Point Cloud Library (PCL) 1.15.0
Loading...
Searching...
No Matches
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//////////////////////////////////////////////////////////////////////////
50template <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//////////////////////////////////////////////////////////////////////////
80template <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//////////////////////////////////////////////////////////////////////////
121template <typename PointT> void
123 const Eigen::VectorXf &model_coefficients, std::vector<double> &distances) const
124{
125 // Needs a valid set of model coefficients
126 if (!isModelValid (model_coefficients))
127 {
128 PCL_ERROR ("[pcl::SampleConsensusModelPlane::getDistancesToModel] Given model is invalid!\n");
129 return;
130 }
131
132 distances.resize (indices_->size ());
133
134 // Iterate through the 3d points and calculate the distances from them to the plane
135 for (std::size_t i = 0; i < indices_->size (); ++i)
136 {
137 // Calculate the distance from the point to the plane normal as the dot product
138 // D = (P-A).N/|N|
139 /*distances[i] = std::abs (model_coefficients[0] * (*input_)[(*indices_)[i]].x +
140 model_coefficients[1] * (*input_)[(*indices_)[i]].y +
141 model_coefficients[2] * (*input_)[(*indices_)[i]].z +
142 model_coefficients[3]);*/
143 Eigen::Vector4f pt ((*input_)[(*indices_)[i]].x,
144 (*input_)[(*indices_)[i]].y,
145 (*input_)[(*indices_)[i]].z,
146 1.0f);
147 distances[i] = std::abs (model_coefficients.dot (pt));
148 }
149}
150
151//////////////////////////////////////////////////////////////////////////
152template <typename PointT> void
154 const Eigen::VectorXf &model_coefficients, const double threshold, Indices &inliers)
155{
156 // Needs a valid set of model coefficients
157 if (!isModelValid (model_coefficients))
158 {
159 PCL_ERROR ("[pcl::SampleConsensusModelPlane::selectWithinDistance] Given model is invalid!\n");
160 return;
161 }
162
163 inliers.clear ();
164 error_sqr_dists_.clear ();
165 inliers.reserve (indices_->size ());
166 error_sqr_dists_.reserve (indices_->size ());
167
168 // Iterate through the 3d points and calculate the distances from them to the plane
169 for (std::size_t i = 0; i < indices_->size (); ++i)
170 {
171 // Calculate the distance from the point to the plane normal as the dot product
172 // D = (P-A).N/|N|
173 Eigen::Vector4f pt ((*input_)[(*indices_)[i]].x,
174 (*input_)[(*indices_)[i]].y,
175 (*input_)[(*indices_)[i]].z,
176 1.0f);
177
178 float distance = std::abs (model_coefficients.dot (pt));
179
180 if (distance < threshold)
181 {
182 // Returns the indices of the points whose distances are smaller than the threshold
183 inliers.push_back ((*indices_)[i]);
184 error_sqr_dists_.push_back (static_cast<double> (distance));
185 }
186 }
187}
188
189//////////////////////////////////////////////////////////////////////////
190template <typename PointT> std::size_t
192 const Eigen::VectorXf &model_coefficients, const double threshold) const
193{
194 // Needs a valid set of model coefficients
195 if (!isModelValid (model_coefficients))
196 {
197 PCL_ERROR ("[pcl::SampleConsensusModelPlane::countWithinDistance] Given model is invalid!\n");
198 return (0);
199 }
200#if defined (__AVX__) && defined (__AVX2__)
201 return countWithinDistanceAVX (model_coefficients, threshold);
202#elif defined (__SSE__) && defined (__SSE2__) && defined (__SSE4_1__)
203 return countWithinDistanceSSE (model_coefficients, threshold);
204#else
205 return countWithinDistanceStandard (model_coefficients, threshold);
206#endif
207}
