Point Cloud Library (PCL)  1.8.0
registration.hpp
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 ///////////////////////////////////////////////////////////////////////////////////////////
42 template <typename PointSource, typename PointTarget, typename Scalar> void
45 {
46  setInputSource (cloud);
47 }
48 
49 ///////////////////////////////////////////////////////////////////////////////////////////
50 template <typename PointSource, typename PointTarget, typename Scalar> typename pcl::Registration<PointSource, PointTarget, Scalar>::PointCloudSourceConstPtr const
52 {
53  return (getInputSource ());
54 }
55 
56 ///////////////////////////////////////////////////////////////////////////////////////////
57 template <typename PointSource, typename PointTarget, typename Scalar> inline void
59 {
60  if (cloud->points.empty ())
61  {
62  PCL_ERROR ("[pcl::%s::setInputTarget] Invalid or empty point cloud dataset given!\n", getClassName ().c_str ());
63  return;
64  }
65  target_ = cloud;
66  target_cloud_updated_ = true;
67 }
68 
69 ///////////////////////////////////////////////////////////////////////////////////////////
70 template <typename PointSource, typename PointTarget, typename Scalar> bool
72 {
73  if (!target_)
74  {
75  PCL_ERROR ("[pcl::registration::%s::compute] No input target dataset was given!\n", getClassName ().c_str ());
76  return (false);
77  }
78 
79  // Only update target kd-tree if a new target cloud was set
80  if (target_cloud_updated_ && !force_no_recompute_)
81  {
82  tree_->setInputCloud (target_);
83  target_cloud_updated_ = false;
84  }
85 
86 
87  // Update the correspondence estimation
88  if (correspondence_estimation_)
89  {
90  correspondence_estimation_->setSearchMethodTarget (tree_, force_no_recompute_);
91  correspondence_estimation_->setSearchMethodSource (tree_reciprocal_, force_no_recompute_reciprocal_);
92  }
93 
94  // Note: we /cannot/ update the search method on all correspondence rejectors, because we know
95  // nothing about them. If they should be cached, they must be cached individually.
96 
98 }
99 
100 ///////////////////////////////////////////////////////////////////////////////////////////
101 template <typename PointSource, typename PointTarget, typename Scalar> bool
103 {
104  if (!input_)
105  {
106  PCL_ERROR ("[pcl::registration::%s::compute] No input source dataset was given!\n", getClassName ().c_str ());
107  return (false);
108  }
109 
110  if (source_cloud_updated_ && !force_no_recompute_reciprocal_)
111  {
112  tree_reciprocal_->setInputCloud (input_);
113  source_cloud_updated_ = false;
114  }
115  return (true);
116 }
117 
118 //////////////////////////////////////////////////////////////////////////////////////////////
119 template <typename PointSource, typename PointTarget, typename Scalar> inline double
121  const std::vector<float> &distances_a,
122  const std::vector<float> &distances_b)
123 {
124  unsigned int nr_elem = static_cast<unsigned int> (std::min (distances_a.size (), distances_b.size ()));
125  Eigen::VectorXf map_a = Eigen::VectorXf::Map (&distances_a[0], nr_elem);
126  Eigen::VectorXf map_b = Eigen::VectorXf::Map (&distances_b[0], nr_elem);
127  return (static_cast<double> ((map_a - map_b).sum ()) / static_cast<double> (nr_elem));
128 }
129 
130 //////////////////////////////////////////////////////////////////////////////////////////////
131 template <typename PointSource, typename PointTarget, typename Scalar> inline double
133 {
134 
135  double fitness_score = 0.0;
136 
137  // Transform the input dataset using the final transformation
138  PointCloudSource input_transformed;
139  transformPointCloud (*input_, input_transformed, final_transformation_);
140 
141  std::vector<int> nn_indices (1);
142  std::vector<float> nn_dists (1);
143 
144  // For each point in the source dataset
145  int nr = 0;
146  for (size_t i = 0; i < input_transformed.points.size (); ++i)
147  {
148  // Find its nearest neighbor in the target
149  tree_->nearestKSearch (input_transformed.points[i], 1, nn_indices, nn_dists);
150 
