Point Cloud Library (PCL)  1.10.0
multiscale_feature_persistence.hpp
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2011, Alexandru-Eugen Ichim
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 #ifndef PCL_FEATURES_IMPL_MULTISCALE_FEATURE_PERSISTENCE_H_
41 #define PCL_FEATURES_IMPL_MULTISCALE_FEATURE_PERSISTENCE_H_
42 
43 #include <pcl/features/multiscale_feature_persistence.h>
44 
45 //////////////////////////////////////////////////////////////////////////////////////////////
46 template <typename PointSource, typename PointFeature>
48  alpha_ (0),
49  distance_metric_ (L1),
50  feature_estimator_ (),
51  features_at_scale_ (),
52  feature_representation_ ()
53 {
54  feature_representation_.reset (new DefaultPointRepresentation<PointFeature>);
55  // No input is needed, hack around the initCompute () check from PCLBase
56  input_.reset (new pcl::PointCloud<PointSource> ());
57 }
58 
59 
60 //////////////////////////////////////////////////////////////////////////////////////////////
61 template <typename PointSource, typename PointFeature> bool
63 {
65  {
66  PCL_ERROR ("[pcl::MultiscaleFeaturePersistence::initCompute] PCLBase::initCompute () failed - no input cloud was given.\n");
67  return false;
68  }
69  if (!feature_estimator_)
70  {
71  PCL_ERROR ("[pcl::MultiscaleFeaturePersistence::initCompute] No feature estimator was set\n");
72  return false;
73  }
74  if (scale_values_.empty ())
75  {
76  PCL_ERROR ("[pcl::MultiscaleFeaturePersistence::initCompute] No scale values were given\n");
77  return false;
78  }
79 
80  mean_feature_.resize (feature_representation_->getNumberOfDimensions ());
81 
82  return true;
83 }
84 
85 
86 //////////////////////////////////////////////////////////////////////////////////////////////
87 template <typename PointSource, typename PointFeature> void
89 {
90  features_at_scale_.clear ();
91  features_at_scale_.reserve (scale_values_.size ());
92  features_at_scale_vectorized_.clear ();
93  features_at_scale_vectorized_.reserve (scale_values_.size ());
94  for (std::size_t scale_i = 0; scale_i < scale_values_.size (); ++scale_i)
95  {
96  FeatureCloudPtr feature_cloud (new FeatureCloud ());
97  computeFeatureAtScale (scale_values_[scale_i], feature_cloud);
98  features_at_scale_[scale_i] = feature_cloud;
99 
100  // Vectorize each feature and insert it into the vectorized feature storage
101  std::vector<std::vector<float> > feature_cloud_vectorized;
102  feature_cloud_vectorized.reserve (feature_cloud->points.size ());
103 
104  for (const auto& feature: feature_cloud->points)
105  {
106  std::vector<float> feature_vectorized (feature_representation_->getNumberOfDimensions ());
107  feature_representation_->vectorize (feature, feature_vectorized);
108  feature_cloud_vectorized.emplace_back (std::move(feature_vectorized));
109  }
110  features_at_scale_vectorized_.emplace_back (std::move(feature_cloud_vectorized));
111  }
112 }
113 
114 
115 //////////////////////////////////////////////////////////////////////////////////////////////
116 template <typename PointSource, typename PointFeature> void
118  FeatureCloudPtr &features)
119 {
120  feature_estimator_->setRadiusSearch (scale);
121  feature_estimator_->compute (*features);
122 }
123 
124 
125 //////////////////////////////////////////////////////////////////////////////////////////////
126 template <typename PointSource, typename PointFeature> float
128  const std::vector<float> &b)
129 {
130  return (pcl::selectNorm<std::vector<float> > (a, b, a.size (), distance_metric_));
131 }
132 
133 
134 //////////////////////////////////////////////////////////////////////////////////////////////
135 template <typename PointSource, typename PointFeature> void
137 {
138  // Reset mean feature
139  for (int i = 0; i < feature_representation_->getNumberOfDimensions (); ++i)
