All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Properties Friends Macros Groups Pages
Classes | Public Member Functions | Protected Attributes | List of all members
te::rp::ClassifierEMStrategy Class Reference
Classification

EM strategy for pixel-based classification. This is an unsupervised and pixel-based classification algorithm. Expectation-Maximization (EM) works iteratively by applying two steps: the E-step (Expectation) and the M-step (Maximization). The method aims to approximate the parameter estimates to real data distribution, along the iterations: More...

#include <ClassifierEMStrategy.h>

Inheritance diagram for te::rp::ClassifierEMStrategy:
te::rp::ClassifierStrategy

Classes

class  Parameters
 Classifier Parameters. More...
 

Public Member Functions

 ClassifierEMStrategy ()
 
bool execute (const te::rst::Raster &inputRaster, const std::vector< unsigned int > &inputRasterBands, const std::vector< te::gm::Polygon * > &inputPolygons, te::rst::Raster &outputRaster, const unsigned int outputRasterBand, const bool enableProgressInterface) throw (te::rp::Exception)
 Executes the classification strategy. More...
 
bool initialize (StrategyParameters const *const strategyParams) throw (te::rp::Exception)
 Initialize the classification strategy. More...
 
 ~ClassifierEMStrategy ()
 

Protected Attributes

bool m_isInitialized
 True if this instance is initialized. More...
 
ClassifierEMStrategy::Parameters m_parameters
 Internal execution parameters. More...
 

Detailed Description

EM strategy for pixel-based classification. This is an unsupervised and pixel-based classification algorithm. Expectation-Maximization (EM) works iteratively by applying two steps: the E-step (Expectation) and the M-step (Maximization). The method aims to approximate the parameter estimates to real data distribution, along the iterations:

  1. The E-step calculates the conditional expectation of the complete a posteriori probability function.
  2. The M-step updates the parameter estimation.

Definition at line 58 of file ClassifierEMStrategy.h.

Constructor & Destructor Documentation

te::rp::ClassifierEMStrategy::ClassifierEMStrategy ( )

Definition at line 88 of file ClassifierEMStrategy.cpp.

References m_isInitialized.

te::rp::ClassifierEMStrategy::~ClassifierEMStrategy ( )

Definition at line 93 of file ClassifierEMStrategy.cpp.

Member Function Documentation

bool te::rp::ClassifierEMStrategy::execute ( const te::rst::Raster &  inputRaster,
const std::vector< unsigned int > &  inputRasterBands,
const std::vector< te::gm::Polygon * > &  inputPolygons,
te::rst::Raster &  outputRaster,
const unsigned int  outputRasterBand,
const bool  enableProgressInterface 
)
throw (te::rp::Exception
)
virtual

Executes the classification strategy.

Parameters
inputRasterInput raster.
inputRasterBandsInput raster bands.
inputPolygonsThe polygons to be classified when using object-based image analysis (OBIA).
outputRasterOutput raster.
outputRasterBandOutput raster band.
enableProgressInterfaceEnable the internal strategy to update the progress interface.
Returns
true if OK, false on errors.

Implements te::rp::ClassifierStrategy.

Definition at line 120 of file ClassifierEMStrategy.cpp.

References te::rst::RasterIterator< T >::begin(), te::rst::PointSetIterator< T >::begin(), te::rst::RasterIterator< T >::end(), te::rst::PointSetIterator< T >::end(), te::rst::RasterIterator< T >::getColumn(), te::common::GetDeterminant(), te::common::GetInverseMatrix(), te::rst::GetRandomPointsInRaster(), te::rst::RasterIterator< T >::getRow(), te::rp::ClassifierEMStrategy::Parameters::m_clustersMeans, te::rp::ClassifierEMStrategy::Parameters::m_epsilon, m_isInitialized, te::rp::ClassifierEMStrategy::Parameters::m_maxInputPoints, te::rp::ClassifierEMStrategy::Parameters::m_maxIterations, te::rp::ClassifierEMStrategy::Parameters::m_numberOfClusters, m_parameters, te::common::TaskProgress::pulse(), te::common::TaskProgress::setCurrentStep(), te::common::TaskProgress::setMessage(), te::common::TaskProgress::setTotalSteps(), TE_TR, TERP_TRUE_OR_RETURN_FALSE, and te::common::TaskProgress::UNDEFINED.

bool te::rp::ClassifierEMStrategy::initialize ( StrategyParameters const *const  strategyParams)
throw (te::rp::Exception
)
virtual

Initialize the classification strategy.

Parameters
strategyParamsA pointer to the user given specific classification strategy parameters ou NULL if no parameters are present.
Returns
true if OK, false on errors.

Implements te::rp::ClassifierStrategy.

Definition at line 97 of file ClassifierEMStrategy.cpp.

References te::rp::ClassifierEMStrategy::Parameters::m_epsilon, m_isInitialized, te::rp::ClassifierEMStrategy::Parameters::m_maxInputPoints, te::rp::ClassifierEMStrategy::Parameters::m_maxIterations, te::rp::ClassifierEMStrategy::Parameters::m_numberOfClusters, m_parameters, TE_TR, and TERP_TRUE_OR_RETURN_FALSE.

Member Data Documentation

bool te::rp::ClassifierEMStrategy::m_isInitialized
protected

True if this instance is initialized.

Definition at line 107 of file ClassifierEMStrategy.h.

Referenced by ClassifierEMStrategy(), execute(), and initialize().

ClassifierEMStrategy::Parameters te::rp::ClassifierEMStrategy::m_parameters
protected

Internal execution parameters.

Definition at line 108 of file ClassifierEMStrategy.h.

Referenced by execute(), and initialize().

The documentation for this class was generated from the following files: