Point Cloud Library (PCL) 1.13.1
Loading...
Searching...
No Matches
transformation_estimation_lm.h
1/*
2 * Software License Agreement (BSD License)
3 *
4 * Point Cloud Library (PCL) - www.pointclouds.org
5 * Copyright (c) 2010-2011, Willow Garage, Inc.
6 * Copyright (c) 2012-, Open Perception, Inc.
7 *
8 * All rights reserved.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 *
14 * * Redistributions of source code must retain the above copyright
15 * notice, this list of conditions and the following disclaimer.
16 * * Redistributions in binary form must reproduce the above
17 * copyright notice, this list of conditions and the following
18 * disclaimer in the documentation and/or other materials provided
19 * with the distribution.
20 * * Neither the name of the copyright holder(s) nor the names of its
21 * contributors may be used to endorse or promote products derived
22 * from this software without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
25 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
26 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
27 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
28 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
29 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
30 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
31 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
32 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
34 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
35 * POSSIBILITY OF SUCH DAMAGE.
36 *
37 * $Id$
38 *
39 */
40
41#pragma once
42
43#include <pcl/registration/transformation_estimation.h>
44#include <pcl/registration/warp_point_rigid.h>
45#include <pcl/memory.h>
46
47namespace pcl {
48namespace registration {
49/** @b TransformationEstimationLM implements Levenberg Marquardt-based
50 * estimation of the transformation aligning the given correspondences.
51 *
52 * \note The class is templated on the source and target point types as well as on the
53 * output scalar of the transformation matrix (i.e., float or double). Default: float.
54 * \author Radu B. Rusu
55 * \ingroup registration
56 */
57template <typename PointSource, typename PointTarget, typename MatScalar = float>
59: public TransformationEstimation<PointSource, PointTarget, MatScalar> {
61 using PointCloudSourcePtr = typename PointCloudSource::Ptr;
62 using PointCloudSourceConstPtr = typename PointCloudSource::ConstPtr;
63
65
66 using PointIndicesPtr = PointIndices::Ptr;
67 using PointIndicesConstPtr = PointIndices::ConstPtr;
68
69public:
70 using Ptr =
71 shared_ptr<TransformationEstimationLM<PointSource, PointTarget, MatScalar>>;
72 using ConstPtr =
73 shared_ptr<const TransformationEstimationLM<PointSource, PointTarget, MatScalar>>;
74
75 using VectorX = Eigen::Matrix<MatScalar, Eigen::Dynamic, 1>;
76 using Vector4 = Eigen::Matrix<MatScalar, 4, 1>;
77 using Matrix4 =
79
80 /** \brief Constructor. */
82
83 /** \brief Copy constructor.
84 * \param[in] src the TransformationEstimationLM object to copy into this
85 */
92
93 /** \brief Copy operator.
94 * \param[in] src the TransformationEstimationLM object to copy into this
95 */
98 {
99 tmp_src_ = src.tmp_src_;
100 tmp_tgt_ = src.tmp_tgt_;
104 return (*this);
105 }
106
107 /** \brief Destructor. */
108 ~TransformationEstimationLM() override = default;
109
110 /** \brief Estimate a rigid rotation transformation between a source and a target
111 * point cloud using LM. \param[in] cloud_src the source point cloud dataset
112 * \param[in] cloud_tgt the target point cloud dataset
113 * \param[out] transformation_matrix the resultant transformation matrix
114 */
115 inline void
117 const pcl::PointCloud<PointTarget>& cloud_tgt,
118 Matrix4& transformation_matrix) const override;
119
120 /** \brief Estimate a rigid rotation transformation between a source and a target
121 * point cloud using LM. \param[in] cloud_src the source point cloud dataset
122 * \param[in] indices_src the vector of indices describing the points of interest in
123 * \a cloud_src
124 * \param[in] cloud_tgt the target point cloud dataset
125 * \param[out] transformation_matrix the resultant transformation matrix
126 */
127 inline void
129 const pcl::Indices& indices_src,
130 const pcl::PointCloud<PointTarget>& cloud_tgt,
131 Matrix4& transformation_matrix) const override;
132
133 /** \brief Estimate a rigid rotation transformation between a source and a target
134 * point cloud using LM. \param[in] cloud_src the source point cloud dataset
135 * \param[in] indices_src the vector of indices describing the points of interest in
136 * \a cloud_src
137 * \param[in] cloud_tgt the target point cloud dataset
138 * \param[in] indices_tgt the vector of indices describing the correspondences of the
139 * interest points from \a indices_src
140 * \param[out] transformation_matrix the resultant transformation matrix
141 */
142 inline void
144 const pcl::Indices& indices_src,
145 const pcl::PointCloud<PointTarget>& cloud_tgt,
146 const pcl::Indices& indices_tgt,
147 Matrix4& transformation_matrix) const override;
148
149 /** \brief Estimate a rigid rotation transformation between a source and a target
150 * point cloud using LM. \param[in] cloud_src the source point cloud dataset
151 * \param[in] cloud_tgt the target point cloud dataset
152 * \param[in] correspondences the vector of correspondences between source and target
153 * point cloud \param[out] transformation_matrix the resultant transformation matrix
154 */
155 inline void
157 const pcl::PointCloud<PointTarget>& cloud_tgt,
158 const pcl::Correspondences& correspondences,
159 Matrix4& transformation_matrix) const override;
160
161 /** \brief Set the function we use to warp points. Defaults to rigid 6D warp.
