Point Cloud Library (PCL)  1.9.1-dev
projection_matrix.hpp
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2012-, Open Perception, Inc.
6  *
7  * All rights reserved.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  *
13  * * Redistributions of source code must retain the above copyright
14  * notice, this list of conditions and the following disclaimer.
15  * * Redistributions in binary form must reproduce the above
16  * copyright notice, this list of conditions and the following
17  * disclaimer in the documentation and/or other materials provided
18  * with the distribution.
19  * * Neither the name of the copyright holder(s) nor the names of its
20  * contributors may be used to endorse or promote products derived
21  * from this software without specific prior written permission.
22  *
23  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
24  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
25  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
26  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
27  * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
28  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
29  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
30  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
31  * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
33  * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
34  * POSSIBILITY OF SUCH DAMAGE.
35  *
36  */
37 
38 #ifndef __PCL_ORGANIZED_PROJECTION_MATRIX_HPP__
39 #define __PCL_ORGANIZED_PROJECTION_MATRIX_HPP__
40 
41 #include <pcl/cloud_iterator.h>
42 
43 ///////////////////////////////////////////////////////////////////////////////////////////
44 namespace pcl
45 {
46  namespace common
47  {
48  namespace internal
49  {
50  template <typename MatrixT> void
51  makeSymmetric (MatrixT& matrix, bool use_upper_triangular = true)
52  {
53  if (use_upper_triangular && (MatrixT::Flags & Eigen::RowMajorBit))
54  {
55  matrix.coeffRef (4) = matrix.coeff (1);
56  matrix.coeffRef (8) = matrix.coeff (2);
57  matrix.coeffRef (9) = matrix.coeff (6);
58  matrix.coeffRef (12) = matrix.coeff (3);
59  matrix.coeffRef (13) = matrix.coeff (7);
60  matrix.coeffRef (14) = matrix.coeff (11);
61  }
62  else
63  {
64  matrix.coeffRef (1) = matrix.coeff (4);
65  matrix.coeffRef (2) = matrix.coeff (8);
66  matrix.coeffRef (6) = matrix.coeff (9);
67  matrix.coeffRef (3) = matrix.coeff (12);
68  matrix.coeffRef (7) = matrix.coeff (13);
69  matrix.coeffRef (11) = matrix.coeff (14);
70  }
71  }
72  }
73  }
74 }
75 
76 //////////////////////////////////////////////////////////////////////////////
77 template <typename PointT> double
79  typename pcl::PointCloud<PointT>::ConstPtr cloud,
80  Eigen::Matrix<float, 3, 4, Eigen::RowMajor>& projection_matrix,
81  const std::vector<int>& indices)
82 {
83  // internally we calculate with double but store the result into float matrices.
84  using Scalar = double;
85  projection_matrix.setZero ();
86  if (cloud->height == 1 || cloud->width == 1)
87  {
88  PCL_ERROR ("[pcl::estimateProjectionMatrix] Input dataset is not organized!\n");
89  return (-1.0);
90  }
91 
92  Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor> A = Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor>::Zero ();
93  Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor> B = Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor>::Zero ();
94  Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor> C = Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor>::Zero ();
95  Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor> D = Eigen::Matrix<Scalar, 4, 4, Eigen::RowMajor>::Zero ();
96 
97  pcl::ConstCloudIterator <PointT> pointIt (*cloud, indices);
98 
99  while (pointIt)
100  {
101  unsigned yIdx = pointIt.getCurrentPointIndex () / cloud->width;
102  unsigned xIdx = pointIt.getCurrentPointIndex () % cloud->width;
103 
104  const PointT& point = *pointIt;
105  if (std::isfinite (point.x))
106  {
107  Scalar xx = point.x * point.x;
108  Scalar xy = point.x * point.y;
109  Scalar xz = point.x * point.z;
110  Scalar yy = point.y * point.y;
111  Scalar yz = point.y * point.z;
112  Scalar zz = point.z * point.z;
113  Scalar xx_yy = xIdx * xIdx + yIdx * yIdx;
114 
115  A.coeffRef (0) += xx;
116  A.coeffRef (1) += xy;
117  A.coeffRef (2) += xz;
118  A.coeffRef (3) += point.x;
119 
120  A.coeffRef (5) += yy;