140  mean_feature_[i] = 0.0f;
141 
142  float normalization_factor = 0.0f;
143  for (const auto& scale: features_at_scale_vectorized_)
144  {
145  normalization_factor += static_cast<float> (scale.size ());
146  for (const auto &feature : scale)
147  for (int dim_i = 0; dim_i < feature_representation_->getNumberOfDimensions (); ++dim_i)
148  mean_feature_[dim_i] += feature[dim_i];
149  }
150 
151  for (int dim_i = 0; dim_i < feature_representation_->getNumberOfDimensions (); ++dim_i)
152  mean_feature_[dim_i] /= normalization_factor;
153 }
154 
155 
156 //////////////////////////////////////////////////////////////////////////////////////////////
157 template <typename PointSource, typename PointFeature> void
159 {
160  unique_features_indices_.clear ();
161  unique_features_table_.clear ();
162  unique_features_indices_.reserve (scale_values_.size ());
163  unique_features_table_.reserve (scale_values_.size ());
164 
165  for (std::size_t scale_i = 0; scale_i < features_at_scale_vectorized_.size (); ++scale_i)
166  {
167  // Calculate standard deviation within the scale
168  float standard_dev = 0.0;
169  std::vector<float> diff_vector (features_at_scale_vectorized_[scale_i].size ());
170  diff_vector.clear();
171 
172  for (const auto& feature: features_at_scale_vectorized_[scale_i])
173  {
174  float diff = distanceBetweenFeatures (feature, mean_feature_);
175  standard_dev += diff * diff;
176  diff_vector.emplace_back (diff);
177  }
178  standard_dev = std::sqrt (standard_dev / static_cast<float> (features_at_scale_vectorized_[scale_i].size ()));
179  PCL_DEBUG ("[pcl::MultiscaleFeaturePersistence::extractUniqueFeatures] Standard deviation for scale %f is %f\n", scale_values_[scale_i], standard_dev);
180 
181  // Select only points outside (mean +/- alpha * standard_dev)
182  std::list<std::size_t> indices_per_scale;
183  std::vector<bool> indices_table_per_scale (features_at_scale_[scale_i]->points.size (), false);
184  for (std::size_t point_i = 0; point_i < features_at_scale_[scale_i]->points.size (); ++point_i)
185  {
186  if (diff_vector[point_i] > alpha_ * standard_dev)
187  {
188  indices_per_scale.emplace_back (point_i);
189  indices_table_per_scale[point_i] = true;
190  }
191  }
192  unique_features_indices_.emplace_back (std::move(indices_per_scale));
193  unique_features_table_.emplace_back (std::move(indices_table_per_scale));
194  }
195 }
196 
197 
198 //////////////////////////////////////////////////////////////////////////////////////////////
199 template <typename PointSource, typename PointFeature> void
201  shared_ptr<std::vector<int> > &output_indices)
202 {
203  if (!initCompute ())
204  return;
205 
206  // Compute the features for all scales with the given feature estimator
207  PCL_DEBUG ("[pcl::MultiscaleFeaturePersistence::determinePersistentFeatures] Computing features ...\n");
208  computeFeaturesAtAllScales ();
209 
210  // Compute mean feature
211  PCL_DEBUG ("[pcl::MultiscaleFeaturePersistence::determinePersistentFeatures] Calculating mean feature ...\n");
212  calculateMeanFeature ();
213 
214  // Get the 'unique' features at each scale
215  PCL_DEBUG ("[pcl::MultiscaleFeaturePersistence::determinePersistentFeatures] Extracting unique features ...\n");
216  extractUniqueFeatures ();
217 
218  PCL_DEBUG ("[pcl::MultiscaleFeaturePersistence::determinePersistentFeatures] Determining persistent features between scales ...\n");
219  // Determine persistent features between scales
220 
221 /*
222  // Method 1: a feature is considered persistent if it is 'unique' in at least 2 different scales
223  for (std::size_t scale_i = 0; scale_i < features_at_scale_vectorized_.size () - 1; ++scale_i)
224  for (std::list<std::size_t>::iterator feature_it = unique_features_indices_[scale_i].begin (); feature_it != unique_features_indices_[scale_i].end (); ++feature_it)
