Point Cloud Library (PCL) 1.13.0
Loading...
Searching...
No Matches
sac_model_line.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_LINE_H_
42#define PCL_SAMPLE_CONSENSUS_IMPL_SAC_MODEL_LINE_H_
43
44#include <pcl/sample_consensus/sac_model_line.h>
45#include <pcl/common/centroid.h>
46#include <pcl/common/concatenate.h>
47#include <pcl/common/eigen.h> // for eigen33
48
49//////////////////////////////////////////////////////////////////////////
50template <typename PointT> bool
52{
53 if (samples.size () != sample_size_)
54 {
55 PCL_ERROR ("[pcl::SampleConsensusModelLine::isSampleGood] Wrong number of samples (is %lu, should be %lu)!\n", samples.size (), sample_size_);
56 return (false);
57 }
58
59 // Make sure that the two sample points are not identical
60 if (
61 std::abs ((*input_)[samples[0]].x - (*input_)[samples[1]].x) <= std::numeric_limits<float>::epsilon ()
62 &&
63 std::abs ((*input_)[samples[0]].y - (*input_)[samples[1]].y) <= std::numeric_limits<float>::epsilon ()
64 &&
65 std::abs ((*input_)[samples[0]].z - (*input_)[samples[1]].z) <= std::numeric_limits<float>::epsilon ())
66 {
67 PCL_ERROR ("[pcl::SampleConsensusModelLine::isSampleGood] The two sample points are (almost) identical!\n");
68 return (false);
69 }
70
71 return (true);
72}
73
74//////////////////////////////////////////////////////////////////////////
75template <typename PointT> bool
77 const Indices &samples, Eigen::VectorXf &model_coefficients) const
78{
79 // Make sure that the samples are valid
80 if (!isSampleGood (samples))
81 {
82 PCL_ERROR ("[pcl::SampleConsensusModelLine::computeModelCoefficients] Invalid set of samples given!\n");
83 return (false);
84 }
85
86 model_coefficients.resize (model_size_);
87 model_coefficients[0] = (*input_)[samples[0]].x;
88 model_coefficients[1] = (*input_)[samples[0]].y;
89 model_coefficients[2] = (*input_)[samples[0]].z;
90
91 model_coefficients[3] = (*input_)[samples[1]].x - model_coefficients[0];
92 model_coefficients[4] = (*input_)[samples[1]].y - model_coefficients[1];
93 model_coefficients[5] = (*input_)[samples[1]].z - model_coefficients[2];
94
95 // This precondition should hold if the samples have been found to be good
96 assert (model_coefficients.template tail<3> ().squaredNorm () > 0.0f);
97
98 model_coefficients.template tail<3> ().normalize ();
99 PCL_DEBUG ("[pcl::SampleConsensusModelLine::computeModelCoefficients] Model is (%g,%g,%g,%g,%g,%g).\n",
100 model_coefficients[0], model_coefficients[1], model_coefficients[2],
101 model_coefficients[3], model_coefficients[4], model_coefficients[5]);
102 return (true);
103}
104
105//////////////////////////////////////////////////////////////////////////
106template <typename PointT> void
108 const Eigen::VectorXf &model_coefficients, std::vector<double> &distances) const
109{
110 // Needs a valid set of model coefficients
111 if (!isModelValid (model_coefficients))
112 {
113 return;
114 }
115
116 distances.resize (indices_->size ());
117
118 // Obtain the line point and direction
119 Eigen::Vector4f line_pt (model_coefficients[0], model_coefficients[1], model_coefficients[2], 0);
120 Eigen::Vector4f line_dir (model_coefficients[3], model_coefficients[4], model_coefficients[5], 0);
121 line_dir.normalize ();
122
123 // Iterate through the 3d points and calculate the distances from them to the line
124 for (std::size_t i = 0; i < indices_->size (); ++i)
125 {
126 // Calculate the distance from the point to the line
127 // D = ||(P2-P1) x (P1-P0)|| / ||P2-P1|| = norm (cross (p2-p1, p2-p0)) / norm(p2-p1)
128 // Need to estimate sqrt here to keep MSAC and friends general
