Point Cloud Library (PCL) 1.15.1-dev
Loading...
Searching...
No Matches
convex_hull.hpp
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 *
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 Willow Garage, Inc. 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 * $Id$
37 *
38 */
39
40#include <pcl/pcl_config.h>
41#ifdef HAVE_QHULL
42
43#ifndef PCL_SURFACE_IMPL_CONVEX_HULL_H_
44#define PCL_SURFACE_IMPL_CONVEX_HULL_H_
45
46#include <pcl/surface/convex_hull.h>
47#include <pcl/common/common.h>
48#include <pcl/common/eigen.h>
49#include <pcl/common/transforms.h>
50#include <pcl/common/io.h>
51#include <cstdio>
52#include <cstdlib>
53#include <pcl/surface/qhull.h>
54
55//////////////////////////////////////////////////////////////////////////
56template <typename PointInT> void
58{
59 PCL_DEBUG ("[pcl::%s::calculateInputDimension] WARNING: Input dimension not specified. Automatically determining input dimension.\n", getClassName ().c_str ());
60 Eigen::Vector4d xyz_centroid;
61 compute3DCentroid (*input_, *indices_, xyz_centroid);
62 EIGEN_ALIGN16 Eigen::Matrix3d covariance_matrix = Eigen::Matrix3d::Zero ();
63 computeCovarianceMatrixNormalized (*input_, *indices_, xyz_centroid, covariance_matrix);
64
65 EIGEN_ALIGN16 Eigen::Vector3d eigen_values;
66 pcl::eigen33 (covariance_matrix, eigen_values);
67
68 if (std::abs (eigen_values[0]) < std::numeric_limits<double>::epsilon () || std::abs (eigen_values[0] / eigen_values[2]) < 1.0e-3)
69 dimension_ = 2;
70 else
71 dimension_ = 3;
72}
73
74//////////////////////////////////////////////////////////////////////////
75template <typename PointInT> void
76pcl::ConvexHull<PointInT>::performReconstruction2D (PointCloud &hull, std::vector<pcl::Vertices> &polygons,
77 bool)
78{
79 int dimension = 2;
80 bool xy_proj_safe = true;
81 bool yz_proj_safe = true;
82 bool xz_proj_safe = true;
83
84 Eigen::Vector4d normal_calc_centroid;
85 Eigen::Matrix3d normal_calc_covariance;
86
87 // Check the input's normal to see which projection to use
88 PointInT p0 = (*input_)[(*indices_)[0]];
89 PointInT p1 = (*input_)[(*indices_)[indices_->size () - 1]];
90 PointInT p2 = (*input_)[(*indices_)[indices_->size () / 2]];
92 (p1.getVector3fMap() - p0.getVector3fMap())
93 .cross(p2.getVector3fMap() - p0.getVector3fMap())
94 .stableNorm() > Eigen::NumTraits<float>::dummy_precision()) {
95 pcl::PointCloud<PointInT> normal_calc_cloud;
96 normal_calc_cloud.resize(3);
97 normal_calc_cloud[0] = p0;
98 normal_calc_cloud[1] = p1;
99 normal_calc_cloud[2] = p2;
100
101 pcl::compute3DCentroid(normal_calc_cloud, normal_calc_centroid);
103 normal_calc_cloud, normal_calc_centroid, normal_calc_covariance);
104 }
105 else {
106 // Three points do not form a valid triangle, fallback to use all points to
107 // calculate the covariance matrix
108 pcl::compute3DCentroid(*input_, *indices_, normal_calc_centroid);
110 *input_, *indices_, normal_calc_centroid, normal_calc_covariance);
111 }
112
113 // Need to set -1 here. See eigen33 for explanations.
