Point Cloud Library (PCL) 1.13.1
Loading...
Searching...
No Matches
cvfh.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#ifndef PCL_FEATURES_IMPL_CVFH_H_
42#define PCL_FEATURES_IMPL_CVFH_H_
43
44#include <pcl/features/cvfh.h>
45#include <pcl/features/normal_3d.h>
46#include <pcl/common/centroid.h>
47
48//////////////////////////////////////////////////////////////////////////////////////////////
49template<typename PointInT, typename PointNT, typename PointOutT> void
51{
53 {
54 output.width = output.height = 0;
55 output.clear ();
56 return;
57 }
58 // Resize the output dataset
59 // Important! We should only allocate precisely how many elements we will need, otherwise
60 // we risk at pre-allocating too much memory which could lead to bad_alloc
61 // (see http://dev.pointclouds.org/issues/657)
62 output.width = output.height = 1;
63 output.resize (1);
64
65 // Perform the actual feature computation
66 computeFeature (output);
67
69}
70
71//////////////////////////////////////////////////////////////////////////////////////////////
72template<typename PointInT, typename PointNT, typename PointOutT> void
76 float tolerance,
78 std::vector<pcl::PointIndices> &clusters,
79 double eps_angle,
80 unsigned int min_pts_per_cluster,
81 unsigned int max_pts_per_cluster)
82{
83 if (tree->getInputCloud ()->size () != cloud.size ())
84 {
85 PCL_ERROR("[pcl::extractEuclideanClusters] Tree built for a different point cloud "
86 "dataset (%zu) than the input cloud (%zu)!\n",
87 static_cast<std::size_t>(tree->getInputCloud()->size()),
88 static_cast<std::size_t>(cloud.size()));
89 return;
90 }
91 if (cloud.size () != normals.size ())
92 {
93 PCL_ERROR("[pcl::extractEuclideanClusters] Number of points in the input point "
94 "cloud (%zu) different than normals (%zu)!\n",
95 static_cast<std::size_t>(cloud.size()),
96 static_cast<std::size_t>(normals.size()));
97 return;
98 }
99
100 // Create a bool vector of processed point indices, and initialize it to false
101 std::vector<bool> processed (cloud.size (), false);
102
103 pcl::Indices nn_indices;
104 std::vector<float> nn_distances;
105 // Process all points in the indices vector
106 for (std::size_t i = 0; i < cloud.size (); ++i)
107 {
108 if (processed[i])
109 continue;
110 processed[i] = true;
111
113 r.header = cloud.header;
114 auto& seed_queue = r.indices;
115
116 seed_queue.push_back (i);
117
118 // loop has an emplace_back, making it difficult to use modern loops
119 for (std::size_t idx = 0; idx != seed_queue.size (); ++idx)
120 {
121 // Search for seed_queue[index]
122 if (!tree->radiusSearch (seed_queue[idx], tolerance, nn_indices, nn_distances))
123 {
124 continue;
125 }
126
127 // skip index 0, since nn_indices[0] == idx, worth it?
128 for (std::size_t j = 1; j < nn_indices.size (); ++j)
129 {
130 if (processed[nn_indices[j]]) // Has this point been processed before ?
131 continue;
132
133 //processed[nn_indices[j]] = true;
134 // [-1;1]
135 const double dot_p = normals[seed_queue[idx]].getNormalVector3fMap().dot(
136 normals[nn_indices[j]].getNormalVector3fMap());
137
138 if (std::acos (dot_p) < eps_angle)
139 {
140 processed[nn_indices[j]] = true;
141 seed_queue.emplace_back (nn_indices[j]);
142 }
143 }
144 }
145
146 // If this queue is satisfactory, add to the clusters
147 if (seed_queue.size () >= min_pts_per_cluster && seed_queue.size () <= max_pts_per_cluster)
148 {
149 std::sort (r.indices.begin (), r.indices.end ());
150 r.indices.erase (std::unique (r.indices.begin (), r.indices.end ()), r.indices.end ());
151
152 // Might be better to work directly in the cluster somehow
153 clusters.emplace_back (std::move(r)); // Trying to avoid a copy by moving
154 }
155 }
156}
157
158//////////////////////////////////////////////////////////////////////////////////////////////
159template<typename PointInT, typename PointNT, typename PointOutT> void
161 const pcl::PointCloud<PointNT> & cloud,
162 pcl::Indices &indices_to_use,
163 pcl::Indices &indices_out,
164 pcl::Indices &indices_in,
165 float threshold)
166{
167 indices_out.resize (cloud.size ());
168 indices_in.resize (cloud.size ());
169
170 std::size_t in, out;
171 in = out = 0;
172
173 for (const auto &index : indices_to_use)
174 {
