g2o::OptimizationAlgorithmWithHessian Class Reference

Base for solvers operating on the approximated Hessian, e.g., Gauss-Newton, Levenberg. More...

#include <optimization_algorithm_with_hessian.h>

Inheritance diagram for g2o::OptimizationAlgorithmWithHessian:

Collaboration diagram for g2o::OptimizationAlgorithmWithHessian:

Public Member Functions
	OptimizationAlgorithmWithHessian (Solver &solver)
virtual	~OptimizationAlgorithmWithHessian ()
virtual bool	init (bool online=false)
virtual bool	computeMarginals (SparseBlockMatrix< MatrixX > &spinv, const std::vector< std::pair< int, int > > &blockIndices)
virtual bool	buildLinearStructure ()
virtual void	updateLinearSystem ()
virtual bool	updateStructure (const std::vector< HyperGraph::Vertex * > &vset, const HyperGraph::EdgeSet &edges)
Solver &	solver ()
	return the underlying solver used to solve the linear system
virtual void	setWriteDebug (bool writeDebug)
virtual bool	writeDebug () const
Public Member Functions inherited from g2o::OptimizationAlgorithm
	OptimizationAlgorithm ()
virtual	~OptimizationAlgorithm ()
virtual SolverResult	solve (int iteration, bool online=false)=0
virtual void	printVerbose (std::ostream &os) const
const SparseOptimizer *	optimizer () const
	return the optimizer operating on
SparseOptimizer *	optimizer ()
void	setOptimizer (SparseOptimizer *optimizer)
const PropertyMap &	properties () const
	return the properties of the solver
bool	updatePropertiesFromString (const std::string &propString)
void	printProperties (std::ostream &os) const

Protected Attributes
Solver &	_solver
Property< bool > *	_writeDebug
Protected Attributes inherited from g2o::OptimizationAlgorithm
SparseOptimizer *	_optimizer
	the optimizer the solver is working on
PropertyMap	_properties

Detailed Description

Base for solvers operating on the approximated Hessian, e.g., Gauss-Newton, Levenberg.

Definition at line 41 of file optimization_algorithm_with_hessian.h.

Constructor & Destructor Documentation

OptimizationAlgorithmWithHessian()

g2o::OptimizationAlgorithmWithHessian::OptimizationAlgorithmWithHessian ( Solver &  solver )

explicit

Definition at line 39 of file optimization_algorithm_with_hessian.cpp.

    : OptimizationAlgorithm(), _solver(solver) {
  _writeDebug = _properties.makeProperty<Property<bool> >("writeDebug", true);
}

References g2o::OptimizationAlgorithm::_properties, _writeDebug, and g2o::PropertyMap::makeProperty().

~OptimizationAlgorithmWithHessian()

g2o::OptimizationAlgorithmWithHessian::~OptimizationAlgorithmWithHessian ( )

virtual

Definition at line 45 of file optimization_algorithm_with_hessian.cpp.

{}

Member Function Documentation

buildLinearStructure()

bool g2o::OptimizationAlgorithmWithHessian::buildLinearStructure ( )

virtual

Definition at line 76 of file optimization_algorithm_with_hessian.cpp.

                                                            {
  return _solver.buildStructure();
}

References _solver, and g2o::Solver::buildStructure().

Referenced by g2o::computeSimpleStars(), and g2o::Star::labelStarEdges().

computeMarginals()

bool g2o::OptimizationAlgorithmWithHessian::computeMarginals ( SparseBlockMatrix< MatrixX > &  spinv, const std::vector< std::pair< int, int > > &  blockIndices  )

virtual

computes the block diagonal elements of the pattern specified in the input and stores them in given SparseBlockMatrix. If your solver does not support computing the marginals, return false.

Implements g2o::OptimizationAlgorithm.

Definition at line 70 of file optimization_algorithm_with_hessian.cpp.

                                                       {
  return _solver.computeMarginals(spinv, blockIndices);
}

References _solver, and g2o::Solver::computeMarginals().

init()

bool g2o::OptimizationAlgorithmWithHessian::init ( bool  online = false )

virtual

initialize the solver, called once before the first call to solve()

Implements g2o::OptimizationAlgorithm.

Definition at line 47 of file optimization_algorithm_with_hessian.cpp.