121  A.coeffRef (6) += yz;
122  A.coeffRef (7) += point.y;
123 
124  A.coeffRef (10) += zz;
125  A.coeffRef (11) += point.z;
126  A.coeffRef (15) += 1.0;
127 
128  B.coeffRef (0) -= xx * xIdx;
129  B.coeffRef (1) -= xy * xIdx;
130  B.coeffRef (2) -= xz * xIdx;
131  B.coeffRef (3) -= point.x * static_cast<double>(xIdx);
132 
133  B.coeffRef (5) -= yy * xIdx;
134  B.coeffRef (6) -= yz * xIdx;
135  B.coeffRef (7) -= point.y * static_cast<double>(xIdx);
136 
137  B.coeffRef (10) -= zz * xIdx;
138  B.coeffRef (11) -= point.z * static_cast<double>(xIdx);
139 
140  B.coeffRef (15) -= xIdx;
141 
142  C.coeffRef (0) -= xx * yIdx;
143  C.coeffRef (1) -= xy * yIdx;
144  C.coeffRef (2) -= xz * yIdx;
145  C.coeffRef (3) -= point.x * static_cast<double>(yIdx);
146 
147  C.coeffRef (5) -= yy * yIdx;
148  C.coeffRef (6) -= yz * yIdx;
149  C.coeffRef (7) -= point.y * static_cast<double>(yIdx);
150 
151  C.coeffRef (10) -= zz * yIdx;
152  C.coeffRef (11) -= point.z * static_cast<double>(yIdx);
153 
154  C.coeffRef (15) -= yIdx;
155 
156  D.coeffRef (0) += xx * xx_yy;
157  D.coeffRef (1) += xy * xx_yy;
158  D.coeffRef (2) += xz * xx_yy;
159  D.coeffRef (3) += point.x * xx_yy;
160 
161  D.coeffRef (5) += yy * xx_yy;
162  D.coeffRef (6) += yz * xx_yy;
163  D.coeffRef (7) += point.y * xx_yy;
164 
165  D.coeffRef (10) += zz * xx_yy;
166  D.coeffRef (11) += point.z * xx_yy;
167 
168  D.coeffRef (15) += xx_yy;
169  }
170 
171  ++pointIt;
172  } // while
173 
178 
179  Eigen::Matrix<Scalar, 12, 12, Eigen::RowMajor> X = Eigen::Matrix<Scalar, 12, 12, Eigen::RowMajor>::Zero ();
180  X.topLeftCorner<4,4> ().matrix () = A;
181  X.block<4,4> (0, 8).matrix () = B;
182  X.block<4,4> (8, 0).matrix () = B;
183  X.block<4,4> (4, 4).matrix () = A;
184  X.block<4,4> (4, 8).matrix () = C;
185  X.block<4,4> (8, 4).matrix () = C;
186  X.block<4,4> (8, 8).matrix () = D;
187 
188  Eigen::SelfAdjointEigenSolver<Eigen::Matrix<Scalar, 12, 12, Eigen::RowMajor> > ei_symm (X);
189  Eigen::Matrix<Scalar, 12, 12, Eigen::RowMajor> eigen_vectors = ei_symm.eigenvectors ();
190 
191  // check whether the residual MSE is low. If its high, the cloud was not captured from a projective device.
192  Eigen::Matrix<Scalar, 1, 1> residual_sqr = eigen_vectors.col (0).transpose () * X * eigen_vectors.col (0);
193 
194  double residual = residual_sqr.coeff (0);
195 
196  projection_matrix.coeffRef (0) = static_cast <float> (eigen_vectors.coeff (0));
197  projection_matrix.coeffRef (1) = static_cast <float> (eigen_vectors.coeff (12));
198  projection_matrix.coeffRef (2) = static_cast <float> (eigen_vectors.coeff (24));
199  projection_matrix.coeffRef (3) = static_cast <float> (eigen_vectors.coeff (36));
200  projection_matrix.coeffRef (4) = static_cast <float> (eigen_vectors.coeff (48));
201  projection_matrix.coeffRef (5) = static_cast <float> (eigen_vectors.coeff (60));
202  projection_matrix.coeffRef (6) = static_cast <float> (eigen_vectors.coeff (72));
203  projection_matrix.coeffRef (7) = static_cast <float> (eigen_vectors.coeff (84));
204  projection_matrix.coeffRef (8) = static_cast <float> (eigen_vectors.coeff (96));
205  projection_matrix.coeffRef (9) = static_cast <float> (eigen_vectors.coeff (108));
206  projection_matrix.coeffRef (10) = static_cast <float> (eigen_vectors.coeff (120));
207  projection_matrix.coeffRef (11) = static_cast <float> (eigen_vectors.coeff (132));
208 
209  if (projection_matrix.coeff (0) < 0)
210  projection_matrix *= -1.0;
211 
212  return (residual);
213 }
214 
215 #endif
Iterator class for point clouds with or without given indices.
This file defines compatibility wrappers for low level I/O functions.
Definition: convolution.h:45
uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:428
uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:426
boost::shared_ptr< const PointCloud< PointT > > ConstPtr
Definition: point_cloud.h:442
void makeSymmetric(MatrixT &matrix, bool use_upper_triangular=true)
double estimateProjectionMatrix(typename pcl::PointCloud< PointT >::ConstPtr cloud, Eigen::Matrix< float, 3, 4, Eigen::RowMajor > &projection_matrix, const std::vector< int > &indices=std::vector< int >())
Estimates the projection matrix P = K * (R|-R*t) from organized point clouds, with K = [[fx...
A point structure representing Euclidean xyz coordinates, and the RGB color.
Definition: norms.h:54
unsigned getCurrentPointIndex() const