175 if (cloud[index].curvature > threshold)
176 {
177 indices_out[out] = index;
178 out++;
179 }
180 else
181 {
182 indices_in[in] = index;
183 in++;
184 }
185 }
186
187 indices_out.resize (out);
188 indices_in.resize (in);
189}
190
191//////////////////////////////////////////////////////////////////////////////////////////////
192template<typename PointInT, typename PointNT, typename PointOutT> void
194{
195 // Check if input was set
196 if (!normals_)
197 {
198 PCL_ERROR ("[pcl::%s::computeFeature] No input dataset containing normals was given!\n", getClassName ().c_str ());
199 output.width = output.height = 0;
200 output.clear ();
201 return;
202 }
203 if (normals_->size () != surface_->size ())
204 {
205 PCL_ERROR ("[pcl::%s::computeFeature] The number of points in the input dataset differs from the number of points in the dataset containing the normals!\n", getClassName ().c_str ());
206 output.width = output.height = 0;
207 output.clear ();
208 return;
209 }
210
211 centroids_dominant_orientations_.clear ();
212
213 // ---[ Step 0: remove normals with high curvature
214 pcl::Indices indices_out;
215 pcl::Indices indices_in;
216 filterNormalsWithHighCurvature (*normals_, *indices_, indices_out, indices_in, curv_threshold_);
217
219 normals_filtered_cloud->width = indices_in.size ();
220 normals_filtered_cloud->height = 1;
221 normals_filtered_cloud->points.resize (normals_filtered_cloud->width);
222
223 for (std::size_t i = 0; i < indices_in.size (); ++i)
224 {
225 (*normals_filtered_cloud)[i].x = (*surface_)[indices_in[i]].x;
226 (*normals_filtered_cloud)[i].y = (*surface_)[indices_in[i]].y;
227 (*normals_filtered_cloud)[i].z = (*surface_)[indices_in[i]].z;
228 }
229
230 std::vector<pcl::PointIndices> clusters;
231
232 if(normals_filtered_cloud->size() >= min_points_)
233 {
234 //recompute normals and use them for clustering
235 KdTreePtr normals_tree_filtered (new pcl::search::KdTree<pcl::PointNormal> (false));
236 normals_tree_filtered->setInputCloud (normals_filtered_cloud);
237
238
240 n3d.setRadiusSearch (radius_normals_);
241 n3d.setSearchMethod (normals_tree_filtered);
242 n3d.setInputCloud (normals_filtered_cloud);
243 n3d.compute (*normals_filtered_cloud);
244
245 KdTreePtr normals_tree (new pcl::search::KdTree<pcl::PointNormal> (false));
246 normals_tree->setInputCloud (normals_filtered_cloud);
247
248 extractEuclideanClustersSmooth (*normals_filtered_cloud,
249 *normals_filtered_cloud,
250 cluster_tolerance_,
251 normals_tree,
252 clusters,
253 eps_angle_threshold_,
254 static_cast<unsigned int> (min_points_));
255
256 }
257
258 VFHEstimator vfh;
259 vfh.setInputCloud (surface_);
260 vfh.setInputNormals (normals_);
261 vfh.setIndices(indices_);
262 vfh.setSearchMethod (this->tree_);
263 vfh.setUseGivenNormal (true);
264 vfh.setUseGivenCentroid (true);
265 vfh.setNormalizeBins (normalize_bins_);
266 vfh.setNormalizeDistance (true);
267 vfh.setFillSizeComponent (true);
268 output.height = 1;
269
270 // ---[ Step 1b : check if any dominant cluster was found
271 if (!clusters.empty ())
272 { // ---[ Step 1b.1 : If yes, compute CVFH using the cluster information
273
274 for (const auto &cluster : clusters) //for each cluster
275 {
276 Eigen::Vector4f avg_normal = Eigen::Vector4f::Zero ();
277 Eigen::Vector4f avg_centroid = Eigen::Vector4f::Zero ();
278
279 for (const auto &index : cluster.indices)
280 {
281 avg_normal += (*normals_filtered_cloud)[index].getNormalVector4fMap ();
282 avg_centroid += (*normals_filtered_cloud)[index].getVector4fMap ();
283 }
284
285 avg_normal /= static_cast<float> (cluster.indices.size ());
286 avg_centroid /= static_cast<float> (cluster.indices.size ());
287
288 Eigen::Vector4f centroid_test;
289 pcl::compute3DCentroid (*normals_filtered_cloud, centroid_test);
290 avg_normal.normalize ();
291
292 Eigen::Vector3f avg_norm (avg_normal[0], avg_normal[1], avg_normal[2]);
293 Eigen::Vector3f avg_dominant_centroid (avg_centroid[0], avg_centroid[1], avg_centroid[2]);
294
295 //append normal and centroid for the clusters
296 dominant_normals_.push_back (avg_norm);
297 centroids_dominant_orientations_.push_back (avg_dominant_centroid);
298 }
299
300 //compute modified VFH for all dominant clusters and add them to the list!
