Point Cloud Library (PCL) 1.13.1
Loading...
Searching...
No Matches
lmeds.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_LMEDS_H_
42#define PCL_SAMPLE_CONSENSUS_IMPL_LMEDS_H_
43
44#include <pcl/sample_consensus/lmeds.h>
45#include <pcl/common/common.h> // for computeMedian
46
47//////////////////////////////////////////////////////////////////////////
48template <typename PointT> bool
50{
51 // Warn and exit if no threshold was set
52 if (threshold_ == std::numeric_limits<double>::max())
53 {
54 PCL_ERROR ("[pcl::LeastMedianSquares::computeModel] No threshold set!\n");
55 return (false);
56 }
57
58 iterations_ = 0;
59 double d_best_penalty = std::numeric_limits<double>::max();
60
61 Indices selection;
62 Eigen::VectorXf model_coefficients (sac_model_->getModelSize ());
63 std::vector<double> distances;
64
65 unsigned skipped_count = 0;
66 // suppress infinite loops by just allowing 10 x maximum allowed iterations for invalid model parameters!
67 const unsigned max_skip = max_iterations_ * 10;
68
69 // Iterate
70 while ((iterations_ < max_iterations_) && (skipped_count < max_skip))
71 {
72 // Get X samples which satisfy the model criteria
73 sac_model_->getSamples (iterations_, selection);
74
75 if (selection.empty ())
76 {
77 PCL_ERROR ("[pcl::LeastMedianSquares::computeModel] No samples could be selected!\n");
78 break;
79 }
80
81 // Search for inliers in the point cloud for the current plane model M
82 if (!sac_model_->computeModelCoefficients (selection, model_coefficients))
83 {
84 //iterations_++;
85 ++skipped_count;
86 PCL_DEBUG ("[pcl::LeastMedianSquares::computeModel] The function computeModelCoefficients failed, so continue with next iteration.\n");
87 continue;
88 }
89
90 double d_cur_penalty;
91 // d_cur_penalty = sum (min (dist, threshold))
92
93 // Iterate through the 3d points and calculate the distances from them to the model
94 sac_model_->getDistancesToModel (model_coefficients, distances);
95
96 // No distances? The model must not respect the user given constraints
97 if (distances.empty ())
98 {
99 //iterations_++;
100 ++skipped_count;
101 continue;
102 }
103 // Move all NaNs in distances to the end
104 const auto new_end = (sac_model_->getInputCloud()->is_dense ? distances.end() : std::partition (distances.begin(), distances.end(), [](double d){return !std::isnan (d);}));
105 const auto nr_valid_dists = std::distance (distances.begin (), new_end);
106
107 // d_cur_penalty = median (distances)
108 PCL_DEBUG ("[pcl::LeastMedianSquares::computeModel] There are %lu valid distances remaining after removing NaN values.\n", nr_valid_dists);
109 if (nr_valid_dists == 0)
110 {
111 //iterations_++;
112 ++skipped_count;
113 continue;
114 }
115 d_cur_penalty = pcl::computeMedian (distances.begin (), new_end, static_cast<double(*)(double)>(std::sqrt));
116
117 // Better match ?
118 if (d_cur_penalty < d_best_penalty)
119 {
120 d_best_penalty = d_cur_penalty;
121
122 // Save the current model/coefficients selection as being the best so far
123 model_ = selection;
124 model_coefficients_ = model_coefficients;
125 }
126
127 ++iterations_;
128 if (debug_verbosity_level > 1)
129 {
130 PCL_DEBUG ("[pcl::LeastMedianSquares::computeModel] Trial %d out of %d. Best penalty is %f.\n", iterations_, max_iterations_, d_best_penalty);
131 }
132 }
133
134 if (model_.empty ())
135 {
136 if (debug_verbosity_level > 0)
137 {
138 PCL_DEBUG ("[pcl::LeastMedianSquares::computeModel] Unable to find a solution!\n");
139 }
140 return (false);
141 }
142
143 // Classify the data points into inliers and outliers
144 // Sigma = 1.4826 * (1 + 5 / (n-d)) * sqrt (M)
145 // @note: See "Robust Regression Methods for Computer Vision: A Review"
146 //double sigma = 1.4826 * (1 + 5 / (sac_model_->getIndices ()->size () - best_model.size ())) * sqrt (d_best_penalty);
147 //double threshold = 2.5 * sigma;
148
149 // Iterate through the 3d points and calculate the distances from them to the model again
150 sac_model_->getDistancesToModel (model_coefficients_, distances);
151 // No distances? The model must not respect the user given constraints
152 if (distances.empty ())
153 {
154 PCL_ERROR ("[pcl::LeastMedianSquares::computeModel] The model found failed to verify against the given constraints!\n");
155 return (false);
156 }
157
158 Indices &indices = *sac_model_->getIndices ();
159
160 if (distances.size () != indices.size ())
161 {
162 PCL_ERROR ("[pcl::LeastMedianSquares::computeModel] Estimated distances (%lu) differs than the normal of indices (%lu).\n", distances.size (), indices.size ());
163 return (false);
164 }
165
166 inliers_.resize (distances.size ());
167 // Get the inliers for the best model found
168 std::size_t n_inliers_count = 0;
169 for (std::size_t i = 0; i < distances.size (); ++i)
170 {
171 if (distances[i] <= threshold_)
172 {
173 inliers_[n_inliers_count++] = indices[i];
174 }
175 }
176
177 // Resize the inliers vector
178 inliers_.resize (n_inliers_count);
179
180 if (debug_verbosity_level > 0)
181 {
182 PCL_DEBUG ("[pcl::LeastMedianSquares::computeModel] Model: %lu size, %lu inliers.\n", model_.size (), n_inliers_count);
183 }
184
185 return (true);
186}
187
188#define PCL_INSTANTIATE_LeastMedianSquares(T) template class PCL_EXPORTS pcl::LeastMedianSquares<T>;
189
190#endif // PCL_SAMPLE_CONSENSUS_IMPL_LMEDS_H_
bool computeModel(int debug_verbosity_level=0) override
Compute the actual model and find the inliers.
Definition lmeds.hpp:49
Define standard C methods and C++ classes that are common to all methods.
auto computeMedian(IteratorT begin, IteratorT end, Functor f) noexcept -> std::result_of_t< Functor(decltype(*begin))>
Compute the median of a list of values (fast).
Definition common.h:285
IndicesAllocator<> Indices
Type used for indices in PCL.
Definition types.h:133