129 distances[i] = sqrt ((line_pt - (*input_)[(*indices_)[i]].getVector4fMap ()).cross3 (line_dir).squaredNorm ());
130 }
131}
132
133//////////////////////////////////////////////////////////////////////////
134template <typename PointT> void
136 const Eigen::VectorXf &model_coefficients, const double threshold, Indices &inliers)
137{
138 // Needs a valid set of model coefficients
139 if (!isModelValid (model_coefficients))
140 return;
141
142 double sqr_threshold = threshold * threshold;
143
144 inliers.clear ();
145 error_sqr_dists_.clear ();
146 inliers.reserve (indices_->size ());
147 error_sqr_dists_.reserve (indices_->size ());
148
149 // Obtain the line point and direction
150 Eigen::Vector4f line_pt (model_coefficients[0], model_coefficients[1], model_coefficients[2], 0);
151 Eigen::Vector4f line_dir (model_coefficients[3], model_coefficients[4], model_coefficients[5], 0);
152 line_dir.normalize ();
153
154 // Iterate through the 3d points and calculate the distances from them to the line
155 for (std::size_t i = 0; i < indices_->size (); ++i)
156 {
157 // Calculate the distance from the point to the line
158 // D = ||(P2-P1) x (P1-P0)|| / ||P2-P1|| = norm (cross (p2-p1, p2-p0)) / norm(p2-p1)
159 double sqr_distance = (line_pt - (*input_)[(*indices_)[i]].getVector4fMap ()).cross3 (line_dir).squaredNorm ();
160
161 if (sqr_distance < sqr_threshold)
162 {
163 // Returns the indices of the points whose squared distances are smaller than the threshold
164 inliers.push_back ((*indices_)[i]);
165 error_sqr_dists_.push_back (sqr_distance);
166 }
167 }
168}
169
170//////////////////////////////////////////////////////////////////////////
171template <typename PointT> std::size_t
173 const Eigen::VectorXf &model_coefficients, const double threshold) const
174{
175 // Needs a valid set of model coefficients
176 if (!isModelValid (model_coefficients))
177 return (0);
178
179 double sqr_threshold = threshold * threshold;
180
181 std::size_t nr_p = 0;
182
183 // Obtain the line point and direction
184 Eigen::Vector4f line_pt (model_coefficients[0], model_coefficients[1], model_coefficients[2], 0.0f);
185 Eigen::Vector4f line_dir (model_coefficients[3], model_coefficients[4], model_coefficients[5], 0.0f);
186 line_dir.normalize ();
187
188 // Iterate through the 3d points and calculate the distances from them to the line
189 for (std::size_t i = 0; i < indices_->size (); ++i)
190 {
191 // Calculate the distance from the point to the line
192 // D = ||(P2-P1) x (P1-P0)|| / ||P2-P1|| = norm (cross (p2-p1, p2-p0)) / norm(p2-p1)
193 double sqr_distance = (line_pt - (*input_)[(*indices_)[i]].getVector4fMap ()).cross3 (line_dir).squaredNorm ();
194
195 if (sqr_distance < sqr_threshold)
196 nr_p++;
197 }
198 return (nr_p);
199}
200
201//////////////////////////////////////////////////////////////////////////
202template <typename PointT> void
204 const Indices &inliers, const Eigen::VectorXf &model_coefficients, Eigen::VectorXf &optimized_coefficients) const
205{
206 // Needs a valid set of model coefficients
207 if (!isModelValid (model_coefficients))
208 {
209 optimized_coefficients = model_coefficients;
210 return;
211 }
212
213 // Need more than the minimum sample size to make a difference
214 if (inliers.size () <= sample_size_)
215 {
216 PCL_ERROR ("[pcl::SampleConsensusModelLine::optimizeModelCoefficients] Not enough inliers to refine/optimize the model's coefficients (%lu)! Returning the same coefficients.\n", inliers.size ());
217 optimized_coefficients = model_coefficients;
218 return;
219 }
220
221 optimized_coefficients.resize (model_size_);
222
223 // Compute the 3x3 covariance matrix
