Point Cloud Library (PCL)  1.12.1-dev
sac_model_plane.hpp
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2009, 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_SAMPLE_CONSENSUS_IMPL_SAC_MODEL_PLANE_H_
42 #define PCL_SAMPLE_CONSENSUS_IMPL_SAC_MODEL_PLANE_H_
43 
44 #include <pcl/sample_consensus/sac_model_plane.h>
45 #include <pcl/common/centroid.h>
46 #include <pcl/common/eigen.h>
47 #include <pcl/common/concatenate.h>
48 
49 //////////////////////////////////////////////////////////////////////////
50 template <typename PointT> bool
52 {
53  if (samples.size () != sample_size_)
54  {
55  PCL_ERROR ("[pcl::SampleConsensusModelPlane::isSampleGood] Wrong number of samples (is %lu, should be %lu)!\n", samples.size (), sample_size_);
56  return (false);
57  }
58 
59  // Check if the sample points are collinear
60  // Similar checks are implemented as precaution in computeModelCoefficients,
61  // so if you find the need to fix something in here, look there, too.
62  pcl::Vector3fMapConst p0 = (*input_)[samples[0]].getVector3fMap ();
63  pcl::Vector3fMapConst p1 = (*input_)[samples[1]].getVector3fMap ();
64  pcl::Vector3fMapConst p2 = (*input_)[samples[2]].getVector3fMap ();
65 
66  // Check if the norm of the cross-product would be non-zero, otherwise
67  // normalization will fail. One could also interpret this as kind of check
68  // if the triangle spanned by those three points would have an area greater
69  // than zero.
70  if ((p1 - p0).cross(p2 - p0).stableNorm() < Eigen::NumTraits<float>::dummy_precision ())
71  {
72  PCL_ERROR ("[pcl::SampleConsensusModelPlane::isSampleGood] Sample points too similar or collinear!\n");
73  return (false);
74  }
75 
76  return (true);
77 }
78 
79 //////////////////////////////////////////////////////////////////////////
80 template <typename PointT> bool
82  const Indices &samples, Eigen::VectorXf &model_coefficients) const
83 {
84  // The checks are redundant with isSampleGood above, but since most of the
85  // computed values are also used to compute the model coefficients, this might
86  // be a situation where this duplication is acceptable.
87  if (samples.size () != sample_size_)
88  {
89  PCL_ERROR ("[pcl::SampleConsensusModelPlane::computeModelCoefficients] Invalid set of samples given (%lu)!\n", samples.size ());
90  return (false);
91  }
92 
93  pcl::Vector3fMapConst p0 = (*input_)[samples[0]].getVector3fMap ();
94  pcl::Vector3fMapConst p1 = (*input_)[samples[1]].getVector3fMap ();
95  pcl::Vector3fMapConst p2 = (*input_)[samples[2]].getVector3fMap ();
96 
97  const Eigen::Vector3f cross = (p1 - p0).cross(p2 - p0);
98  const float crossNorm = cross.stableNorm();
99 
100  // Checking for collinearity here
101  if (crossNorm < Eigen::NumTraits<float>::dummy_precision ())
102  {
103  PCL_ERROR ("[pcl::SampleConsensusModelPlane::computeModelCoefficients] Chosen samples are collinear!\n");
104  return (false);
105  }
106 
107  // Compute the plane coefficients from the 3 given points in a straightforward manner
108  // calculate the plane normal n = (p2-p1) x (p3-p1) = cross (p2-p1, p3-p1)
109  model_coefficients.resize (model_size_);
110  model_coefficients.template head<3>() = cross / crossNorm;
111 
112  // ... + d = 0
113  model_coefficients[3] = -1.0f * (model_coefficients.template head<3>().dot (p0));
114 
115  PCL_DEBUG ("[pcl::SampleConsensusModelPlane::computeModelCoefficients] Model is (%g,%g,%g,%g).\n",
116  model_coefficients[0], model_coefficients[1], model_coefficients[2], model_coefficients[3]);
117  return (true);
118 }
119 
120 #define AT(POS) ((*input_)[(*indices_)[(POS)]])
121 
122 #ifdef __AVX__
123 // This function computes the distances of 8 points to the plane
124 template <typename PointT> inline __m256 pcl::SampleConsensusModelPlane<PointT>::dist8 (const std::size_t i, const __m256 &a_vec, const __m256 &b_vec, const __m256 &c_vec, const __m256 &d_vec, const __m256 &abs_help) const
125 {
126  // The andnot-function realizes an abs-operation: the sign bit is removed
127  return _mm256_andnot_ps (abs_help,
128  _mm256_add_ps (_mm256_add_ps (_mm256_mul_ps (a_vec, _mm256_set_ps (AT(i ).x, AT(i+1).x, AT(i+2).x, AT(i+3).x, AT(i+4).x, AT(i+5).x, AT(i+6).x, AT(i+7).x)),
129  _mm256_mul_ps (b_vec, _mm256_set_ps (AT(i ).y, AT(i+1).y, AT(i+2).y, AT(i+3).y, AT(i+4).y, AT(i+5).y, AT(i+6).y, AT(i+7).y))),
130  _mm256_add_ps (_mm256_mul_ps (c_vec, _mm256_set_ps (AT(i ).z, AT(i+1).z, AT(i+2).z, AT(i+3).z, AT(i+4).z, AT(i+5).z, AT(i+6).z, AT(i+7).z)),
131  d_vec))); // TODO this could be replaced by three fmadd-instructions (if available), but the speed gain would probably be minimal
132 }
133 #endif // ifdef __AVX__
134 
135 #ifdef __SSE__
136 // This function computes the distances of 4 points to the plane
137 template <typename PointT> inline __m128 pcl::SampleConsensusModelPlane<PointT>::dist4 (const std::size_t i, const __m128 &a_vec, const __m128 &b_vec, const __m128 &c_vec, const __m128 &d_vec, const __m128 &abs_help) const
138 {
139  // The andnot-function realizes an abs-operation: the sign bit is removed
140  return _mm_andnot_ps (abs_help,
141  _mm_add_ps (_mm_add_ps (_mm_mul_ps (a_vec, _mm_set_ps (AT(i ).x, AT(i+1).x, AT(i+2).x, AT(i+3).x)),
142  _mm_mul_ps (b_vec, _mm_set_ps (AT(i ).y, AT(i+1).y, AT(i+2).y, AT(i+3).y))),
143  _mm_add_ps (_mm_mul_ps (c_vec, _mm_set_ps (AT(i ).z, AT(i+1).z, AT(i+2).z, AT(i+3).z)),
144  d_vec))); // TODO this could be replaced by three fmadd-instructions (if available), but the speed gain would probably be minimal
145 }
146 #endif // ifdef __SSE__
147 
148 #undef AT
149 
150 //////////////////////////////////////////////////////////////////////////
151 template <typename PointT> void
153  const Eigen::VectorXf &model_coefficients, std::vector<double> &distances) const
154 {
155  // Needs a valid set of model coefficients
156  if (!isModelValid (model_coefficients))
157  {
158  PCL_ERROR ("[pcl::SampleConsensusModelPlane::getDistancesToModel] Given model is invalid!\n");
159  return;
160  }
161 
162  distances.resize (indices_->size ());
163 
164  // Iterate through the 3d points and calculate the distances from them to the plane
165  for (std::size_t i = 0; i < indices_->size (); ++i)
166  {
167  // Calculate the distance from the point to the plane normal as the dot product
168  // D = (P-A).N/|N|
169  /*distances[i] = std::abs (model_coefficients[0] * (*input_)[(*indices_)[i]].x +
170  model_coefficients[1] * (*input_)[(*indices_)[i]].y +
171  model_coefficients[2] * (*input_)[(*indices_)[i]].z +
172  model_coefficients[3]);*/
173  Eigen::Vector4f pt ((*input_)[(*indices_)[i]].x,
174  (*input_)[(*indices_)[i]].y,
175  (*input_)[(*indices_)[i]].z,
176  1.0f);
177  distances[i] = std::abs (model_coefficients.dot (pt));
178  }
179 }
180 
181 //////////////////////////////////////////////////////////////////////////
182 template <typename PointT> void
184  const Eigen::VectorXf &model_coefficients, const double threshold, Indices &inliers)
185 {
186  // Needs a valid set of model coefficients
187  if (!isModelValid (model_coefficients))
188  {
189  PCL_ERROR ("[pcl::SampleConsensusModelPlane::selectWithinDistance] Given model is invalid!\n");
190  return;
191  }
192 
193  inliers.clear ();
194  error_sqr_dists_.clear ();
195  inliers.reserve (indices_->size ());
196  error_sqr_dists_.reserve (indices_->size ());
197 
198  // Iterate through the 3d points and calculate the distances from them to the plane
199  for (std::size_t i = 0; i < indices_->size (); ++i)
200  {
201  // Calculate the distance from the point to the plane normal as the dot product
202  // D = (P-A).N/|N|
203  Eigen::Vector4f pt ((*input_)[(*indices_)[i]].x,
204  (*input_)[(*indices_)[i]].y,
205  (*input_)[(*indices_)[i]].z,