114 Eigen::Vector3d::Scalar eigen_value;
115 Eigen::Vector3d plane_params;
116 pcl::eigen33 (normal_calc_covariance, eigen_value, plane_params);
117 float theta_x = std::abs (static_cast<float> (plane_params.dot (x_axis_)));
118 float theta_y = std::abs (static_cast<float> (plane_params.dot (y_axis_)));
119 float theta_z = std::abs (static_cast<float> (plane_params.dot (z_axis_)));
120
121 // Check for degenerate cases of each projection
122 // We must avoid projections in which the plane projects as a line
123 if (theta_z > projection_angle_thresh_)
124 {
125 xz_proj_safe = false;
126 yz_proj_safe = false;
127 }
128 if (theta_x > projection_angle_thresh_)
129 {
130 xz_proj_safe = false;
131 xy_proj_safe = false;
132 }
133 if (theta_y > projection_angle_thresh_)
134 {
135 xy_proj_safe = false;
136 yz_proj_safe = false;
137 }
138
139 // True if qhull should free points in qh_freeqhull() or reallocation
140 boolT ismalloc = True;
141 // output from qh_produce_output(), use NULL to skip qh_produce_output()
142 FILE *outfile = nullptr;
143
145 outfile = stderr;
146
147 // option flags for qhull, see qh_opt.htm
148 const char* flags = qhull_flags.c_str ();
149 // error messages from qhull code
150 FILE *errfile = stderr;
151
152 // Array of coordinates for each point
153 coordT *points = reinterpret_cast<coordT*> (calloc (indices_->size () * dimension, sizeof (coordT)));
154
155 // Build input data, using appropriate projection
156 int j = 0;
157 if (xy_proj_safe)
158 {
159 for (std::size_t i = 0; i < indices_->size (); ++i, j+=dimension)
160 {
161 points[j + 0] = static_cast<coordT> ((*input_)[(*indices_)[i]].x);
162 points[j + 1] = static_cast<coordT> ((*input_)[(*indices_)[i]].y);
163 }
164 }
165 else if (yz_proj_safe)
166 {
167 for (std::size_t i = 0; i < indices_->size (); ++i, j+=dimension)
168 {
169 points[j + 0] = static_cast<coordT> ((*input_)[(*indices_)[i]].y);
170 points[j + 1] = static_cast<coordT> ((*input_)[(*indices_)[i]].z);
171 }
172 }
173 else if (xz_proj_safe)
174 {
175 for (std::size_t i = 0; i < indices_->size (); ++i, j+=dimension)
176 {
177 points[j + 0] = static_cast<coordT> ((*input_)[(*indices_)[i]].x);
178 points[j + 1] = static_cast<coordT> ((*input_)[(*indices_)[i]].z);
179 }
180 }
181 else
182 {
183 // This should only happen if we had invalid input
184 PCL_ERROR ("[pcl::%s::performReconstruction2D] Invalid input!\n", getClassName ().c_str ());
185 }
186
187 qhT qh_qh;
188 qhT* qh = &qh_qh;
189 QHULL_LIB_CHECK
190 qh_zero(qh, errfile);
191
192 // Compute convex hull
193 int exitcode = qh_new_qhull (qh, dimension, static_cast<int> (indices_->size ()), points, ismalloc, const_cast<char*> (flags), outfile, errfile);
194 if (compute_area_)
195 {
196 qh_prepare_output(qh);
197 }
198
199 // 0 if no error from qhull or it doesn't find any vertices
200 if (exitcode != 0 || qh->num_vertices == 0)
201 {
202 PCL_ERROR ("[pcl::%s::performReconstrution2D] ERROR: qhull was unable to compute a convex hull for the given point cloud (%lu)!\n", getClassName ().c_str (), indices_->size ());
203
204 hull.resize (0);
205 hull.width = hull.height = 0;
206 polygons.resize (0);
207
208 qh_freeqhull (qh, !qh_ALL);
209 int curlong, totlong;
210 qh_memfreeshort (qh, &curlong, &totlong);
211
212 return;
213 }
214
215 // Qhull returns the area in volume for 2D
216 if (compute_area_)
217 {
218 total_area_ = qh->totvol;
219 total_volume_ = 0.0;
220 }
221
222 int num_vertices = qh->num_vertices;
223
224 hull.clear();
225 hull.resize(num_vertices, PointInT{});
226
227 vertexT * vertex;
228 int i = 0;
229
230 AlignedVector<std::pair<int, Eigen::Vector4f>> idx_points (num_vertices);
231
