Point Cloud Library (PCL) 1.13.1
Loading...
Searching...
No Matches
prosac.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_PROSAC_H_
42#define PCL_SAMPLE_CONSENSUS_IMPL_PROSAC_H_
43
44#if defined __GNUC__
45# pragma GCC system_header
46#endif
47
48#include <limits>
49
50#include <boost/math/distributions/binomial.hpp>
51#include <pcl/sample_consensus/prosac.h>
52
53//////////////////////////////////////////////////////////////////////////
54// Variable naming uses capital letters to make the comparison with the original paper easier
55template<typename PointT> bool
57{
58 // Warn and exit if no threshold was set
59 if (threshold_ == std::numeric_limits<double>::max())
60 {
61 PCL_ERROR ("[pcl::ProgressiveSampleConsensus::computeModel] No threshold set!\n");
62 return (false);
63 }
64
65 // Initialize some PROSAC constants
66 const int T_N = 200000;
67 const std::size_t N = sac_model_->indices_->size ();
68 const std::size_t m = sac_model_->getSampleSize ();
69 float T_n = static_cast<float> (T_N);
70 for (unsigned int i = 0; i < m; ++i)
71 T_n *= static_cast<float> (m - i) / static_cast<float> (N - i);
72 float T_prime_n = 1.0f;
73 std::size_t I_N_best = 0;
74 float n = static_cast<float> (m);
75
76 // Define the n_Start coefficients from Section 2.2
77 float n_star = static_cast<float> (N);
78 float epsilon_n_star = 0.0;
79 std::size_t k_n_star = T_N;
80
81 // Compute the I_n_star_min of Equation 8
82 std::vector<unsigned int> I_n_star_min (N);
83
84 // Initialize the usual RANSAC parameters
85 iterations_ = 0;
86
87 Indices inliers;
88 Indices selection;
89 Eigen::VectorXf model_coefficients (sac_model_->getModelSize ());
90
91 // We will increase the pool so the indices_ vector can only contain m elements at first
92 Indices index_pool;
93 index_pool.reserve (N);
94 for (unsigned int i = 0; i < n; ++i)
95 index_pool.push_back (sac_model_->indices_->operator[](i));
96
97 // Iterate
98 while (static_cast<unsigned int> (iterations_) < k_n_star)
99 {
100 // Choose the samples
101
102 // Step 1
103 // According to Equation 5 in the text text, not the algorithm
104 if ((iterations_ == T_prime_n) && (n < n_star))
105 {
106 // Increase the pool
107 ++n;
108 if (n >= N)
109 break;
110 index_pool.push_back (sac_model_->indices_->at(static_cast<unsigned int> (n - 1)));
111 // Update other variables
112 float T_n_minus_1 = T_n;
113 T_n *= (static_cast<float>(n) + 1.0f) / (static_cast<float>(n) + 1.0f - static_cast<float>(m));
114 T_prime_n += std::ceil (T_n - T_n_minus_1);
115 }
116
117 // Step 2
118 sac_model_->indices_->swap (index_pool);
119 selection.clear ();
120 sac_model_->getSamples (iterations_, selection);
121 if (T_prime_n < iterations_)
122 {
123 selection.pop_back ();
124 selection.push_back (sac_model_->indices_->at(static_cast<unsigned int> (n - 1)));
125 }
126
127 // Make sure we use the right indices for testing
128 sac_model_->indices_->swap (index_pool);
129
130 if (selection.empty ())
131 {
132 PCL_ERROR ("[pcl::ProgressiveSampleConsensus::computeModel] No samples could be selected!\n");
133 break;
134 }
135
136 // Search for inliers in the point cloud for the current model
137 if (!sac_model_->computeModelCoefficients (selection, model_coefficients))
138 {
139 ++iterations_;
140 continue;
141 }
142
143 // Select the inliers that are within threshold_ from the model
144 inliers.clear ();
145 sac_model_->selectWithinDistance (model_coefficients, threshold_, inliers);
146
147 std::size_t I_N = inliers.size ();
148
149 // If we find more inliers than before
150 if (I_N > I_N_best)
151 {
152 I_N_best = I_N;
153
154 // Save the current model/inlier/coefficients selection as being the best so far