206  1.0f);
207 
208  float distance = std::abs (model_coefficients.dot (pt));
209 
210  if (distance < threshold)
211  {
212  // Returns the indices of the points whose distances are smaller than the threshold
213  inliers.push_back ((*indices_)[i]);
214  error_sqr_dists_.push_back (static_cast<double> (distance));
215  }
216  }
217 }
218 
219 //////////////////////////////////////////////////////////////////////////
220 template <typename PointT> std::size_t
222  const Eigen::VectorXf &model_coefficients, const double threshold) const
223 {
224  // Needs a valid set of model coefficients
225  if (!isModelValid (model_coefficients))
226  {
227  PCL_ERROR ("[pcl::SampleConsensusModelPlane::countWithinDistance] Given model is invalid!\n");
228  return (0);
229  }
230 #if defined (__AVX__) && defined (__AVX2__)
231  return countWithinDistanceAVX (model_coefficients, threshold);
232 #elif defined (__SSE__) && defined (__SSE2__) && defined (__SSE4_1__)
233  return countWithinDistanceSSE (model_coefficients, threshold);
234 #else
235  return countWithinDistanceStandard (model_coefficients, threshold);
236 #endif
237 }
238 
239 //////////////////////////////////////////////////////////////////////////
240 template <typename PointT> std::size_t
242  const Eigen::VectorXf &model_coefficients, const double threshold, std::size_t i) const
243 {
244  std::size_t nr_p = 0;
245  // Iterate through the 3d points and calculate the distances from them to the plane
246  for (; i < indices_->size (); ++i)
247  {
248  // Calculate the distance from the point to the plane normal as the dot product
249  // D = (P-A).N/|N|
250  Eigen::Vector4f pt ((*input_)[(*indices_)[i]].x,
251  (*input_)[(*indices_)[i]].y,
252  (*input_)[(*indices_)[i]].z,
253  1.0f);
254  if (std::abs (model_coefficients.dot (pt)) < threshold)
255  {
256  nr_p++;
257  }
258  }
259  return (nr_p);
260 }
261 
262 //////////////////////////////////////////////////////////////////////////
263 #if defined (__SSE__) && defined (__SSE2__) && defined (__SSE4_1__)
264 template <typename PointT> std::size_t
266  const Eigen::VectorXf &model_coefficients, const double threshold, std::size_t i) const
267 {
268  std::size_t nr_p = 0;
269  const __m128 a_vec = _mm_set1_ps (model_coefficients[0]);
270  const __m128 b_vec = _mm_set1_ps (model_coefficients[1]);
271  const __m128 c_vec = _mm_set1_ps (model_coefficients[2]);
272  const __m128 d_vec = _mm_set1_ps (model_coefficients[3]);
273  const __m128 threshold_vec = _mm_set1_ps (threshold);
274  const __m128 abs_help = _mm_set1_ps (-0.0F); // -0.0F (negative zero) means that all bits are 0, only the sign bit is 1
275  __m128i res = _mm_set1_epi32(0); // This corresponds to nr_p: 4 32bit integers that, summed together, hold the number of inliers
276  for (; (i + 4) <= indices_->size (); i += 4)
277  {
278  const __m128 mask = _mm_cmplt_ps (dist4 (i, a_vec, b_vec, c_vec, d_vec, abs_help), threshold_vec); // The mask contains 1 bits if the corresponding points are inliers, else 0 bits
279  res = _mm_add_epi32 (res, _mm_and_si128 (_mm_set1_epi32 (1), _mm_castps_si128 (mask))); // The latter part creates a vector with ones (as 32bit integers) where the points are inliers
280  //const int res = _mm_movemask_ps (mask);
281  //if (res & 1) nr_p++;
282  //if (res & 2) nr_p++;
283  //if (res & 4) nr_p++;
284  //if (res & 8) nr_p++;
285  }
286  nr_p += _mm_extract_epi32 (res, 0);
287  nr_p += _mm_extract_epi32 (res, 1);
288  nr_p += _mm_extract_epi32 (res, 2);
289  nr_p += _mm_extract_epi32 (res, 3);
290 
291  // Process the remaining points (at most 3)
292  nr_p += countWithinDistanceStandard(model_coefficients, threshold, i);
293  return (nr_p);
294 }
295 #endif
296 
297 //////////////////////////////////////////////////////////////////////////
298 #if defined (__AVX__) && defined (__AVX2__)