224 Eigen::Vector4f centroid;
225 if (0 == compute3DCentroid (*input_, inliers, centroid))
226 {
227 PCL_WARN ("[pcl::SampleConsensusModelLine::optimizeModelCoefficients] compute3DCentroid failed (returned 0) because there are no valid inliers.\n");
228 optimized_coefficients = model_coefficients;
229 return;
230 }
231 Eigen::Matrix3f covariance_matrix;
232 computeCovarianceMatrix (*input_, inliers, centroid, covariance_matrix);
233 optimized_coefficients[0] = centroid[0];
234 optimized_coefficients[1] = centroid[1];
235 optimized_coefficients[2] = centroid[2];
236
237 // Extract the eigenvalues and eigenvectors
238 EIGEN_ALIGN16 Eigen::Vector3f eigen_values;
239 EIGEN_ALIGN16 Eigen::Vector3f eigen_vector;
240 pcl::eigen33 (covariance_matrix, eigen_values);
241 pcl::computeCorrespondingEigenVector (covariance_matrix, eigen_values [2], eigen_vector);
242 //pcl::eigen33 (covariance_matrix, eigen_vectors, eigen_values);
243
244 optimized_coefficients.template tail<3> ().matrix () = eigen_vector;
245}
246
247//////////////////////////////////////////////////////////////////////////
248template <typename PointT> void
250 const Indices &inliers, const Eigen::VectorXf &model_coefficients, PointCloud &projected_points, bool copy_data_fields) const
251{
252 // Needs a valid model coefficients
253 if (!isModelValid (model_coefficients))
254 return;
255
256 // Obtain the line point and direction
257 Eigen::Vector4f line_pt (model_coefficients[0], model_coefficients[1], model_coefficients[2], 0.0f);
258 Eigen::Vector4f line_dir (model_coefficients[3], model_coefficients[4], model_coefficients[5], 0.0f);
259
260 projected_points.header = input_->header;
261 projected_points.is_dense = input_->is_dense;
262
263 // Copy all the data fields from the input cloud to the projected one?
264 if (copy_data_fields)
265 {
266 // Allocate enough space and copy the basics
267 projected_points.resize (input_->size ());
268 projected_points.width = input_->width;
269 projected_points.height = input_->height;
270
271 using FieldList = typename pcl::traits::fieldList<PointT>::type;
272 // Iterate over each point
273 for (std::size_t i = 0; i < projected_points.size (); ++i)
274 // Iterate over each dimension
275 pcl::for_each_type <FieldList> (NdConcatenateFunctor <PointT, PointT> ((*input_)[i], projected_points[i]));
276
277 // Iterate through the 3d points and calculate the distances from them to the line
278 for (const auto &inlier : inliers)
279 {
280 Eigen::Vector4f pt ((*input_)[inlier].x, (*input_)[inlier].y, (*input_)[inlier].z, 0.0f);
281 // double k = (DOT_PROD_3D (points[i], p21) - dotA_B) / dotB_B;
282 float k = (pt.dot (line_dir) - line_pt.dot (line_dir)) / line_dir.dot (line_dir);
283
284 Eigen::Vector4f pp = line_pt + k * line_dir;
285 // Calculate the projection of the point on the line (pointProj = A + k * B)
286 projected_points[inlier].x = pp[0];
287 projected_points[inlier].y = pp[1];
288 projected_points[inlier].z = pp[2];
289 }
290 }
291 else
292 {
293 // Allocate enough space and copy the basics
294 projected_points.resize (inliers.size ());
295 projected_points.width = inliers.size ();
296 projected_points.height = 1;
297
298 using FieldList = typename pcl::traits::fieldList<PointT>::type;
299 // Iterate over each point
300 for (std::size_t i = 0; i < inliers.size (); ++i)
301 // Iterate over each dimension
302 pcl::for_each_type <FieldList> (NdConcatenateFunctor <PointT, PointT> ((*input_)[inliers[i]], projected_points[i]));
303
304 // Iterate through the 3d points and calculate the distances from them to the line
305 for (std::size_t i = 0; i < inliers.size (); ++i)
306 {
307 Eigen::Vector4f pt ((*input_)[inliers[i]].x, (*input_)[inliers[i]].y, (*input_)[inliers[i]].z, 0.0f);
308 // double k = (DOT_PROD_3D (points[i], p21) - dotA_B) / dotB_B;
309 float k = (pt.dot (line_dir) - line_pt.dot (line_dir)) / line_dir.dot (line_dir);
310
311 Eigen::Vector4f pp = line_pt + k * line_dir;
312 // Calculate the projection of the point on the line (pointProj = A + k * B)
313 projected_points[i].x = pp[0];
314 projected_points[i].y = pp[1];
315 projected_points[i].z = pp[2];
316 }
317 }
318}
319
320//////////////////////////////////////////////////////////////////////////
321template <typename PointT> bool
323 const std::set<index_t> &indices, const Eigen::VectorXf &model_coefficients, const double threshold) const
324{
325 // Needs a valid set of model coefficients
326 if (!isModelValid (model_coefficients))
327 return (false);
328
329 // Obtain the line point and direction
330 Eigen::Vector4f line_pt (model_coefficients[0], model_coefficients[1], model_coefficients[2], 0.0f);
331 Eigen::Vector4f line_dir (model_coefficients[3], model_coefficients[4], model_coefficients[5], 0.0f);
332 line_dir.normalize ();
333
334 double sqr_threshold = threshold * threshold;
335 // Iterate through the 3d points and calculate the distances from them to the line
336 for (const auto &index : indices)
337 {
338 // Calculate the distance from the point to the line
339 // D = ||(P2-P1) x (P1-P0)|| / ||P2-P1|| = norm (cross (p2-p1, p2-p0)) / norm(p2-p1)
340 if ((line_pt - (*input_)[index].getVector4fMap ()).cross3 (line_dir).squaredNorm () > sqr_threshold)
341 return (false);
342 }
343
344 return (true);
345}
346
347#define PCL_INSTANTIATE_SampleConsensusModelLine(T) template class PCL_EXPORTS pcl::SampleConsensusModelLine<T>;
348
349#endif // PCL_SAMPLE_CONSENSUS_IMPL_SAC_MODEL_LINE_H_
350
Define methods for centroid estimation and covariance matrix calculus.
bool computeModelCoefficients(const Indices &samples, Eigen::VectorXf &model_coefficients) const override
Check whether the given index samples can form a valid line model, compute the model coefficients fro...
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.
void getDistancesToModel(const Eigen::VectorXf &model_coefficients, std::vector< double > &distances) const override
Compute all squared distances from the cloud data to a given line model.
bool isSampleGood(const Indices &samples) const override
Check if a sample of indices results in a good sample of points indices.
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 line model coefficients.
void optimizeModelCoefficients(const Indices &inliers, const Eigen::VectorXf &model_coefficients, Eigen::VectorXf &optimized_coefficients) const override
Recompute the line coefficients using the given inlier set and return them to the user.
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 line model.
typename SampleConsensusModel< PointT >::PointCloud PointCloud
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 computeCorrespondingEigenVector(const Matrix &mat, const typename Matrix::Scalar &eigenvalue, Vector &eigenvector)
determines the corresponding eigenvector to the given eigenvalue of the symmetric positive semi defin...
Definition eigen.hpp:226
unsigned int computeCovarianceMatrix(const pcl::PointCloud< PointT > &cloud, const Eigen::Matrix< Scalar, 4, 1 > &centroid, Eigen::Matrix< Scalar, 3, 3 > &covariance_matrix)
Compute the 3x3 covariance matrix of a given set of points.
Definition centroid.hpp:191
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
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