Point Cloud Library (PCL)
1.3.1
|
00001 /* 00002 * Software License Agreement (BSD License) 00003 * 00004 * Copyright (c) 2009, Willow Garage, Inc. 00005 * All rights reserved. 00006 * 00007 * Redistribution and use in source and binary forms, with or without 00008 * modification, are permitted provided that the following conditions 00009 * are met: 00010 * 00011 * * Redistributions of source code must retain the above copyright 00012 * notice, this list of conditions and the following disclaimer. 00013 * * Redistributions in binary form must reproduce the above 00014 * copyright notice, this list of conditions and the following 00015 * disclaimer in the documentation and/or other materials provided 00016 * with the distribution. 00017 * * Neither the name of Willow Garage, Inc. nor the names of its 00018 * contributors may be used to endorse or promote products derived 00019 * from this software without specific prior written permission. 00020 * 00021 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 00022 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 00023 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS 00024 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 00025 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 00026 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, 00027 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 00028 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER 00029 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 00030 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN 00031 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 00032 * POSSIBILITY OF SUCH DAMAGE. 00033 * 00034 * $Id: fpfh_omp.hpp 3022 2011-11-01 03:42:22Z rusu $ 00035 * 00036 */ 00037 00038 #ifndef PCL_FEATURES_IMPL_FPFH_OMP_H_ 00039 #define PCL_FEATURES_IMPL_FPFH_OMP_H_ 00040 00041 #include "pcl/features/fpfh_omp.h" 00042 00044 template <typename PointInT, typename PointNT, typename PointOutT> void 00045 pcl::FPFHEstimationOMP<PointInT, PointNT, PointOutT>::computeFeature (PointCloudOut &output) 00046 { 00047 std::vector<int> spfh_indices_vec; 00048 std::vector<int> spfh_hist_lookup (surface_->points.size ()); 00049 00050 // Build a list of (unique) indices for which we will need to compute SPFH signatures 00051 // (We need an SPFH signature for every point that is a neighbor of any point in input_[indices_]) 00052 if (surface_ != input_ || 00053 indices_->size () != surface_->points.size ()) 00054 { 00055 std::vector<int> nn_indices (k_); // \note These resizes are irrelevant for a radiusSearch (). 00056 std::vector<float> nn_dists (k_); 00057 00058 std::set<int> spfh_indices_set; 00059 for (size_t idx = 0; idx < indices_->size (); ++idx) 00060 { 00061 int p_idx = (*indices_)[idx]; 00062 this->searchForNeighbors (p_idx, search_parameter_, nn_indices, nn_dists); 00063 00064 spfh_indices_set.insert (nn_indices.begin (), nn_indices.end ()); 00065 } 00066 spfh_indices_vec.resize (spfh_indices_set.size ()); 00067 std::copy (spfh_indices_set.begin (), spfh_indices_set.end (), spfh_indices_vec.begin ()); 00068 } 00069 else 00070 { 00071 // Special case: When a feature must be computed at every point, there is no need for a neighborhood search 00072 spfh_indices_vec.resize (indices_->size ()); 00073 for (size_t idx = 0; idx < indices_->size (); ++idx) 00074 spfh_indices_vec[idx] = idx; 00075 } 00076 00077 // Initialize the arrays that will store the SPFH signatures 00078 size_t data_size = spfh_indices_vec.size (); 00079 hist_f1_.setZero (data_size, nr_bins_f1_); 00080 hist_f2_.setZero (data_size, nr_bins_f2_); 00081 hist_f3_.setZero (data_size, nr_bins_f3_); 00082 00083 // Compute SPFH signatures for every point that needs them 00084 00085 #pragma omp parallel for schedule (dynamic, threads_) 00086 for (int i = 0; i < (int) spfh_indices_vec.size (); ++i) 00087 { 00088 // Get the next point index 00089 int p_idx = spfh_indices_vec[i]; 00090 00091 // Find the neighborhood around p_idx 00092 std::vector<int> nn_indices (k_); // \note These resizes are irrelevant for a radiusSearch (). 