299 template <typename PointT> std::size_t
301  const Eigen::VectorXf &model_coefficients, const double threshold, std::size_t i) const
302 {
303  std::size_t nr_p = 0;
304  const __m256 a_vec = _mm256_set1_ps (model_coefficients[0]);
305  const __m256 b_vec = _mm256_set1_ps (model_coefficients[1]);
306  const __m256 c_vec = _mm256_set1_ps (model_coefficients[2]);
307  const __m256 d_vec = _mm256_set1_ps (model_coefficients[3]);
308  const __m256 threshold_vec = _mm256_set1_ps (threshold);
309  const __m256 abs_help = _mm256_set1_ps (-0.0F); // -0.0F (negative zero) means that all bits are 0, only the sign bit is 1
310  __m256i res = _mm256_set1_epi32(0); // This corresponds to nr_p: 8 32bit integers that, summed together, hold the number of inliers
311  for (; (i + 8) <= indices_->size (); i += 8)
312  {
313  const __m256 mask = _mm256_cmp_ps (dist8 (i, a_vec, b_vec, c_vec, d_vec, abs_help), threshold_vec, _CMP_LT_OQ); // The mask contains 1 bits if the corresponding points are inliers, else 0 bits
314  res = _mm256_add_epi32 (res, _mm256_and_si256 (_mm256_set1_epi32 (1), _mm256_castps_si256 (mask))); // The latter part creates a vector with ones (as 32bit integers) where the points are inliers
315  //const int res = _mm256_movemask_ps (mask);
316  //if (res & 1) nr_p++;
317  //if (res & 2) nr_p++;
318  //if (res & 4) nr_p++;
319  //if (res & 8) nr_p++;
320  //if (res & 16) nr_p++;
321  //if (res & 32) nr_p++;
322  //if (res & 64) nr_p++;
323  //if (res & 128) nr_p++;
324  }
325  nr_p += _mm256_extract_epi32 (res, 0);
326  nr_p += _mm256_extract_epi32 (res, 1);
327  nr_p += _mm256_extract_epi32 (res, 2);
328  nr_p += _mm256_extract_epi32 (res, 3);
329  nr_p += _mm256_extract_epi32 (res, 4);
330  nr_p += _mm256_extract_epi32 (res, 5);
331  nr_p += _mm256_extract_epi32 (res, 6);
332  nr_p += _mm256_extract_epi32 (res, 7);
333 
334  // Process the remaining points (at most 7)
335  nr_p += countWithinDistanceStandard(model_coefficients, threshold, i);
336  return (nr_p);
337 }
338 #endif
339 
340 //////////////////////////////////////////////////////////////////////////
341 template <typename PointT> void
343  const Indices &inliers, const Eigen::VectorXf &model_coefficients, Eigen::VectorXf &optimized_coefficients) const
344 {
345  // Needs a valid set of model coefficients
346  if (!isModelValid (model_coefficients))
347  {
348  PCL_ERROR ("[pcl::SampleConsensusModelPlane::optimizeModelCoefficients] Given model is invalid!\n");
349  optimized_coefficients = model_coefficients;
350  return;
351  }
352 
353  // Need more than the minimum sample size to make a difference
354  if (inliers.size () <= sample_size_)
355  {
356  PCL_ERROR ("[pcl::SampleConsensusModelPlane::optimizeModelCoefficients] Not enough inliers found to optimize model coefficients (%lu)! Returning the same coefficients.\n", inliers.size ());
357  optimized_coefficients = model_coefficients;
358  return;
359  }
360 
361  Eigen::Vector4f plane_parameters;
362 
363  // Use Least-Squares to fit the plane through all the given sample points and find out its coefficients
364  EIGEN_ALIGN16 Eigen::Matrix3f covariance_matrix;
365  Eigen::Vector4f xyz_centroid;
366 
367  if (0 == computeMeanAndCovarianceMatrix (*input_, inliers, covariance_matrix, xyz_centroid))
368  {
369  PCL_ERROR ("[pcl::SampleConsensusModelPlane::optimizeModelCoefficients] computeMeanAndCovarianceMatrix failed (returned 0) because there are no valid inliers.\n");
370  optimized_coefficients = model_coefficients;
371  return;
372  }
373 
374  // Compute the model coefficients
375  EIGEN_ALIGN16 Eigen::Vector3f::Scalar eigen_value;
376  EIGEN_ALIGN16 Eigen::Vector3f eigen_vector;
377  pcl::eigen33 (covariance_matrix, eigen_value, eigen_vector);
378 
379  // Hessian form (D = nc . p_plane (centroid here) + p)
380  optimized_coefficients.resize (model_size_);
381  optimized_coefficients[0] = eigen_vector [0];