232 FORALLvertices
233 {
234 hull[i] = (*input_)[(*indices_)[qh_pointid (qh, vertex->point)]];
235 idx_points[i].first = qh_pointid (qh, vertex->point);
236 ++i;
237 }
238
239 // Sort
240 Eigen::Vector4f centroid;
241 pcl::compute3DCentroid (hull, centroid);
242 if (xy_proj_safe)
243 {
244 for (std::size_t j = 0; j < hull.size (); j++)
245 {
246 idx_points[j].second[0] = hull[j].x - centroid[0];
247 idx_points[j].second[1] = hull[j].y - centroid[1];
248 }
249 }
250 else if (yz_proj_safe)
251 {
252 for (std::size_t j = 0; j < hull.size (); j++)
253 {
254 idx_points[j].second[0] = hull[j].y - centroid[1];
255 idx_points[j].second[1] = hull[j].z - centroid[2];
256 }
257 }
258 else if (xz_proj_safe)
259 {
260 for (std::size_t j = 0; j < hull.size (); j++)
261 {
262 idx_points[j].second[0] = hull[j].x - centroid[0];
263 idx_points[j].second[1] = hull[j].z - centroid[2];
264 }
265 }
266 std::sort (idx_points.begin (), idx_points.end (), comparePoints2D);
267
268 polygons.resize (1);
269 polygons[0].vertices.resize (hull.size ());
270
271 hull_indices_.header = input_->header;
272 hull_indices_.indices.clear ();
273 hull_indices_.indices.reserve (hull.size ());
274
275 for (int j = 0; j < static_cast<int> (hull.size ()); j++)
276 {
277 hull_indices_.indices.push_back ((*indices_)[idx_points[j].first]);
278 hull[j] = (*input_)[(*indices_)[idx_points[j].first]];
279 polygons[0].vertices[j] = static_cast<unsigned int> (j);
280 }
281
282 qh_freeqhull (qh, !qh_ALL);
283 int curlong, totlong;
284 qh_memfreeshort (qh, &curlong, &totlong);
285
286 hull.width = hull.size ();
287 hull.height = 1;
288 hull.is_dense = false;
289 return;
290}
291
292#ifdef __GNUC__
293#pragma GCC diagnostic ignored "-Wold-style-cast"
294#endif
295//////////////////////////////////////////////////////////////////////////
296template <typename PointInT> void
298 PointCloud &hull, std::vector<pcl::Vertices> &polygons, bool fill_polygon_data)
299{
300 int dimension = 3;
301
302 // True if qhull should free points in qh_freeqhull() or reallocation
303 boolT ismalloc = True;
304 // output from qh_produce_output(), use NULL to skip qh_produce_output()
305 FILE *outfile = nullptr;
306
308 outfile = stderr;
309
310 // option flags for qhull, see qh_opt.htm
311 const char *flags = qhull_flags.c_str ();
312 // error messages from qhull code
313 FILE *errfile = stderr;
314
315 // Array of coordinates for each point
316 coordT *points = reinterpret_cast<coordT*> (calloc (indices_->size () * dimension, sizeof (coordT)));
317
318 int j = 0;
319 for (std::size_t i = 0; i < indices_->size (); ++i, j+=dimension)
320 {
321 points[j + 0] = static_cast<coordT> ((*input_)[(*indices_)[i]].x);
322 points[j + 1] = static_cast<coordT> ((*input_)[(*indices_)[i]].y);
323 points[j + 2] = static_cast<coordT> ((*input_)[(*indices_)[i]].z);
324 }
325
326 qhT qh_qh;
327 qhT* qh = &qh_qh;
328 QHULL_LIB_CHECK
329 qh_zero(qh, errfile);
330
331 // Compute convex hull
332 int exitcode = qh_new_qhull (qh, dimension, static_cast<int> (indices_->size ()), points, ismalloc, const_cast<char*> (flags), outfile, errfile);
333 if (compute_area_)
334 {
335 qh_prepare_output(qh);
336 }
337
338 // 0 if no error from qhull
339 if (exitcode != 0)
340 {
341 PCL_ERROR("[pcl::%s::performReconstrution3D] ERROR: qhull was unable to compute a "
342 "convex hull for the given point cloud (%zu)!\n",
343 getClassName().c_str(),
344 static_cast<std::size_t>(input_->size()));
345
346 hull.resize (0);
347 hull.width = hull.height = 0;
348 polygons.resize (0);
349
350 qh_freeqhull (qh, !qh_ALL);
351 int curlong, totlong;
352 qh_memfreeshort (qh, &curlong, &totlong);
353
354 return;
355 }