301 output.resize (dominant_normals_.size ());
302 output.width = dominant_normals_.size ();
303
304 for (std::size_t i = 0; i < dominant_normals_.size (); ++i)
305 {
306 //configure VFH computation for CVFH
307 vfh.setNormalToUse (dominant_normals_[i]);
308 vfh.setCentroidToUse (centroids_dominant_orientations_[i]);
310 vfh.compute (vfh_signature);
311 output[i] = vfh_signature[0];
312 }
313 }
314 else
315 { // ---[ Step 1b.1 : If no, compute CVFH using all the object points
316 Eigen::Vector4f avg_centroid;
317 pcl::compute3DCentroid (*surface_, avg_centroid);
318 Eigen::Vector3f cloud_centroid (avg_centroid[0], avg_centroid[1], avg_centroid[2]);
319 centroids_dominant_orientations_.push_back (cloud_centroid);
320
321 //configure VFH computation for CVFH using all object points
322 vfh.setCentroidToUse (cloud_centroid);
323 vfh.setUseGivenNormal (false);
324
326 vfh.compute (vfh_signature);
327
328 output.resize (1);
329 output.width = 1;
330
331 output[0] = vfh_signature[0];
332 }
333}
334
335#define PCL_INSTANTIATE_CVFHEstimation(T,NT,OutT) template class PCL_EXPORTS pcl::CVFHEstimation<T,NT,OutT>;
336
337#endif // PCL_FEATURES_IMPL_VFH_H_
Define methods for centroid estimation and covariance matrix calculus.
CVFHEstimation estimates the Clustered Viewpoint Feature Histogram (CVFH) descriptor for a given poin...
Definition cvfh.h:63
typename Feature< PointInT, PointOutT >::PointCloudOut PointCloudOut
Definition cvfh.h:76
void filterNormalsWithHighCurvature(const pcl::PointCloud< PointNT > &cloud, pcl::Indices &indices_to_use, pcl::Indices &indices_out, pcl::Indices &indices_in, float threshold)
Removes normals with high curvature caused by real edges or noisy data.
Definition cvfh.hpp:160
void compute(PointCloudOut &output)
Overloaded computed method from pcl::Feature.
Definition cvfh.hpp:50
Feature represents the base feature class.
Definition feature.h:107
void setRadiusSearch(double radius)
Set the sphere radius that is to be used for determining the nearest neighbors used for the feature e...
Definition feature.h:198
void setSearchMethod(const KdTreePtr &tree)
Provide a pointer to the search object.
Definition feature.h:164
void compute(PointCloudOut &output)
Base method for feature estimation for all points given in <setInputCloud (), setIndices ()> using th...
Definition feature.hpp:194
NormalEstimation estimates local surface properties (surface normals and curvatures)at each 3D point.
Definition normal_3d.h:244
void setInputCloud(const PointCloudConstPtr &cloud) override
Provide a pointer to the input dataset.
Definition normal_3d.h:332
PointCloud represents the base class in PCL for storing collections of 3D points.
void resize(std::size_t count)
Resizes the container to contain count elements.
reference emplace_back(Args &&...args)
Emplace a new point in the cloud, at the end of the container.
pcl::PCLHeader header
The point cloud header.
std::size_t size() const
shared_ptr< PointCloud< PointT > > Ptr
search::KdTree is a wrapper class which inherits the pcl::KdTree class for performing search function...
Definition kdtree.h:62
shared_ptr< pcl::search::Search< PointT > > Ptr
Definition search.h:81
virtual PointCloudConstPtr getInputCloud() const
Get a pointer to the input point cloud dataset.
Definition search.h:123
virtual int radiusSearch(const PointT &point, double radius, Indices &k_indices, std::vector< float > &k_sqr_distances, unsigned int max_nn=0) const =0
Search for all the nearest neighbors of the query point in a given radius.
unsigned int compute3DCentroid(ConstCloudIterator< PointT > &cloud_iterator, Eigen::Matrix< Scalar, 4, 1 > &centroid)
Compute the 3D (X-Y-Z) centroid of a set of points and return it as a 3D vector.
Definition centroid.hpp:56
IndicesAllocator<> Indices
Type used for indices in PCL.
Definition types.h:133
::pcl::PCLHeader header