                                                       {
  assert(_optimizer && "_optimizer not set");
  _solver.setWriteDebug(_writeDebug->value());
  bool useSchur = false;
  for (OptimizableGraph::VertexContainer::const_iterator it =
           _optimizer->activeVertices().begin();
       it != _optimizer->activeVertices().end(); ++it) {
    OptimizableGraph::Vertex* v = *it;
    if (v->marginalized()) {
      useSchur = true;
      break;
    }
  }
  if (useSchur) {
    if (_solver.supportsSchur()) _solver.setSchur(true);
  } else {
    if (_solver.supportsSchur()) _solver.setSchur(false);
  }
 
  bool initState = _solver.init(_optimizer, online);
  return initState;
}

References g2o::OptimizationAlgorithm::_optimizer, _solver, _writeDebug, g2o::SparseOptimizer::activeVertices(), g2o::Solver::init(), g2o::OptimizableGraph::Vertex::marginalized(), g2o::Solver::setSchur(), g2o::Solver::setWriteDebug(), g2o::Solver::supportsSchur(), and g2o::Property< T >::value().

setWriteDebug()

void g2o::OptimizationAlgorithmWithHessian::setWriteDebug ( bool  writeDebug )

virtual

write debug output of the Hessian if system is not positive definite

Definition at line 90 of file optimization_algorithm_with_hessian.cpp.

                                                                    {
  _writeDebug->setValue(writeDebug);
}

References _writeDebug, g2o::Property< T >::setValue(), and writeDebug().

solver()

Solver & g2o::OptimizationAlgorithmWithHessian::solver ( )

inline

return the underlying solver used to solve the linear system

Definition at line 61 of file optimization_algorithm_with_hessian.h.

{ return _solver; }

Referenced by g2o::SparseOptimizerIncremental::initSolver(), and g2o::SparseOptimizerOnline::initSolver().

updateLinearSystem()

void g2o::OptimizationAlgorithmWithHessian::updateLinearSystem ( )

virtual

Definition at line 80 of file optimization_algorithm_with_hessian.cpp.

                                                          {
  _solver.buildSystem();
}

References _solver, and g2o::Solver::buildSystem().

Referenced by g2o::computeSimpleStars(), and g2o::Star::labelStarEdges().

updateStructure()

bool g2o::OptimizationAlgorithmWithHessian::updateStructure ( const std::vector< HyperGraph::Vertex * > &  vset, const HyperGraph::EdgeSet &  edges  )

virtual

update the structures for online processing

Implements g2o::OptimizationAlgorithm.

Definition at line 84 of file optimization_algorithm_with_hessian.cpp.

                                    {
  return _solver.updateStructure(vset, edges);
}

References _solver, and g2o::Solver::updateStructure().

writeDebug()

virtual bool g2o::OptimizationAlgorithmWithHessian::writeDebug ( ) const

inlinevirtual

Definition at line 67 of file optimization_algorithm_with_hessian.h.

{ return _writeDebug->value(); }

Referenced by setWriteDebug().

Member Data Documentation

_solver

Solver& g2o::OptimizationAlgorithmWithHessian::_solver

protected

Definition at line 70 of file optimization_algorithm_with_hessian.h.

Referenced by buildLinearStructure(), computeMarginals(), g2o::OptimizationAlgorithmLevenberg::computeScale(), init(), g2o::OptimizationAlgorithmGaussNewton::printVerbose(), g2o::OptimizationAlgorithmLevenberg::printVerbose(), g2o::OptimizationAlgorithmDogleg::solve(), g2o::OptimizationAlgorithmGaussNewton::solve(), g2o::OptimizationAlgorithmLevenberg::solve(), updateLinearSystem(), and updateStructure().

_writeDebug

Property<bool>* g2o::OptimizationAlgorithmWithHessian::_writeDebug

protected

Definition at line 71 of file optimization_algorithm_with_hessian.h.

Referenced by init(), OptimizationAlgorithmWithHessian(), and setWriteDebug().

---

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

• /build/reproducible-path/g2o-0~20230806/g2o/core/optimization_algorithm_with_hessian.h
• /build/reproducible-path/g2o-0~20230806/g2o/core/optimization_algorithm_with_hessian.cpp