SRBA: Sparser Relative Bundle Adjustment
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends
Public Member Functions | Public Attributes
srba::RbaEngine< KF2KF_POSE_TYPE, LM_TYPE, OBS_TYPE, RBA_OPTIONS >::TSRBAParameters Struct Reference

Detailed Description

template<class KF2KF_POSE_TYPE, class LM_TYPE, class OBS_TYPE, class RBA_OPTIONS = RBA_OPTIONS_DEFAULT>
struct srba::RbaEngine< KF2KF_POSE_TYPE, LM_TYPE, OBS_TYPE, RBA_OPTIONS >::TSRBAParameters

Different parameters for the SRBA methods

Definition at line 424 of file RbaEngine.h.

#include <RbaEngine.h>

Inheritance diagram for srba::RbaEngine< KF2KF_POSE_TYPE, LM_TYPE, OBS_TYPE, RBA_OPTIONS >::TSRBAParameters:
mrpt::utils::CLoadableOptions

List of all members.

Public Member Functions

virtual void loadFromConfigFile (const mrpt::utils::CConfigFileBase &source, const std::string &section)
virtual void saveToConfigFile (mrpt::utils::CConfigFileBase &out, const std::string &section) const

Public Attributes

topo_dist_t max_tree_depth
topo_dist_t max_optimize_depth
bool optimize_new_edges_alone
 (Default:true) Before running a whole "local area" optimization, try to optimize new edges one by one to have a better starting point.
bool use_robust_kernel
bool use_robust_kernel_stage1
double kernel_param
size_t max_iters
double max_error_per_obs_to_stop
 default: 1e-9
double max_rho
 default: 1.0
double max_lambda
 default: 1e20
double min_error_reduction_ratio_to_relinearize
 default 0.01
bool numeric_jacobians
 (Default:false) Use a numeric approximation of the Jacobians (very slow!) instead of analytical ones.
void(* feedback_user_iteration )(unsigned int iter, const double total_sq_err, const double mean_sqroot_error)
bool compute_condition_number
 Compute and return to the user the Hessian condition number of k2k edges (default=false)
bool compute_sparsity_stats
 Compute stats on the sparsity of the problem matrices (default=false)
double max_rmse_show_red_warning
 Minimum RSME to show optimization error in red color (default=0.5)
TCovarianceRecoveryPolicy cov_recovery
 Recover covariance? What method to use? (Default: crpLandmarksApprox)

Member Function Documentation

template<class KF2KF_POSE_TYPE , class LM_TYPE , class OBS_TYPE , class RBA_OPTIONS >
void srba::RbaEngine< KF2KF_POSE_TYPE, LM_TYPE, OBS_TYPE, RBA_OPTIONS >::TSRBAParameters::loadFromConfigFile ( const mrpt::utils::CConfigFileBase source,
const std::string &  section 
) [virtual]

See docs of mrpt::utils::CLoadableOptions

Implements mrpt::utils::CLoadableOptions.

Definition at line 60 of file rba_problem_common.h.

template<class KF2KF_POSE_TYPE , class LM_TYPE , class OBS_TYPE , class RBA_OPTIONS >
void srba::RbaEngine< KF2KF_POSE_TYPE, LM_TYPE, OBS_TYPE, RBA_OPTIONS >::TSRBAParameters::saveToConfigFile ( mrpt::utils::CConfigFileBase out,
const std::string &  section 
) const [virtual]

See docs of mrpt::utils::CLoadableOptions

Definition at line 79 of file rba_problem_common.h.


Member Data Documentation

template<class KF2KF_POSE_TYPE , class LM_TYPE , class OBS_TYPE , class RBA_OPTIONS = RBA_OPTIONS_DEFAULT>
topo_dist_t srba::RbaEngine< KF2KF_POSE_TYPE, LM_TYPE, OBS_TYPE, RBA_OPTIONS >::TSRBAParameters::max_optimize_depth

The maximum topological distance of keyframes to be optimized around the most recent keyframe.

Definition at line 437 of file RbaEngine.h.

template<class KF2KF_POSE_TYPE , class LM_TYPE , class OBS_TYPE , class RBA_OPTIONS = RBA_OPTIONS_DEFAULT>
topo_dist_t srba::RbaEngine< KF2KF_POSE_TYPE, LM_TYPE, OBS_TYPE, RBA_OPTIONS >::TSRBAParameters::max_tree_depth

Maximum depth for maintained spanning trees.

Definition at line 434 of file RbaEngine.h.


The documentation for this struct was generated from the following files:
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends