Point Cloud Library (PCL)
1.7.0
|
00001 /* 00002 * Software License Agreement (BSD License) 00003 * 00004 * Point Cloud Library (PCL) - www.pointclouds.org 00005 * Copyright (c) 2009, Willow Garage, Inc. 00006 * Copyright (c) 2012-, Open Perception, Inc. 00007 * 00008 * All rights reserved. 00009 * 00010 * Redistribution and use in source and binary forms, with or without 00011 * modification, are permitted provided that the following conditions 00012 * are met: 00013 * 00014 * * Redistributions of source code must retain the above copyright 00015 * notice, this list of conditions and the following disclaimer. 00016 * * Redistributions in binary form must reproduce the above 00017 * copyright notice, this list of conditions and the following 00018 * disclaimer in the documentation and/or other materials provided 00019 * with the distribution. 00020 * * Neither the name of the copyright holder(s) nor the names of its 00021 * contributors may be used to endorse or promote products derived 00022 * from this software without specific prior written permission. 00023 * 00024 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 00025 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 00026 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS 00027 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 00028 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 00029 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, 00030 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 00031 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER 00032 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 00033 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN 00034 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 00035 * POSSIBILITY OF SUCH DAMAGE. 00036 * 00037 * $Id$ 00038 * 00039 */ 00040 00041 #ifndef PCL_FEATURES_IMPL_RSD_H_ 00042 #define PCL_FEATURES_IMPL_RSD_H_ 00043 00044 #include <cfloat> 00045 #include <pcl/features/rsd.h> 00046 00047 ////////////////////////////////////////////////////////////////////////////////////////////// 00048 template <typename PointInT, typename PointNT, typename PointOutT> Eigen::MatrixXf 00049 pcl::computeRSD (boost::shared_ptr<const pcl::PointCloud<PointInT> > &surface, boost::shared_ptr<const pcl::PointCloud<PointNT> > &normals, 00050 const std::vector<int> &indices, double max_dist, 00051 int nr_subdiv, double plane_radius, PointOutT &radii, bool compute_histogram) 00052 { 00053 // Check if the full histogram has to be saved or not 00054 Eigen::MatrixXf histogram; 00055 if (compute_histogram) 00056 histogram = Eigen::MatrixXf::Zero (nr_subdiv, nr_subdiv); 00057 00058 // Check if enough points are provided or not 00059 if (indices.size () < 2) 00060 { 00061 radii.r_max = 0; 00062 radii.r_min = 0; 00063 return histogram; 00064 } 00065 00066 // Initialize minimum and maximum angle values in each distance bin 00067 std::vector<std::vector<double> > min_max_angle_by_dist (nr_subdiv); 00068 min_max_angle_by_dist[0].resize (2); 00069 min_max_angle_by_dist[0][0] = min_max_angle_by_dist[0][1] = 0.0; 00070 for (int di=1; di<nr_subdiv; di++) 00071 { 00072 min_max_angle_by_dist[di].resize (2); 00073 min_max_angle_by_dist[di][0] = +DBL_MAX; 00074 min_max_angle_by_dist[di][1] = -DBL_MAX; 00075 } 00076 00077 // Compute distance by normal angle distribution for points 00078 std::vector<int>::const_iterator i, begin (indices.begin()), end (indices.end()); 00079 for(i = begin+1; i != end; ++i) 00080 { 00081 // compute angle between the two lines going through normals (disregard orientation!) 00082 double cosine = normals->points[*i].normal[0] * normals->points[*begin].normal[0] + 00083 normals->points[*i].normal[1] * normals->points[*begin].normal[1] + 00084 normals->points[*i].normal[2] * normals->points[*begin].normal[2]; 00085 if (cosine > 1) cosine = 1; 00086 if (cosine < -1) cosine = -1; 00087 double angle = acos (cosine); 00088 if (angle > M_PI/2) angle = M_PI - angle; /// \note: orientation is neglected! 00089 00090 // Compute point to point distance 00091 double dist = sqrt ((surface->points[*i].x - surface->points[*begin].x) * (surface->points[*i].x - surface->points[*begin].x) + 00092 (surface->points[*i].y - surface->points[*begin].y) * (surface->points[*i].y - surface->points[*begin].y) + 00093 (surface->points[*i].z - surface->points[*begin].z) * (surface->points[*i].z - surface->points[*begin].z)); 00094 00095 if (dist > max_dist) 00096 continue; /// \note: we neglect points that are outside the specified interval! 00097 00098 // compute bins and increase 00099 int bin_d = static_cast<int> (floor (nr_subdiv * dist / max_dist)); 00100 if (compute_histogram) 00101 { 00102 int bin_a = std::min (nr_subdiv-1, static_cast<int> (floor (nr_subdiv * angle / (M_PI/2)))); 00103 histogram(bin_a, bin_d)++; 00104 } 00105 00106 // update min-max values for distance bins 00107 if (min_max_angle_by_dist[bin_d][0] > angle) min_max_angle_by_dist[bin_d][0] = angle; 00108 if (min_max_angle_by_dist[bin_d][1] < angle) min_max_angle_by_dist[bin_d][1] = angle; 00109 } 00110 00111 // Estimate radius from min and max lines 00112 double Amint_Amin = 0, Amint_d = 0; 00113 double Amaxt_Amax = 0, Amaxt_d = 0; 00114 for (int di=0; di<nr_subdiv; di++) 00115 { 00116 // combute the members of A'*A*r = A'*D 00117 if (min_max_angle_by_dist[di][1] >= 0) 00118 { 00119 double p_min = min_max_angle_by_dist[di][0]; 00120 double p_max = min_max_angle_by_dist[di][1]; 00121 double f = (di+0.5)*max_dist/nr_subdiv; 00122 Amint_Amin += p_min * p_min; 00123 Amint_d += p_min * f; 00124 Amaxt_Amax += p_max * p_max; 00125 Amaxt_d += p_max * f; 00126 } 00127 } 00128 float min_radius = Amint_Amin == 0.0f ? float (plane_radius) : float (std::min (Amint_d/Amint_Amin, plane_radius)); 00129 float max_radius = Amaxt_Amax == 0.0f ? float (plane_radius) : float (std::min (Amaxt_d/Amaxt_Amax, plane_radius)); 00130 00131 // Small correction of the systematic error of the estimation (based on analysis with nr_subdiv_ = 5) 00132 min_radius *= 1.1f; 00133 max_radius *= 0.9f; 00134 if (min_radius < max_radius) 00135 { 00136 radii.r_min = min_radius; 00137 radii.r_max = max_radius; 00138 } 00139 else 00140 { 00141 radii.r_max = min_radius; 00142 radii.r_min = max_radius; 00143 } 00144 00145 return histogram; 00146 } 00147 00148 ////////////////////////////////////////////////////////////////////////////////////////////// 00149 template <typename PointNT, typename PointOutT> Eigen::MatrixXf 00150 pcl::computeRSD (boost::shared_ptr<const pcl::PointCloud<PointNT> > &normals, 00151 const std::vector<int> &indices, const std::vector<float> &sqr_dists, double max_dist, 00152 int nr_subdiv, double plane_radius, PointOutT &radii, bool compute_histogram) 00153 { 00154 // Check if the full histogram has to be saved or not 00155 Eigen::MatrixXf histogram; 00156 if (compute_histogram) 00157 histogram = Eigen::MatrixXf::Zero (nr_subdiv, nr_subdiv); 00158 00159 // Check if enough points are provided or not 00160 if (indices.size () < 2) 00161 { 00162 radii.r_max = 0; 00163 radii.r_min = 0; 00164 return histogram; 00165 } 00166 00167 // Initialize minimum and maximum angle values in each distance bin 00168 std::vector<std::vector<double> > min_max_angle_by_dist (nr_subdiv); 00169 min_max_angle_by_dist[0].resize (2); 00170 min_max_angle_by_dist[0][0] = min_max_angle_by_dist[0][1] = 0.0; 00171 for (int di=1; di<nr_subdiv; di++) 00172 { 00173 min_max_angle_by_dist[di].resize (2); 00174 min_max_angle_by_dist[di][0] = +DBL_MAX; 00175 min_max_angle_by_dist[di][1] = -DBL_MAX; 00176 } 00177 00178 // Compute distance by normal angle distribution for points 00179 std::vector<int>::const_iterator i, begin (indices.begin()), end (indices.end()); 00180 for(i = begin+1; i != end; ++i) 00181 { 00182 // compute angle between the two lines going through normals (disregard orientation!) 00183 double cosine = normals->points[*i].normal[0] * normals->points[*begin].normal[0] + 00184 normals->points[*i].normal[1] * normals->points[*begin].normal[1] + 00185 normals->points[*i].normal[2] * normals->points[*begin].normal[2]; 00186 if (cosine > 1) cosine = 1; 00187 if (cosine < -1) cosine = -1; 00188 double angle = acos (cosine); 00189 if (angle > M_PI/2) angle = M_PI - angle; /// \note: orientation is neglected! 00190 00191 // Compute point to point distance 00192 double dist = sqrt (sqr_dists[i-begin]); 00193 00194 if (dist > max_dist) 00195 continue; /// \note: we neglect points that are outside the specified interval! 00196 00197 // compute bins and increase 00198 int bin_d = static_cast<int> (floor (nr_subdiv * dist / max_dist)); 00199 if (compute_histogram) 00200 { 00201 int bin_a = std::min (nr_subdiv-1, static_cast<int> (floor (nr_subdiv * angle / (M_PI/2)))); 00202 histogram(bin_a, bin_d)++; 00203 } 00204 00205 // update min-max values for distance bins 00206 if (min_max_angle_by_dist[bin_d][0] > angle) min_max_angle_by_dist[bin_d][0] = angle; 00207 if (min_max_angle_by_dist[bin_d][1] < angle) min_max_angle_by_dist[bin_d][1] = angle; 00208 } 00209 00210 // Estimate radius from min and max lines 00211 double Amint_Amin = 0, Amint_d = 0; 00212 double Amaxt_Amax = 0, Amaxt_d = 0; 00213 for (int di=0; di<nr_subdiv; di++) 00214 { 00215 // combute the members of A'*A*r = A'*D 00216 if (min_max_angle_by_dist[di][1] >= 0) 00217 { 00218 double p_min = min_max_angle_by_dist[di][0]; 00219 double p_max = min_max_angle_by_dist[di][1]; 00220 double f = (di+0.5)*max_dist/nr_subdiv; 00221 Amint_Amin += p_min * p_min; 00222 Amint_d += p_min * f; 00223 Amaxt_Amax += p_max * p_max; 00224 Amaxt_d += p_max * f; 00225 } 00226 } 00227 float min_radius = Amint_Amin == 0.0f ? float (plane_radius) : float (std::min (Amint_d/Amint_Amin, plane_radius)); 00228 float max_radius = Amaxt_Amax == 0.0f ? float (plane_radius) : float (std::min (Amaxt_d/Amaxt_Amax, plane_radius)); 00229 00230 // Small correction of the systematic error of the estimation (based on analysis with nr_subdiv_ = 5) 00231 min_radius *= 1.1f; 00232 max_radius *= 0.9f; 00233 if (min_radius < max_radius) 00234 { 00235 radii.r_min = min_radius; 00236 radii.r_max = max_radius; 00237 } 00238 else 00239 { 00240 radii.r_max = min_radius; 00241 radii.r_min = max_radius; 00242 } 00243 00244 return histogram; 00245 } 00246 00247 ////////////////////////////////////////////////////////////////////////////////////////////// 00248 template <typename PointInT, typename PointNT, typename PointOutT> void 00249 pcl::RSDEstimation<PointInT, PointNT, PointOutT>::computeFeature (PointCloudOut &output) 00250 { 00251 // Check if search_radius_ was set 00252 if (search_radius_ < 0) 00253 { 00254 PCL_ERROR ("[pcl::%s::computeFeature] A search radius needs to be set!\n", getClassName ().c_str ()); 00255 output.width = output.height = 0; 00256 output.points.clear (); 00257 return; 00258 } 00259 00260 // List of indices and corresponding squared distances for a neighborhood 00261 // \note resize is irrelevant for a radiusSearch (). 00262 std::vector<int> nn_indices; 00263 std::vector<float> nn_sqr_dists; 00264 00265 // Check if the full histogram has to be saved or not 00266 if (save_histograms_) 00267 { 00268 // Reserve space for the output histogram dataset 00269 histograms_.reset (new std::vector<Eigen::MatrixXf, Eigen::aligned_allocator<Eigen::MatrixXf> >); 00270 histograms_->reserve (output.points.size ()); 00271 00272 // Iterating over the entire index vector 00273 for (size_t idx = 0; idx < indices_->size (); ++idx) 00274 { 00275 // Compute and store r_min and r_max in the output cloud 00276 this->searchForNeighbors ((*indices_)[idx], search_parameter_, nn_indices, nn_sqr_dists); 00277 //histograms_->push_back (computeRSD (surface_, normals_, nn_indices, search_radius_, nr_subdiv_, plane_radius_, output.points[idx], true)); 00278 histograms_->push_back (computeRSD (normals_, nn_indices, nn_sqr_dists, search_radius_, nr_subdiv_, plane_radius_, output.points[idx], true)); 00279 } 00280 } 00281 else 00282 { 00283 // Iterating over the entire index vector 00284 for (size_t idx = 0; idx < indices_->size (); ++idx) 00285 { 00286 // Compute and store r_min and r_max in the output cloud 00287 this->searchForNeighbors ((*indices_)[idx], search_parameter_, nn_indices, nn_sqr_dists); 00288 //computeRSD (surface_, normals_, nn_indices, search_radius_, nr_subdiv_, plane_radius_, output.points[idx], false); 00289 computeRSD (normals_, nn_indices, nn_sqr_dists, search_radius_, nr_subdiv_, plane_radius_, output.points[idx], false); 00290 } 00291 } 00292 } 00293 00294 #define PCL_INSTANTIATE_RSDEstimation(T,NT,OutT) template class PCL_EXPORTS pcl::RSDEstimation<T,NT,OutT>; 00295 00296 #endif // PCL_FEATURES_IMPL_VFH_H_