382  optimized_coefficients[1] = eigen_vector [1];
383  optimized_coefficients[2] = eigen_vector [2];
384  optimized_coefficients[3] = 0.0f;
385  optimized_coefficients[3] = -1.0f * optimized_coefficients.dot (xyz_centroid);
386 
387  // Make sure it results in a valid model
388  if (!isModelValid (optimized_coefficients))
389  {
390  optimized_coefficients = model_coefficients;
391  }
392 }
393 
394 //////////////////////////////////////////////////////////////////////////
395 template <typename PointT> void
397  const Indices &inliers, const Eigen::VectorXf &model_coefficients, PointCloud &projected_points, bool copy_data_fields) const
398 {
399  // Needs a valid set of model coefficients
400  if (!isModelValid (model_coefficients))
401  {
402  PCL_ERROR ("[pcl::SampleConsensusModelPlane::projectPoints] Given model is invalid!\n");
403  return;
404  }
405 
406  projected_points.header = input_->header;
407  projected_points.is_dense = input_->is_dense;
408 
409  Eigen::Vector4f mc (model_coefficients[0], model_coefficients[1], model_coefficients[2], 0.0f);
410 
411  // normalize the vector perpendicular to the plane...
412  mc.normalize ();
413  // ... and store the resulting normal as a local copy of the model coefficients
414  Eigen::Vector4f tmp_mc = model_coefficients;
415  tmp_mc[0] = mc[0];
416  tmp_mc[1] = mc[1];
417  tmp_mc[2] = mc[2];
418 
419  // Copy all the data fields from the input cloud to the projected one?
420  if (copy_data_fields)
421  {
422  // Allocate enough space and copy the basics
423  projected_points.resize (input_->size ());
424  projected_points.width = input_->width;
425  projected_points.height = input_->height;
426 
427  using FieldList = typename pcl::traits::fieldList<PointT>::type;
428  // Iterate over each point
429  for (std::size_t i = 0; i < input_->size (); ++i)
430  // Iterate over each dimension
431  pcl::for_each_type <FieldList> (NdConcatenateFunctor <PointT, PointT> ((*input_)[i], projected_points[i]));
432 
433  // Iterate through the 3d points and calculate the distances from them to the plane
434  for (const auto &inlier : inliers)
435  {
436  // Calculate the distance from the point to the plane
437  Eigen::Vector4f p ((*input_)[inlier].x,
438  (*input_)[inlier].y,
439  (*input_)[inlier].z,
440  1);
441  // use normalized coefficients to calculate the scalar projection
442  float distance_to_plane = tmp_mc.dot (p);
443 
444  pcl::Vector4fMap pp = projected_points[inlier].getVector4fMap ();
445  pp.matrix () = p - mc * distance_to_plane; // mc[3] = 0, therefore the 3rd coordinate is safe
446  }
447  }
448  else
449  {
450  // Allocate enough space and copy the basics
451  projected_points.resize (inliers.size ());
452  projected_points.width = inliers.size ();
453  projected_points.height = 1;
454 
455  using FieldList = typename pcl::traits::fieldList<PointT>::type;
456  // Iterate over each point
457  for (std::size_t i = 0; i < inliers.size (); ++i)
458  {
459  // Iterate over each dimension
460  pcl::for_each_type <FieldList> (NdConcatenateFunctor <PointT, PointT> ((*input_)[inliers[i]], projected_points[i]));
461  }
462 
463  // Iterate through the 3d points and calculate the distances from them to the plane
464  for (std::size_t i = 0; i < inliers.size (); ++i)
465  {
466  // Calculate the distance from the point to the plane
467  Eigen::Vector4f p ((*input_)[inliers[i]].x,
468  (*input_)[inliers[i]].y,
469  (*input_)[inliers[i]].z,
470  1.0f);
471  // use normalized coefficients to calculate the scalar projection
472  float distance_to_plane = tmp_mc.dot (p);
473 
474  pcl::Vector4fMap pp = projected_points[i].getVector4fMap ();
475  pp.matrix () = p - mc * distance_to_plane; // mc[3] = 0, therefore the 3rd coordinate is safe
476  }
477  }
478 }
479 
480 //////////////////////////////////////////////////////////////////////////