208
209//////////////////////////////////////////////////////////////////////////
210template <typename PointT> std::size_t
212 const Eigen::VectorXf &model_coefficients, const double threshold, std::size_t i) const
213{
214 std::size_t nr_p = 0;
215 // Iterate through the 3d points and calculate the distances from them to the plane
216 for (; i < indices_->size (); ++i)
217 {
218 // Calculate the distance from the point to the plane normal as the dot product
219 // D = (P-A).N/|N|
220 Eigen::Vector4f pt ((*input_)[(*indices_)[i]].x,
221 (*input_)[(*indices_)[i]].y,
222 (*input_)[(*indices_)[i]].z,
223 1.0f);
224 if (std::abs (model_coefficients.dot (pt)) < threshold)
225 {
226 nr_p++;
227 }
228 }
229 return (nr_p);
230}
231
232//////////////////////////////////////////////////////////////////////////
233#if defined (__SSE__) && defined (__SSE2__) && defined (__SSE4_1__)
234template <typename PointT> std::size_t
236 const Eigen::VectorXf &model_coefficients, const double threshold, std::size_t i) const
237{
238 std::size_t nr_p = 0;
239 const __m128 a_vec = _mm_set1_ps (model_coefficients[0]);
240 const __m128 b_vec = _mm_set1_ps (model_coefficients[1]);
241 const __m128 c_vec = _mm_set1_ps (model_coefficients[2]);
242 const __m128 d_vec = _mm_set1_ps (model_coefficients[3]);
243 const __m128 threshold_vec = _mm_set1_ps (threshold);
244 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
245 __m128i res = _mm_set1_epi32(0); // This corresponds to nr_p: 4 32bit integers that, summed together, hold the number of inliers
246 for (; (i + 4) <= indices_->size (); i += 4)
247 {
248 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
249 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
250 //const int res = _mm_movemask_ps (mask);
251 //if (res & 1) nr_p++;
252 //if (res & 2) nr_p++;
253 //if (res & 4) nr_p++;
254 //if (res & 8) nr_p++;
255 }
256 nr_p += _mm_extract_epi32 (res, 0);
257 nr_p += _mm_extract_epi32 (res, 1);
258 nr_p += _mm_extract_epi32 (res, 2);
259 nr_p += _mm_extract_epi32 (res, 3);
260
261 // Process the remaining points (at most 3)
262 nr_p += countWithinDistanceStandard(model_coefficients, threshold, i);
263 return (nr_p);
264}
265#endif
266
267//////////////////////////////////////////////////////////////////////////
268#if defined (__AVX__) && defined (__AVX2__)
269template <typename PointT> std::size_t
271 const Eigen::VectorXf &model_coefficients, const double threshold, std::size_t i) const
272{
273 std::size_t nr_p = 0;
274 const __m256 a_vec = _mm256_set1_ps (model_coefficients[0]);
275 const __m256 b_vec = _mm256_set1_ps (model_coefficients[1]);
276 const __m256 c_vec = _mm256_set1_ps (model_coefficients[2]);
277 const __m256 d_vec = _mm256_set1_ps (model_coefficients[3]);
278 const __m256 threshold_vec = _mm256_set1_ps (threshold);
279 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
280 __m256i res = _mm256_set1_epi32(0); // This corresponds to nr_p: 8 32bit integers that, summed together, hold the number of inliers
281 for (; (i + 8) <= indices_->size (); i += 8)
282 {
283 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
284 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
285 //const int res = _mm256_movemask_ps (mask);
286 //if (res & 1) nr_p++;
287 //if (res & 2) nr_p++;
288 //if (res & 4) nr_p++;
289 //if (res & 8) nr_p++;
290 //if (res & 16) nr_p++;
291 //if (res & 32) nr_p++;
292 //if (res & 64) nr_p++;
293 //if (res & 128) nr_p++;
294 }
295 nr_p += _mm256_extract_epi32 (res, 0);
296 nr_p += _mm256_extract_epi32 (res, 1);
297 nr_p += _mm256_extract_epi32 (res, 2);
298 nr_p += _mm256_extract_epi32 (res, 3);