225  {
226  if (unique_features_table_[scale_i][*feature_it] == true)
227  {
228  output_features.points.push_back (features_at_scale[scale_i]->points[*feature_it]);
229  output_indices->push_back (feature_estimator_->getIndices ()->at (*feature_it));
230  }
231  }
232 */
233  // Method 2: a feature is considered persistent if it is 'unique' in all the scales
234  for (const auto& feature: unique_features_indices_.front ())
235  {
236  bool present_in_all = true;
237  for (std::size_t scale_i = 0; scale_i < features_at_scale_.size (); ++scale_i)
238  present_in_all = present_in_all && unique_features_table_[scale_i][feature];
239 
240  if (present_in_all)
241  {
242  output_features.points.emplace_back (features_at_scale_.front ()->points[feature]);
243  output_indices->emplace_back (feature_estimator_->getIndices ()->at (feature));
244  }
245  }
246 
247  // Consider that output cloud is unorganized
248  output_features.header = feature_estimator_->getInputCloud ()->header;
249  output_features.is_dense = feature_estimator_->getInputCloud ()->is_dense;
250  output_features.width = static_cast<std::uint32_t> (output_features.points.size ());
251  output_features.height = 1;
252 }
253 
254 
255 #define PCL_INSTANTIATE_MultiscaleFeaturePersistence(InT, Feature) template class PCL_EXPORTS pcl::MultiscaleFeaturePersistence<InT, Feature>;
256 
257 #endif /* PCL_FEATURES_IMPL_MULTISCALE_FEATURE_PERSISTENCE_H_ */
pcl::uint32_t
std::uint32_t uint32_t
Definition: pcl_macros.h:96
pcl::PointCloud::height
std::uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:402
pcl::PointCloud::points
std::vector< PointT, Eigen::aligned_allocator< PointT > > points
The point data.
Definition: point_cloud.h:397
pcl::PCLBase< PointSource >::input_
PointCloudConstPtr input_
The input point cloud dataset.
Definition: pcl_base.h:151
pcl::MultiscaleFeaturePersistence::computeFeaturesAtAllScales
void computeFeaturesAtAllScales()
Method that calls computeFeatureAtScale () for each scale parameter.
Definition: multiscale_feature_persistence.hpp:88
pcl::DefaultPointRepresentation
DefaultPointRepresentation extends PointRepresentation to define default behavior for common point ty...
Definition: point_representation.h:177
pcl::L1
@ L1
Definition: norms.h:54
pcl::PCLBase
PCL base class.
Definition: pcl_base.h:69
pcl::PointCloud< PointSource >
pcl::PointCloud::width
std::uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:400
pcl::selectNorm
float selectNorm(FloatVectorT a, FloatVectorT b, int dim, NormType norm_type)
Method that calculates any norm type available, based on the norm_type variable.
Definition: norms.hpp:49
pcl::PointCloud::is_dense
bool is_dense
True if no points are invalid (e.g., have NaN or Inf values in any of their floating point fields).
Definition: point_cloud.h:405
pcl::MultiscaleFeaturePersistence::FeatureCloudPtr
typename pcl::PointCloud< PointFeature >::Ptr FeatureCloudPtr
Definition: multiscale_feature_persistence.h:69
pcl::MultiscaleFeaturePersistence::determinePersistentFeatures
void determinePersistentFeatures(FeatureCloud &output_features, shared_ptr< std::vector< int > > &output_indices)
Central function that computes the persistent features.
Definition: multiscale_feature_persistence.hpp:200
pcl::PointCloud::header
pcl::PCLHeader header
The point cloud header.
Definition: point_cloud.h:394
pcl::MultiscaleFeaturePersistence
Generic class for extracting the persistent features from an input point cloud It can be given any Fe...
Definition: multiscale_feature_persistence.h:63
pcl::MultiscaleFeaturePersistence::MultiscaleFeaturePersistence
MultiscaleFeaturePersistence()
Empty constructor.
Definition: multiscale_feature_persistence.hpp:47
pcl::shared_ptr
boost::shared_ptr< T > shared_ptr
Alias for boost::shared_ptr.
Definition: pcl_macros.h:90