356
357 qh_triangulate (qh);
358
359 int num_facets = qh->num_facets;
360
361 int num_vertices = qh->num_vertices;
362 hull.resize (num_vertices);
363
364 vertexT * vertex;
365 int i = 0;
366 // Max vertex id
367 unsigned int max_vertex_id = 0;
368 FORALLvertices
369 {
370 if (vertex->id + 1 > max_vertex_id)
371 max_vertex_id = vertex->id + 1;
372 }
373
374 ++max_vertex_id;
375 std::vector<int> qhid_to_pcidx (max_vertex_id);
376
377 hull_indices_.header = input_->header;
378 hull_indices_.indices.clear ();
379 hull_indices_.indices.reserve (num_vertices);
380
381 FORALLvertices
382 {
383 // Add vertices to hull point_cloud and store index
384 hull_indices_.indices.push_back ((*indices_)[qh_pointid (qh, vertex->point)]);
385 hull[i] = (*input_)[hull_indices_.indices.back ()];
386
387 qhid_to_pcidx[vertex->id] = i; // map the vertex id of qhull to the point cloud index
388 ++i;
389 }
390
391 if (compute_area_)
392 {
393 total_area_ = qh->totarea;
394 total_volume_ = qh->totvol;
395 }
396
397 if (fill_polygon_data)
398 {
399 polygons.resize (num_facets);
400 int dd = 0;
401
402 facetT * facet;
403 FORALLfacets
404 {
405 polygons[dd].vertices.resize (3);
406
407 // Needed by FOREACHvertex_i_
408 int vertex_n, vertex_i;
409 FOREACHvertex_i_ (qh, (*facet).vertices)
410 //facet_vertices.vertices.push_back (qhid_to_pcidx[vertex->id]);
411 polygons[dd].vertices[vertex_i] = qhid_to_pcidx[vertex->id];
412 ++dd;
413 }
414 }
415 // Deallocates memory (also the points)
416 qh_freeqhull (qh, !qh_ALL);
417 int curlong, totlong;
418 qh_memfreeshort (qh, &curlong, &totlong);
419
420 hull.width = hull.size ();
421 hull.height = 1;
422 hull.is_dense = false;
423}
424#ifdef __GNUC__
425#pragma GCC diagnostic warning "-Wold-style-cast"
426#endif
427
428//////////////////////////////////////////////////////////////////////////
429template <typename PointInT> void
430pcl::ConvexHull<PointInT>::performReconstruction (PointCloud &hull, std::vector<pcl::Vertices> &polygons,
431 bool fill_polygon_data)
432{
433 if (dimension_ == 0)
434 calculateInputDimension ();
435 if (dimension_ == 2)
436 performReconstruction2D (hull, polygons, fill_polygon_data);
437 else if (dimension_ == 3)
438 performReconstruction3D (hull, polygons, fill_polygon_data);
439 else
440 PCL_ERROR ("[pcl::%s::performReconstruction] Error: invalid input dimension requested: %d\n",getClassName ().c_str (),dimension_);
441}
442
443//////////////////////////////////////////////////////////////////////////
444template <typename PointInT> void
446{
447 points.header = input_->header;
448 if (!initCompute () || input_->points.empty () || indices_->empty ())
449 {
450 points.clear ();
451 return;
452 }
453
454 // Perform the actual surface reconstruction
455 std::vector<pcl::Vertices> polygons;
456 performReconstruction (points, polygons, false);
457
458 points.width = points.size ();
459 points.height = 1;
460 points.is_dense = true;
461
462 deinitCompute ();
463}
464
465
466//////////////////////////////////////////////////////////////////////////
467template <typename PointInT> void
469{
470 // Perform reconstruction
471 pcl::PointCloud<PointInT> hull_points;
472 performReconstruction (hull_points, output.polygons, true);
473
474 // Convert the PointCloud into a PCLPointCloud2
475 pcl::toPCLPointCloud2 (hull_points, output.cloud);
476}
477
478//////////////////////////////////////////////////////////////////////////
479template <typename PointInT> void
480pcl::ConvexHull<PointInT>::performReconstruction (std::vector<pcl::Vertices> &polygons)
481{
482 pcl::PointCloud<PointInT> hull_points;
483 performReconstruction (hull_points, polygons, true);
484}
