here it is done with the \doxygen{otb}{MRFEnergyPotts} which implement
  // a Potts model.
  // The second energy is for the fidelity to the original data. Here it is done with an
  // \doxygen{otb}{MRFEnergyGaussianClassification} class, which defines a gaussian model for the data.

  using EnergyRegularizationType = otb::MRFEnergyPotts<LabelledImageType, LabelledImageType>;
  using EnergyFidelityType       = otb::MRFEnergyGaussianClassification<InputImageType, LabelledImageType>;

  // The different filters composing our pipeline are created by invoking their
  // \code{New()} methods, assigning the results to smart pointers.

  MarkovRandomFieldFilterType::Pointer markovFilter         = MarkovRandomFieldFilterType::New();
  EnergyRegularizationType::Pointer    energyRegularization = EnergyRegularizationType::New();
  EnergyFidelityType::Pointer          energyFidelity       = EnergyFidelityType::New();
  OptimizerType::Pointer               optimizer            = OptimizerType::New();
  SamplerType::Pointer                 sampler              = SamplerType::New();

  // Parameter for the \doxygen{otb}{MRFEnergyGaussianClassification} class, meand
  // and standard deviation are created.

  if ((bool)(atoi(argv[6])) == true)
    // Overpass random calculation(for test only):

  unsigned int nClass = 4;
  energyFidelity->SetNumberOfParameters(2 * nClass);
  EnergyFidelityType::ParametersType parameters;
  parameters[0] = 10.0;  // Class 0 mean
  parameters[1] = 10.0;  // Class 0 stdev
  parameters[2] = 80.0;  // Class 1 mean
  parameters[3] = 10.0;  // Class 1 stdev
  parameters[4] = 150.0; // Class 2 mean
  parameters[5] = 10.0;  // Class 2 stdev
  parameters[6] = 220.0; // Class 3 mean
  parameters[7] = 10.0;  // Class 3 stde

  // Parameters are given to the different class an the sampler, optimizer and
  // energies are connected with the Markov filter.

  OptimizerType::ParametersType param(1);


  // The pipeline is connected. An \doxygen{itk}{RescaleIntensityImageFilter}
  // rescale the classified image before saving it.


  using RescaleType                  = itk::RescaleIntensityImageFilter<LabelledImageType, LabelledImageType>;
  RescaleType::Pointer rescaleFilter = RescaleType::New();



  // Finally, the pipeline execution is triggered.


  // Figure~\ref{fig:MRF_CLASSIFICATION1} shows the output of the Markov Random
  // Field classification after 20 iterations with a
  // random sampler and a Metropolis optimizer.
  // \begin{figure}
  // \center
  // \includegraphics[width=0.44\textwidth]{QB_Suburb.eps}
  // \includegraphics[width=0.44\textwidth]{MarkovRandomField1.eps}
  // \itkcaption[MRF restoration]{Result of applying
  // the \doxygen{otb}{MarkovRandomFieldFilter} to an extract from a PAN Quickbird
  // image for classification. The result is obtained after 20 iterations with a
  // random sampler and a Metropolis optimizer. From left to right : original image,
  // classification.}
  // \label{fig:MRF_CLASSIFICATION1}
  // \end{figure}

  return EXIT_SUCCESS;