481 template <typename PointT> bool
483  const std::set<index_t> &indices, const Eigen::VectorXf &model_coefficients, const double threshold) const
484 {
485  // Needs a valid set of model coefficients
486  if (!isModelValid (model_coefficients))
487  {
488  PCL_ERROR ("[pcl::SampleConsensusModelPlane::doSamplesVerifyModel] Given model is invalid!\n");
489  return (false);
490  }
491 
492  for (const auto &index : indices)
493  {
494  Eigen::Vector4f pt ((*input_)[index].x,
495  (*input_)[index].y,
496  (*input_)[index].z,
497  1.0f);
498  if (std::abs (model_coefficients.dot (pt)) > threshold)
499  {
500  return (false);
501  }
502  }
503 
504  return (true);
505 }
506 
507 #define PCL_INSTANTIATE_SampleConsensusModelPlane(T) template class PCL_EXPORTS pcl::SampleConsensusModelPlane<T>;
508 
509 #endif // PCL_SAMPLE_CONSENSUS_IMPL_SAC_MODEL_PLANE_H_
510 
Define methods for centroid estimation and covariance matrix calculus.
PointCloud represents the base class in PCL for storing collections of 3D points.
Definition: point_cloud.h:173
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:403
void resize(std::size_t count)
Resizes the container to contain count elements.
Definition: point_cloud.h:462
std::uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:398
pcl::PCLHeader header
The point cloud header.
Definition: point_cloud.h:392
std::uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:400
SampleConsensusModelPlane defines a model for 3D plane segmentation.
void optimizeModelCoefficients(const Indices &inliers, const Eigen::VectorXf &model_coefficients, Eigen::VectorXf &optimized_coefficients) const override
Recompute the plane coefficients using the given inlier set and return them to the user.
void getDistancesToModel(const Eigen::VectorXf &model_coefficients, std::vector< double > &distances) const override
Compute all distances from the cloud data to a given plane model.
bool doSamplesVerifyModel(const std::set< index_t > &indices, const Eigen::VectorXf &model_coefficients, const double threshold) const override
Verify whether a subset of indices verifies the given plane model coefficients.
void selectWithinDistance(const Eigen::VectorXf &model_coefficients, const double threshold, Indices &inliers) override
Select all the points which respect the given model coefficients as inliers.
void projectPoints(const Indices &inliers, const Eigen::VectorXf &model_coefficients, PointCloud &projected_points, bool copy_data_fields=true) const override
Create a new point cloud with inliers projected onto the plane model.
bool computeModelCoefficients(const Indices &samples, Eigen::VectorXf &model_coefficients) const override
Check whether the given index samples can form a valid plane model, compute the model coefficients fr...
std::size_t countWithinDistanceStandard(const Eigen::VectorXf &model_coefficients, const double threshold, std::size_t i=0) const
This implementation uses no SIMD instructions.
std::size_t countWithinDistance(const Eigen::VectorXf &model_coefficients, const double threshold) const override
Count all the points which respect the given model coefficients as inliers.
unsigned int computeMeanAndCovarianceMatrix(const pcl::PointCloud< PointT > &cloud, Eigen::Matrix< Scalar, 3, 3 > &covariance_matrix, Eigen::Matrix< Scalar, 4, 1 > &centroid)
Compute the normalized 3x3 covariance matrix and the centroid of a given set of points in a single lo...
Definition: centroid.hpp:508
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:296
float distance(const PointT &p1, const PointT &p2)
Definition: geometry.h:60
Eigen::Map< Eigen::Vector4f, Eigen::Aligned > Vector4fMap
const Eigen::Map< const Eigen::Vector3f > Vector3fMapConst
IndicesAllocator<> Indices
Type used for indices in PCL.
Definition: types.h:133
Helper functor structure for concatenate.
Definition: concatenate.h:50