162 * \param[in] warp_fcn a shared pointer to an object that warps points
163 */
164 void
167 {
168 warp_point_ = warp_fcn;
169 }
170
171protected:
172 /** \brief Compute the distance between a source point and its corresponding target
173 * point \param[in] p_src The source point \param[in] p_tgt The target point \return
174 * The distance between \a p_src and \a p_tgt
175 *
176 * \note Older versions of PCL used this method internally for calculating the
177 * optimization gradient. Since PCL 1.7, a switch has been made to the
178 * computeDistance method using Vector4 types instead. This method is only
179 * kept for API compatibility reasons.
180 */
181 virtual MatScalar
182 computeDistance(const PointSource& p_src, const PointTarget& p_tgt) const
183 {
184 Vector4 s(p_src.x, p_src.y, p_src.z, 0);
185 Vector4 t(p_tgt.x, p_tgt.y, p_tgt.z, 0);
186 return ((s - t).norm());
187 }
188
189 /** \brief Compute the distance between a source point and its corresponding target
190 * point \param[in] p_src The source point \param[in] p_tgt The target point \return
191 * The distance between \a p_src and \a p_tgt
192 *
193 * \note A different distance function can be defined by creating a subclass of
194 * TransformationEstimationLM and overriding this method.
195 * (See \a TransformationEstimationPointToPlane)
196 */
197 virtual MatScalar
198 computeDistance(const Vector4& p_src, const PointTarget& p_tgt) const
199 {
200 Vector4 t(p_tgt.x, p_tgt.y, p_tgt.z, 0);
201 return ((p_src - t).norm());
202 }
203
204 /** \brief Temporary pointer to the source dataset. */
206
207 /** \brief Temporary pointer to the target dataset. */
209
210 /** \brief Temporary pointer to the source dataset indices. */
212
213 /** \brief Temporary pointer to the target dataset indices. */
215
216 /** \brief The parameterized function used to warp the source to the target. */
219
220 /** Base functor all the models that need non linear optimization must
221 * define their own one and implement operator() (const Eigen::VectorXd& x,
222 * Eigen::VectorXd& fvec) or operator() (const Eigen::VectorXf& x, Eigen::VectorXf&
223 * fvec) depending on the chosen _Scalar
224 */
225 template <typename _Scalar, int NX = Eigen::Dynamic, int NY = Eigen::Dynamic>
226 struct Functor {
227 using Scalar = _Scalar;
229 using InputType = Eigen::Matrix<_Scalar, InputsAtCompileTime, 1>;
230 using ValueType = Eigen::Matrix<_Scalar, ValuesAtCompileTime, 1>;
232 Eigen::Matrix<_Scalar, ValuesAtCompileTime, InputsAtCompileTime>;
233
234 /** \brief Empty Constructor. */
236
237 /** \brief Constructor
238 * \param[in] m_data_points number of data points to evaluate.
239 */
240 Functor(int m_data_points) : m_data_points_(m_data_points) {}
241
242 /** \brief Destructor. */
243 virtual ~Functor() = default;
244
245 /** \brief Get the number of values. */
246 int
247 values() const
248 {
249 return (m_data_points_);
250 }
251
252 protected:
254 };
255
256 struct OptimizationFunctor : public Functor<MatScalar> {
257 using Functor<MatScalar>::values;
258
259 /** Functor constructor
260 * \param[in] m_data_points the number of data points to evaluate
261 * \param[in,out] estimator pointer to the estimator object
262 */
263 OptimizationFunctor(int m_data_points, const TransformationEstimationLM* estimator)
264 : Functor<MatScalar>(m_data_points), estimator_(estimator)
265 {}
266
267 /** Copy constructor
268 * \param[in] src the optimization functor to copy into this
269 */
271 : Functor<MatScalar>(src.m_data_points_), estimator_()
272 {
273 *this = src;
274 }
275
276 /** Copy operator
277 * \param[in] src the optimization functor to copy into this
278 */
279 inline OptimizationFunctor&
281 {
284 return (*this);
285 }
286
287 /** \brief Destructor. */
288 ~OptimizationFunctor() override = default;
289
290 /** Fill fvec from x. For the current state vector x fill the f values
291 * \param[in] x state vector
292 * \param[out] fvec f values vector
293 */
294 int
295 operator()(const VectorX& x, VectorX& fvec) const;
296
298 };
299
300 struct OptimizationFunctorWithIndices : public Functor<MatScalar> {
301 using Functor<MatScalar>::values;
302
303 /** Functor constructor
304 * \param[in] m_data_points the number of data points to evaluate
305 * \param[in,out] estimator pointer to the estimator object
306 */
308 const TransformationEstimationLM* estimator)
309 : Functor<MatScalar>(m_data_points), estimator_(estimator)
310 {}
311
312 /** Copy constructor
313 * \param[in] src the optimization functor to copy into this
314 */
316 : Functor<MatScalar>(src.m_data_points_), estimator_()
317 {
318 *this = src;
319 }
320
321 /** Copy operator
322 * \param[in] src the optimization functor to copy into this
323 */
326 {
329 return (*this);
330 }
331
332 /** \brief Destructor. */
334
335 /** Fill fvec from x. For the current state vector x fill the f values
336 * \param[in] x state vector
337 * \param[out] fvec f values vector
338 */
339 int
340 operator()(const VectorX& x, VectorX& fvec) const;
341
343 };
344
345public:
347};
348} // namespace registration
349} // namespace pcl
350
351#include <pcl/registration/impl/transformation_estimation_lm.hpp>
shared_ptr< PointCloud< PointSource > > Ptr
shared_ptr< const PointCloud< PointSource > > ConstPtr
TransformationEstimation represents the base class for methods for transformation estimation based on...