151  // Deal with occlusions (incomplete targets)
152  if (nn_dists[0] <= max_range)
153  {
154  // Add to the fitness score
155  fitness_score += nn_dists[0];
156  nr++;
157  }
158  }
159 
160  if (nr > 0)
161  return (fitness_score / nr);
162  else
163  return (std::numeric_limits<double>::max ());
164 
165 }
166 
167 //////////////////////////////////////////////////////////////////////////////////////////////
168 template <typename PointSource, typename PointTarget, typename Scalar> inline void
170 {
171  align (output, Matrix4::Identity ());
172 }
173 
174 //////////////////////////////////////////////////////////////////////////////////////////////
175 template <typename PointSource, typename PointTarget, typename Scalar> inline void
177 {
178  if (!initCompute ())
179  return;
180 
181  // Resize the output dataset
182  if (output.points.size () != indices_->size ())
183  output.points.resize (indices_->size ());
184  // Copy the header
185  output.header = input_->header;
186  // Check if the output will be computed for all points or only a subset
187  if (indices_->size () != input_->points.size ())
188  {
189  output.width = static_cast<uint32_t> (indices_->size ());
190  output.height = 1;
191  }
192  else
193  {
194  output.width = static_cast<uint32_t> (input_->width);
195  output.height = input_->height;
196  }
197  output.is_dense = input_->is_dense;
198 
199  // Copy the point data to output
200  for (size_t i = 0; i < indices_->size (); ++i)
201  output.points[i] = input_->points[(*indices_)[i]];
202 
203  // Set the internal point representation of choice unless otherwise noted
204  if (point_representation_ && !force_no_recompute_)
205  tree_->setPointRepresentation (point_representation_);
206 
207  // Perform the actual transformation computation
208  converged_ = false;
209  final_transformation_ = transformation_ = previous_transformation_ = Matrix4::Identity ();
210 
211  // Right before we estimate the transformation, we set all the point.data[3] values to 1 to aid the rigid
212  // transformation
213  for (size_t i = 0; i < indices_->size (); ++i)
214  output.points[i].data[3] = 1.0;
215 
216  computeTransformation (output, guess);
217 
218  deinitCompute ();
219 }
220 
bool initComputeReciprocal()
Internal computation when reciprocal lookup is needed.
void align(PointCloudSource &output)
Call the registration algorithm which estimates the transformation and returns the transformed source...
std::vector< PointT, Eigen::aligned_allocator< PointT > > points
The point data.
Definition: point_cloud.h:410
void setInputCloud(const PointCloudSourceConstPtr &cloud)
Provide a pointer to the input source (e.g., the point cloud that we want to align to the target) ...
PointCloudSourceConstPtr const getInputCloud()
Get a pointer to the input point cloud dataset target.
PointCloudSource::ConstPtr PointCloudSourceConstPtr
Definition: registration.h:84
double getFitnessScore(double max_range=std::numeric_limits< double >::max())
Obtain the Euclidean fitness score (e.g., sum of squared distances from the source to the target) ...
virtual void setInputTarget(const PointCloudTargetConstPtr &cloud)
Provide a pointer to the input target (e.g., the point cloud that we want to align the input source t...
uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:415
Eigen::Matrix< float, 4, 4 > Matrix4
Definition: registration.h:65
uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:413
void transformPointCloud(const pcl::PointCloud< PointT > &cloud_in, pcl::PointCloud< PointT > &cloud_out, const Eigen::Transform< Scalar, 3, Eigen::Affine > &transform, bool copy_all_fields=true)
Apply an affine transform defined by an Eigen Transform.
Definition: transforms.hpp:42
bool initCompute()
Internal computation initalization.
PCL base class.
Definition: pcl_base.h:68
pcl::PCLHeader header
The point cloud header.
Definition: point_cloud.h:407
PointCloudTarget::ConstPtr PointCloudTargetConstPtr
Definition: registration.h:88
bool is_dense
True if no points are invalid (e.g., have NaN or Inf values).
Definition: point_cloud.h:418