299 nr_p += _mm256_extract_epi32 (res, 4);
300 nr_p += _mm256_extract_epi32 (res, 5);
301 nr_p += _mm256_extract_epi32 (res, 6);
302 nr_p += _mm256_extract_epi32 (res, 7);
303
304 // Process the remaining points (at most 7)
305 nr_p += countWithinDistanceStandard(model_coefficients, threshold, i);
306 return (nr_p);
307}
308#endif
309
310//////////////////////////////////////////////////////////////////////////
311template <typename PointT> void
313 const Indices &inliers, const Eigen::VectorXf &model_coefficients, Eigen::VectorXf &optimized_coefficients) const
314{
315 // Needs a valid set of model coefficients
316 if (!isModelValid (model_coefficients))
317 {
318 PCL_ERROR ("[pcl::SampleConsensusModelPlane::optimizeModelCoefficients] Given model is invalid!\n");
319 optimized_coefficients = model_coefficients;
320 return;
321 }
322
323 // Need more than the minimum sample size to make a difference
324 if (inliers.size () <= sample_size_)
325 {
326 PCL_ERROR ("[pcl::SampleConsensusModelPlane::optimizeModelCoefficients] Not enough inliers found to optimize model coefficients (%lu)! Returning the same coefficients.\n", inliers.size ());
327 optimized_coefficients = model_coefficients;
328 return;
329 }
330
331 Eigen::Vector4f plane_parameters;
332
333 // Use Least-Squares to fit the plane through all the given sample points and find out its coefficients
334 EIGEN_ALIGN16 Eigen::Matrix3f covariance_matrix;
335 Eigen::Vector4f xyz_centroid;
336
337 if (0 == computeMeanAndCovarianceMatrix (*input_, inliers, covariance_matrix, xyz_centroid))
338 {
339 PCL_ERROR ("[pcl::SampleConsensusModelPlane::optimizeModelCoefficients] computeMeanAndCovarianceMatrix failed (returned 0) because there are no valid inliers.\n");
340 optimized_coefficients = model_coefficients;
341 return;
342 }
343
344 // Compute the model coefficients
345 EIGEN_ALIGN16 Eigen::Vector3f::Scalar eigen_value;
346 EIGEN_ALIGN16 Eigen::Vector3f eigen_vector;
347 pcl::eigen33 (covariance_matrix, eigen_value, eigen_vector);
348
349 // Hessian form (D = nc . p_plane (centroid here) + p)
350 optimized_coefficients.resize (model_size_);
351 optimized_coefficients[0] = eigen_vector [0];
352 optimized_coefficients[1] = eigen_vector [1];
353 optimized_coefficients[2] = eigen_vector [2];
354 optimized_coefficients[3] = 0.0f;
355 optimized_coefficients[3] = -1.0f * optimized_coefficients.dot (xyz_centroid);
356
357 // Make sure it results in a valid model
358 if (!isModelValid (optimized_coefficients))
359 {
360 optimized_coefficients = model_coefficients;
361 }
362}
363
364//////////////////////////////////////////////////////////////////////////
365template <typename PointT> void
367 const Indices &inliers, const Eigen::VectorXf &model_coefficients, PointCloud &projected_points, bool copy_data_fields) const
368{
369 // Needs a valid set of model coefficients
370 if (!isModelValid (model_coefficients))
371 {
372 PCL_ERROR ("[pcl::SampleConsensusModelPlane::projectPoints] Given model is invalid!\n");
373 return;
374 }
375
376 projected_points.header = input_->header;
377 projected_points.is_dense = input_->is_dense;
378
379 Eigen::Vector4f mc (model_coefficients[0], model_coefficients[1], model_coefficients[2], 0.0f);
380
381 // normalize the vector perpendicular to the plane...
382 mc.normalize ();
383 // ... and store the resulting normal as a local copy of the model coefficients
384 Eigen::Vector4f tmp_mc = model_coefficients;
385 tmp_mc[0] = mc[0];
386 tmp_mc[1] = mc[1];
387 tmp_mc[2] = mc[2];
388
389 // Copy all the data fields from the input cloud to the projected one?