TransformationEstimationLM implements Levenberg Marquardt-based estimation of the transformation alig...
TransformationEstimationLM & operator=(const TransformationEstimationLM &src)
Copy operator.
TransformationEstimationLM(const TransformationEstimationLM &src)
Copy constructor.
shared_ptr< const TransformationEstimationLM< PointSource, PointTarget, MatScalar > > ConstPtr
virtual MatScalar computeDistance(const PointSource &p_src, const PointTarget &p_tgt) const
Compute the distance between a source point and its corresponding target point.
shared_ptr< TransformationEstimationLM< PointSource, PointTarget, MatScalar > > Ptr
void estimateRigidTransformation(const pcl::PointCloud< PointSource > &cloud_src, const pcl::PointCloud< PointTarget > &cloud_tgt, Matrix4 &transformation_matrix) const override
Estimate a rigid rotation transformation between a source and a target point cloud using LM.
virtual MatScalar computeDistance(const Vector4 &p_src, const PointTarget &p_tgt) const
Compute the distance between a source point and its corresponding target point.
Eigen::Matrix< MatScalar, Eigen::Dynamic, 1 > VectorX
const pcl::Indices * tmp_idx_tgt_
Temporary pointer to the target dataset indices.
const pcl::Indices * tmp_idx_src_
Temporary pointer to the source dataset indices.
const PointCloudSource * tmp_src_
Temporary pointer to the source dataset.
~TransformationEstimationLM() override=default
Destructor.
typename TransformationEstimation< PointSource, PointTarget, MatScalar >::Matrix4 Matrix4
const PointCloudTarget * tmp_tgt_
Temporary pointer to the target dataset.
pcl::registration::WarpPointRigid< PointSource, PointTarget, MatScalar >::Ptr warp_point_
The parameterized function used to warp the source to the target.
void setWarpFunction(const typename WarpPointRigid< PointSource, PointTarget, MatScalar >::Ptr &warp_fcn)
Set the function we use to warp points.
shared_ptr< WarpPointRigid< PointSourceT, PointTargetT, Scalar > > Ptr
#define PCL_MAKE_ALIGNED_OPERATOR_NEW
Macro to signal a class requires a custom allocator.
Definition memory.h:63
Defines functions, macros and traits for allocating and using memory.
std::vector< pcl::Correspondence, Eigen::aligned_allocator< pcl::Correspondence > > Correspondences
IndicesAllocator<> Indices
Type used for indices in PCL.
Definition types.h:133
shared_ptr< ::pcl::PointIndices > Ptr
shared_ptr< const ::pcl::PointIndices > ConstPtr
Base functor all the models that need non linear optimization must define their own one and implement...
Eigen::Matrix< _Scalar, InputsAtCompileTime, 1 > InputType
Eigen::Matrix< _Scalar, ValuesAtCompileTime, 1 > ValueType
Eigen::Matrix< _Scalar, ValuesAtCompileTime, InputsAtCompileTime > JacobianType
OptimizationFunctor(int m_data_points, const TransformationEstimationLM *estimator)
Functor constructor.
int operator()(const VectorX &x, VectorX &fvec) const
Fill fvec from x.
OptimizationFunctor & operator=(const OptimizationFunctor &src)
Copy operator.
const TransformationEstimationLM< PointSource, PointTarget, MatScalar > * estimator_
OptimizationFunctorWithIndices(const OptimizationFunctorWithIndices &src)
Copy constructor.
OptimizationFunctorWithIndices(int m_data_points, const TransformationEstimationLM *estimator)
Functor constructor.
const TransformationEstimationLM< PointSource, PointTarget, MatScalar > * estimator_
OptimizationFunctorWithIndices & operator=(const OptimizationFunctorWithIndices &src)
Copy operator.