155 inliers_ = inliers;
156 model_ = selection;
157 model_coefficients_ = model_coefficients;
158
159 // We estimate I_n_star for different possible values of n_star by using the inliers
160 std::sort (inliers.begin (), inliers.end ());
161
162 // Try to find a better n_star
163 // We minimize k_n_star and therefore maximize epsilon_n_star = I_n_star / n_star
164 std::size_t possible_n_star_best = N, I_possible_n_star_best = I_N;
165 float epsilon_possible_n_star_best = static_cast<float>(I_possible_n_star_best) / static_cast<float>(possible_n_star_best);
166
167 // We only need to compute possible better epsilon_n_star for when _n is just about to be removed an inlier
168 std::size_t I_possible_n_star = I_N;
169 for (auto last_inlier = inliers.crbegin (), inliers_end = inliers.crend ();
170 last_inlier != inliers_end;
171 ++last_inlier, --I_possible_n_star)
172 {
173 // The best possible_n_star for a given I_possible_n_star is the index of the last inlier
174 unsigned int possible_n_star = (*last_inlier) + 1;
175 if (possible_n_star <= m)
176 break;
177
178 // If we find a better epsilon_n_star
179 float epsilon_possible_n_star = static_cast<float>(I_possible_n_star) / static_cast<float>(possible_n_star);
180 // Make sure we have a better epsilon_possible_n_star
181 if ((epsilon_possible_n_star > epsilon_n_star) && (epsilon_possible_n_star > epsilon_possible_n_star_best))
182 {
183 // Typo in Equation 7, not (n-m choose i-m) but (n choose i-m)
184 std::size_t I_possible_n_star_min = m
185 + static_cast<std::size_t> (std::ceil (boost::math::quantile (boost::math::complement (boost::math::binomial_distribution<float>(static_cast<float> (possible_n_star), 0.1f), 0.05))));
186 // If Equation 9 is not verified, exit
187 if (I_possible_n_star < I_possible_n_star_min)
188 break;
189
190 possible_n_star_best = possible_n_star;
191 I_possible_n_star_best = I_possible_n_star;
192 epsilon_possible_n_star_best = epsilon_possible_n_star;
193 }
194 }
195
196 // Check if we get a better epsilon
197 if (epsilon_possible_n_star_best > epsilon_n_star)
198 {
199 // update the best value
200 epsilon_n_star = epsilon_possible_n_star_best;
201
202 // Compute the new k_n_star
203 float bottom_log = 1 - std::pow (epsilon_n_star, static_cast<float>(m));
204 if (bottom_log == 0)
205 k_n_star = 1;
206 else if (bottom_log == 1)
207 k_n_star = T_N;
208 else
209 k_n_star = static_cast<int> (std::ceil (std::log (0.05) / std::log (bottom_log)));
210 // It seems weird to have very few iterations, so do have a few (totally empirical)
211 k_n_star = (std::max)(k_n_star, 2 * m);
212 }
213 }
214
215 ++iterations_;
216 if (debug_verbosity_level > 1)
217 PCL_DEBUG ("[pcl::ProgressiveSampleConsensus::computeModel] Trial %d out of %d: %d inliers (best is: %d so far).\n", iterations_, k_n_star, I_N, I_N_best);
218 if (iterations_ > max_iterations_)
219 {
220 if (debug_verbosity_level > 0)
221 PCL_DEBUG ("[pcl::ProgressiveSampleConsensus::computeModel] RANSAC reached the maximum number of trials.\n");
222 break;
223 }
224 }
225
226 if (debug_verbosity_level > 0)
227 PCL_DEBUG ("[pcl::ProgressiveSampleConsensus::computeModel] Model: %lu size, %d inliers.\n", model_.size (), I_N_best);
228
229 if (model_.empty ())
230 {
231 inliers_.clear ();
232 return (false);
233 }
234
235 return (true);
236}
237
238#define PCL_INSTANTIATE_ProgressiveSampleConsensus(T) template class PCL_EXPORTS pcl::ProgressiveSampleConsensus<T>;
239
240#endif // PCL_SAMPLE_CONSENSUS_IMPL_PROSAC_H_
bool computeModel(int debug_verbosity_level=0) override
Compute the actual model and find the inliers.
Definition prosac.hpp:56
IndicesAllocator<> Indices
Type used for indices in PCL.
Definition types.h:133