390 if (copy_data_fields)
391 {
392 // Allocate enough space and copy the basics
393 projected_points.resize (input_->size ());
394 projected_points.width = input_->width;
395 projected_points.height = input_->height;
396
397 using FieldList = typename pcl::traits::fieldList<PointT>::type;
398 // Iterate over each point
399 for (std::size_t i = 0; i < input_->size (); ++i)
400 // Iterate over each dimension
401 pcl::for_each_type <FieldList> (NdConcatenateFunctor <PointT, PointT> ((*input_)[i], projected_points[i]));
402
403 // Iterate through the 3d points and calculate the distances from them to the plane
404 for (const auto &inlier : inliers)
405 {
406 // Calculate the distance from the point to the plane
407 Eigen::Vector4f p ((*input_)[inlier].x,
408 (*input_)[inlier].y,
409 (*input_)[inlier].z,
410 1);
411 // use normalized coefficients to calculate the scalar projection
412 float distance_to_plane = tmp_mc.dot (p);
413
414 pcl::Vector4fMap pp = projected_points[inlier].getVector4fMap ();
415 pp.matrix () = p - mc * distance_to_plane; // mc[3] = 0, therefore the 3rd coordinate is safe
416 }
417 }
418 else
419 {
420 // Allocate enough space and copy the basics
421 projected_points.resize (inliers.size ());
422 projected_points.width = inliers.size ();
423 projected_points.height = 1;
424
425 using FieldList = typename pcl::traits::fieldList<PointT>::type;
426 // Iterate over each point
427 for (std::size_t i = 0; i < inliers.size (); ++i)
428 {
429 // Iterate over each dimension
430 pcl::for_each_type <FieldList> (NdConcatenateFunctor <PointT, PointT> ((*input_)[inliers[i]], projected_points[i]));
431 }
432
433 // Iterate through the 3d points and calculate the distances from them to the plane
434 for (std::size_t i = 0; i < inliers.size (); ++i)
435 {
436 // Calculate the distance from the point to the plane
437 Eigen::Vector4f p ((*input_)[inliers[i]].x,
438 (*input_)[inliers[i]].y,
439 (*input_)[inliers[i]].z,
440 1.0f);
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[i].getVector4fMap ();
445 pp.matrix () = p - mc * distance_to_plane; // mc[3] = 0, therefore the 3rd coordinate is safe
446 }
447 }
448}
449
450//////////////////////////////////////////////////////////////////////////
451template <typename PointT> bool
453 const std::set<index_t> &indices, const Eigen::VectorXf &model_coefficients, const double threshold) const
454{
455 // Needs a valid set of model coefficients
456 if (!isModelValid (model_coefficients))
457 {
458 PCL_ERROR ("[pcl::SampleConsensusModelPlane::doSamplesVerifyModel] Given model is invalid!\n");
459 return (false);
460 }
461
462 for (const auto &index : indices)
463 {
464 Eigen::Vector4f pt ((*input_)[index].x,
465 (*input_)[index].y,
466 (*input_)[index].z,
467 1.0f);
468 if (std::abs (model_coefficients.dot (pt)) > threshold)
469 {
470 return (false);
471 }
472 }
473
474 return (true);
475}
476
477#define PCL_INSTANTIATE_SampleConsensusModelPlane(T) template class PCL_EXPORTS pcl::SampleConsensusModelPlane<T>;
478
479#endif // PCL_SAMPLE_CONSENSUS_IMPL_SAC_MODEL_PLANE_H_
480
Define methods for centroid estimation and covariance matrix calculus.
SampleConsensusModelPlane defines a model for 3D plane segmentation.
typename SampleConsensusModel< PointT >::PointCloud PointCloud
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:509
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
__device__ __host__ __forceinline__ float3 cross(const float3 &v1, const float3 &v2)
Definition utils.hpp:107
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