485
486//////////////////////////////////////////////////////////////////////////
487template <typename PointInT> void
488pcl::ConvexHull<PointInT>::reconstruct (PointCloud &points, std::vector<pcl::Vertices> &polygons)
489{
490 points.header = input_->header;
491 if (!initCompute () || input_->points.empty () || indices_->empty ())
492 {
493 points.clear ();
494 return;
495 }
496
497 // Perform the actual surface reconstruction
498 performReconstruction (points, polygons, true);
499
500 points.width = points.size ();
501 points.height = 1;
502 points.is_dense = true;
503
504 deinitCompute ();
505}
506//////////////////////////////////////////////////////////////////////////
507template <typename PointInT> void
509{
510 hull_point_indices = hull_indices_;
511}
512
513#define PCL_INSTANTIATE_ConvexHull(T) template class PCL_EXPORTS pcl::ConvexHull<T>;
514
515#endif // PCL_SURFACE_IMPL_CONVEX_HULL_H_
516#endif
void calculateInputDimension()
Automatically determines the dimension of input data - 2D or 3D.
void performReconstruction2D(PointCloud &points, std::vector< pcl::Vertices > &polygons, bool fill_polygon_data=false)
The reconstruction method for 2D data.
void getHullPointIndices(pcl::PointIndices &hull_point_indices) const
Retrieve the indices of the input point cloud that for the convex hull.
void performReconstruction(PointCloud &points, std::vector< pcl::Vertices > &polygons, bool fill_polygon_data=false)
The actual reconstruction method.
void reconstruct(PointCloud &points, std::vector< pcl::Vertices > &polygons)
Compute a convex hull for all points given.
void performReconstruction3D(PointCloud &points, std::vector< pcl::Vertices > &polygons, bool fill_polygon_data=false)
The reconstruction method for 3D data.
bool is_dense
True if no points are invalid (e.g., have NaN or Inf values in any of their floating point fields).
void resize(std::size_t count)
Resizes the container to contain count elements.
std::uint32_t width
The point cloud width (if organized as an image-structure).
pcl::PCLHeader header
The point cloud header.
std::uint32_t height
The point cloud height (if organized as an image-structure).
void clear()
Removes all points in a cloud and sets the width and height to 0.
std::size_t size() const
Define standard C methods and C++ classes that are common to all methods.
unsigned int computeCovarianceMatrixNormalized(const pcl::PointCloud< PointT > &cloud, const Eigen::Matrix< Scalar, 4, 1 > &centroid, Eigen::Matrix< Scalar, 3, 3 > &covariance_matrix)
Compute normalized the 3x3 covariance matrix of a given set of points.
Definition centroid.hpp:269
void eigen33(const Matrix &mat, typename Matrix::Scalar &eigenvalue, Vector &eigenvector)
determines the eigenvector and eigenvalue of the smallest eigenvalue of the symmetric positive semi d...
Definition eigen.hpp:295
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:57
bool comparePoints2D(const std::pair< int, Eigen::Vector4f > &p1, const std::pair< int, Eigen::Vector4f > &p2)
Sort 2D points in a vector structure.
Definition convex_hull.h:59
PCL_EXPORTS bool isVerbosityLevelEnabled(VERBOSITY_LEVEL severity)
is verbosity level enabled?
void toPCLPointCloud2(const pcl::PointCloud< PointT > &cloud, pcl::PCLPointCloud2 &msg, bool padding)
Convert a pcl::PointCloud<T> object to a PCLPointCloud2 binary data blob.
std::vector< T, Eigen::aligned_allocator< T > > AlignedVector
Type used for aligned vector of Eigen objects in PCL.
Definition types.h:139
constexpr bool isXYZFinite(const PointT &) noexcept
std::vector< ::pcl::Vertices > polygons
Definition PolygonMesh.h:22
::pcl::PCLPointCloud2 cloud
Definition PolygonMesh.h:20