00093 std::vector<float> nn_dists (k_); 00094 this->searchForNeighbors (*surface_, p_idx, search_parameter_, nn_indices, nn_dists); 00095 00096 // Estimate the SPFH signature around p_idx 00097 computePointSPFHSignature (*surface_, *normals_, p_idx, i, nn_indices, hist_f1_, hist_f2_, hist_f3_); 00098 00099 // Populate a lookup table for converting a point index to its corresponding row in the spfh_hist_* matrices 00100 spfh_hist_lookup[p_idx] = i; 00101 } 00102 00103 // Intialize the array that will store the FPFH signature 00104 int nr_bins = nr_bins_f1_ + nr_bins_f2_ + nr_bins_f3_; 00105 00106 // Iterate over the entire index vector 00107 #pragma omp parallel for schedule (dynamic, threads_) 00108 for (int idx = 0; idx < (int) indices_->size (); ++idx) 00109 { 00110 // Find the indices of point idx's neighbors... 00111 std::vector<int> nn_indices (k_); // \note These resizes are irrelevant for a radiusSearch (). 00112 std::vector<float> nn_dists (k_); 00113 this->searchForNeighbors ((*indices_)[idx], search_parameter_, nn_indices, nn_dists); 00114 00115 // ... and remap the nn_indices values so that they represent row indices in the spfh_hist_* matrices 00116 // instead of indices into surface_->points 00117 for (size_t i = 0; i < nn_indices.size (); ++i) 00118 nn_indices[i] = spfh_hist_lookup[nn_indices[i]]; 00119 00120 // Compute the FPFH signature (i.e. compute a weighted combination of local SPFH signatures) ... 00121 Eigen::VectorXf fpfh_histogram = Eigen::VectorXf::Zero (nr_bins); 00122 weightPointSPFHSignature (hist_f1_, hist_f2_, hist_f3_, nn_indices, nn_dists, fpfh_histogram); 00123 00124 // ...and copy it into the output cloud 00125 for (int d = 0; d < fpfh_histogram.size (); ++d) 00126 output.points[idx].histogram[d] = fpfh_histogram[d]; 00127 } 00128 00129 } 00130 00131 /* 00133 template <typename PointInT, typename PointNT, typename PointOutT> void 00134 pcl::FPFHEstimationOMP<PointInT, PointNT, PointOutT>::computeFeature (PointCloudOut &output) 00135 { 00136 int data_size = indices_->size (); 00137 // Reset the whole thing 00138 hist_f1_.setZero (data_size, nr_bins_f1_); 00139 hist_f2_.setZero (data_size, nr_bins_f2_); 00140 hist_f3_.setZero (data_size, nr_bins_f3_); 00141 00142 int nr_bins = nr_bins_f1_ + nr_bins_f2_ + nr_bins_f3_; 00143 00144 // Iterating over the entire index vector 00145 #pragma omp parallel for schedule (dynamic, threads_) 00146 for (int idx = 0; idx < data_size; ++idx) 00147 { 00148 // Allocate enough space to hold the results 00149 // \note This resize is irrelevant for a radiusSearch (). 00150 std::vector<int> nn_indices (k_); 00151 std::vector<float> nn_dists (k_); 00152 00153 this->searchForNeighbors ((*indices_)[idx], search_parameter_, nn_indices, nn_dists); 00154 00155 // Estimate the FPFH signature at each patch 00156 this->computePointSPFHSignature (*surface_, *normals_, (*indices_)[idx], nn_indices, 00157 hist_f1_, hist_f2_, hist_f3_); 00158 } 00159 00160 // Iterating over the entire index vector 00161 #pragma omp parallel for schedule (dynamic, threads_) 00162 for (int idx = 0; idx < data_size; ++idx) 00163 { 00164 // Allocate enough space to hold the results 00165 // \note This resize is irrelevant for a radiusSearch (). 00166 std::vector<int> nn_indices (k_); 00167 std::vector<float> nn_dists (k_); 00168 00169 Eigen::VectorXf fpfh_histogram = Eigen::VectorXf::Zero (nr_bins); 00170 00171 this->searchForNeighbors ((*indices_)[idx], search_parameter_, nn_indices, nn_dists); 00172 00173 weightPointSPFHSignature (hist_f1_, hist_f2_, hist_f3_, nn_indices, nn_dists, fpfh_histogram); 00174 00175 // Copy into the resultant cloud 00176 for (int d = 0; d < fpfh_histogram.size (); ++d) 00177 output.points[idx].histogram[d] = fpfh_histogram[d]; 00178 } 00179 } 00180 */ 00181 00182 #define PCL_INSTANTIATE_FPFHEstimationOMP(T,NT,OutT) template class PCL_EXPORTS pcl::FPFHEstimationOMP<T,NT,OutT>; 00183 00184 #endif // PCL_FEATURES_IMPL_FPFH_OMP_H_ 00185