MVE - Multi-View Environment mve-devel
Loading...
Searching...
No Matches
Classes | Public Types | Public Member Functions | List of all members
sfm::ba::LinearSolver Class Reference

#include <ba_linear_solver.h>

Classes

struct  Options
 
struct  Status
 

Public Types

typedef DenseVector< double > DenseVectorType
 
typedef SparseMatrix< double > SparseMatrixType
 

Public Member Functions

 LinearSolver (Options const &options)
 
Status solve (SparseMatrixType const &jac_cams, SparseMatrixType const &jac_points, DenseVectorType const &vector_f, DenseVectorType *delta_x)
 Solve the system J^T J x = -J^T f based on the bundle adjustment mode.
 

Detailed Description

Definition at line 22 of file ba_linear_solver.h.

Member Typedef Documentation

◆ DenseVectorType

Definition at line 44 of file ba_linear_solver.h.

◆ SparseMatrixType

Definition at line 43 of file ba_linear_solver.h.

Constructor & Destructor Documentation

◆ LinearSolver()

sfm::ba::LinearSolver::LinearSolver ( Options const &  options)
inline

Definition at line 105 of file ba_linear_solver.h.

Member Function Documentation

◆ solve()

LinearSolver::Status sfm::ba::LinearSolver::solve ( SparseMatrixType const &  jac_cams,
SparseMatrixType const &  jac_points,
DenseVectorType const &  vector_f,
DenseVectorType delta_x 
)

Solve the system J^T J x = -J^T f based on the bundle adjustment mode.

If the Jacobian for cameras is empty, only points are optimized. If the Jacobian for points is empty, only cameras are optimized. If both, Jacobian for cams and points is given, the Schur complement trick is used to solve the linear system.

Definition at line 121 of file ba_linear_solver.cc.


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