Point Cloud Library (PCL)  1.12.0-dev
vector_average.h
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2010-2012, 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 
38 #pragma once
39 
40 #include <Eigen/Core> // for Matrix
41 
42 #include <pcl/memory.h>
43 #include <pcl/pcl_macros.h>
44 
45 namespace pcl
46 {
47  /** \brief Calculates the weighted average and the covariance matrix
48  *
49  * A class to calculate the weighted average and the covariance matrix of a set of vectors with given weights.
50  * The original data is not saved. Mean and covariance are calculated iteratively.
51  * \author Bastian Steder
52  * \ingroup common
53  */
54  template <typename real, int dimension>
56  {
57  public:
58  using VectorType = Eigen::Matrix<real, dimension, 1>;
59  using MatrixType = Eigen::Matrix<real, dimension, dimension>;
60  //-----CONSTRUCTOR&DESTRUCTOR-----
61  /** Constructor - dimension gives the size of the vectors to work with. */
62  VectorAverage ();
63 
64  //-----METHODS-----
65  /** Reset the object to work with a new data set */
66  inline void
67  reset ();
68 
69  /** Get the mean of the added vectors */
70  inline const
71  VectorType& getMean () const { return mean_;}
72 
73  /** Get the covariance matrix of the added vectors */
74  inline const
75  MatrixType& getCovariance () const { return covariance_;}
76 
77  /** Get the summed up weight of all added vectors */
78  inline real
80 
81  /** Get the number of added vectors */
82  inline unsigned int
84 
85  /** Add a new sample */
86  inline void
87  add (const VectorType& sample, real weight=1.0);
88 
89  /** Do Principal component analysis */
90  inline void
91  doPCA (VectorType& eigen_values, VectorType& eigen_vector1,
92  VectorType& eigen_vector2, VectorType& eigen_vector3) const;
93 
94  /** Do Principal component analysis */
95  inline void
96  doPCA (VectorType& eigen_values) const;
97 
98  /** Get the eigenvector corresponding to the smallest eigenvalue */
99  inline void
100  getEigenVector1 (VectorType& eigen_vector1) const;
101 
103 
104  //-----VARIABLES-----
105 
106 
107  protected:
108  //-----METHODS-----
109  //-----VARIABLES-----
110  unsigned int noOfSamples_ = 0;
112  VectorType mean_ = VectorType::Identity ();
113  MatrixType covariance_ = MatrixType::Identity ();
114  };
115 
119 } // END namespace
120 
121 #include <pcl/common/impl/vector_average.hpp>
pcl::VectorAverage::getAccumulatedWeight
real getAccumulatedWeight() const
Get the summed up weight of all added vectors.
Definition: vector_average.h:79
pcl_macros.h
Defines all the PCL and non-PCL macros used.
pcl
Definition: convolution.h:46
pcl::VectorAverage::getEigenVector1
void getEigenVector1(VectorType &eigen_vector1) const
Get the eigenvector corresponding to the smallest eigenvalue.
Definition: vector_average.hpp:121
pcl::VectorAverage::getCovariance
const MatrixType & getCovariance() const
Get the covariance matrix of the added vectors.
Definition: vector_average.h:75
pcl::VectorAverage::getMean
const VectorType & getMean() const
Get the mean of the added vectors.
Definition: vector_average.h:71
pcl::VectorAverage::VectorAverage
VectorAverage()
Constructor - dimension gives the size of the vectors to work with.
Definition: vector_average.hpp:48
pcl::VectorAverage::mean_
VectorType mean_
Definition: vector_average.h:112
pcl::VectorAverage::VectorType
Eigen::Matrix< real, dimension, 1 > VectorType
Definition: vector_average.h:58
pcl::VectorAverage::accumulatedWeight_
real accumulatedWeight_
Definition: vector_average.h:111
pcl::VectorAverage::reset
void reset()
Reset the object to work with a new data set.
Definition: vector_average.hpp:54
pcl::VectorAverage::getNoOfSamples
unsigned int getNoOfSamples()
Get the number of added vectors.
Definition: vector_average.h:83
pcl::VectorAverage::doPCA
void doPCA(VectorType &eigen_values, VectorType &eigen_vector1, VectorType &eigen_vector2, VectorType &eigen_vector3) const
Do Principal component analysis.
Definition: vector_average.hpp:84
PCL_MAKE_ALIGNED_OPERATOR_NEW
#define PCL_MAKE_ALIGNED_OPERATOR_NEW
Macro to signal a class requires a custom allocator.
Definition: memory.h:63
pcl::VectorAverage
Calculates the weighted average and the covariance matrix.
Definition: vector_average.h:55
pcl::VectorAverage::covariance_
MatrixType covariance_
Definition: vector_average.h:113
pcl::VectorAverage::noOfSamples_
unsigned int noOfSamples_
Definition: vector_average.h:110
pcl::VectorAverage::add
void add(const VectorType &sample, real weight=1.0)
Add a new sample.
Definition: vector_average.hpp:63
pcl::VectorAverage::MatrixType
Eigen::Matrix< real, dimension, dimension > MatrixType
Definition: vector_average.h:59
memory.h
Defines functions, macros and traits for allocating and using memory.