template <typename PointInT, typename PointOutT> void
pcl::MovingLeastSquares<PointInT, PointOutT>::addProjectedPointNormal (int index,
const Eigen::Vector3d &point,
const Eigen::Vector3d &normal,
double curvature,
PointCloudOut &projected_points,
NormalCloud &projected_points_normals,
PointIndices &corresponding_input_indices) const
{
PointOutT aux;
aux.x = static_cast<float> (point[0]);
aux.y = static_cast<float> (point[1]);
aux.z = static_cast<float> (point[2]);
// Copy additional point information if available
copyMissingFields (input_->points[index], aux);
projected_points.push_back (aux);
corresponding_input_indices.indices.push_back (index);
if (compute_normals_)
{
pcl::Normal aux_normal;
aux_normal.normal_x = static_cast<float> (normal[0]);
aux_normal.normal_y = static_cast<float> (normal[1]);
aux_normal.normal_z = static_cast<float> (normal[2]);
aux_normal.curvature = curvature;
projected_points_normals.push_back (aux_normal);
}
}
template <typename PointInT, typename PointOutT> void
pcl::MovingLeastSquares<PointInT, PointOutT>::performProcessing (PointCloudOut &output)
{
// Compute the number of coefficients
nr_coeff_ = (order_ + 1) * (order_ + 2) / 2;
// Allocate enough space to hold the results of nearest neighbor searches
// \note resize is irrelevant for a radiusSearch ().
std::vector<int> nn_indices;
std::vector<float> nn_sqr_dists;
// For all points
for (size_t cp = 0; cp < indices_->size (); ++cp)
{
// Get the initial estimates of point positions and their neighborhoods
if (!searchForNeighbors ((*indices_)[cp], nn_indices, nn_sqr_dists))
continue;
// Check the number of nearest neighbors for normal estimation (and later
// for polynomial fit as well)
if (nn_indices.size () < 3)
continue;
PointCloudOut projected_points;
NormalCloud projected_points_normals;
// Get a plane approximating the local surface's tangent and project point onto it
int index = (*indices_)[cp];
computeMLSPointNormal (index, nn_indices, nn_sqr_dists, projected_points, projected_points_normals, *corresponding_input_indices_, mls_results_[index]);
// Copy all information from the input cloud to the output points (not doing any interpolation)
for (size_t pp = 0; pp < projected_points.size (); ++pp)
copyMissingFields (input_->points[(*indices_)[cp]], projected_points[pp]);
// Append projected points to output
output.insert (output.end (), projected_points.begin (), projected_points.end ());
if (compute_normals_)
normals_->insert (normals_->end (), projected_points_normals.begin (), projected_points_normals.end ());
}
// Perform the distinct-cloud or voxel-grid upsampling
performUpsampling (output);
}
template <typename PointInT, typename PointOutT> void
pcl::MovingLeastSquares<PointInT, PointOutT>::performUpsampling (PointCloudOut &output)
{
if (upsample_method_ == DISTINCT_CLOUD)
{
for (size_t dp_i = 0; dp_i < distinct_cloud_->size (); ++dp_i) // dp_i = distinct_point_i
{
// Distinct cloud may have nan points, skip them
if (!pcl_isfinite (distinct_cloud_->points[dp_i].x))
continue;
// Get 3D position of point
//Eigen::Vector3f pos = distinct_cloud_->points[dp_i].getVector3fMap ();
std::vector<int> nn_indices;
std::vector<float> nn_dists;
tree_->nearestKSearch (distinct_cloud_->points[dp_i], 1, nn_indices, nn_dists);
int input_index = nn_indices.front ();
// If the closest point did not have a valid MLS fitting result
// OR if it is too far away from the sampled point
if (mls_results_[input_index].valid == false)
continue;
Eigen::Vector3d add_point = distinct_cloud_->points[dp_i].getVector3fMap ().template cast<double> ();
float u_disp = static_cast<float> ((add_point - mls_results_[input_index].mean).dot (mls_results_[input_index].u_axis)),
v_disp = static_cast<float> ((add_point - mls_results_[input_index].mean).dot (mls_results_[input_index].v_axis));
PointOutT result_point;
pcl::Normal result_normal;
projectPointToMLSSurface (u_disp, v_disp,
mls_results_[input_index].u_axis, mls_results_[input_index].v_axis,
mls_results_[input_index].plane_normal,
mls_results_[input_index].mean,
mls_results_[input_index].curvature,
mls_results_[input_index].c_vec,
mls_results_[input_index].num_neighbors,
result_point, result_normal);
// Copy additional point information if available
copyMissingFields (input_->points[input_index], result_point);
// Store the id of the original point
corresponding_input_indices_->indices.push_back (input_index);
output.push_back (result_point);
if (compute_normals_)
normals_->push_back (result_normal);
}
}
// For the voxel grid upsampling method, generate the voxel grid and dilate it
// Then, project the newly obtained points to the MLS surface
if (upsample_method_ == VOXEL_GRID_DILATION)
{
MLSVoxelGrid voxel_grid (input_, indices_, voxel_size_);
for (int iteration = 0; iteration < dilation_iteration_num_; ++iteration)
voxel_grid.dilate ();
for (typename MLSVoxelGrid::HashMap::iterator m_it = voxel_grid.voxel_grid_.begin (); m_it != voxel_grid.voxel_grid_.end (); ++m_it)
{
// Get 3D position of point
Eigen::Vector3f pos;
voxel_grid.getPosition (m_it->first, pos);
PointInT p;
p.x = pos[0];
p.y = pos[1];
p.z = pos[2];
std::vector<int> nn_indices;
std::vector<float> nn_dists;
tree_->nearestKSearch (p, 1, nn_indices, nn_dists);
int input_index = nn_indices.front ();
// If the closest point did not have a valid MLS fitting result
// OR if it is too far away from the sampled point
if (mls_results_[input_index].valid == false)
continue;
Eigen::Vector3d add_point = p.getVector3fMap ().template cast<double> ();
float u_disp = static_cast<float> ((add_point - mls_results_[input_index].mean).dot (mls_results_[input_index].u_axis)),
v_disp = static_cast<float> ((add_point - mls_results_[input_index].mean).dot (mls_results_[input_index].v_axis));
PointOutT result_point;
pcl::Normal result_normal;
projectPointToMLSSurface (u_disp, v_disp,
mls_results_[input_index].u_axis, mls_results_[input_index].v_axis,
mls_results_[input_index].plane_normal,
mls_results_[input_index].mean,
mls_results_[input_index].curvature,
mls_results_[input_index].c_vec,
mls_results_[input_index].num_neighbors,
result_point, result_normal);
// Copy additional point information if available
copyMissingFields (input_->points[input_index], result_point);
// Store the id of the original point
corresponding_input_indices_->indices.push_back (input_index);
output.push_back (result_point);
if (compute_normals_)
normals_->push_back (result_normal);
}
}
}
template <typename PointInT, typename PointOutT> void
pcl::MovingLeastSquares<PointInT, PointOutT>::performProcessing (PointCloudOut &output)
{
// Compute the number of coefficients
nr_coeff_ = (order_ + 1) * (order_ + 2) / 2;
#ifdef _OPENMP
// (Maximum) number of threads
const unsigned int threads = threads_ == 0 ? 1 : threads_;
// Create temporaries for each thread in order to avoid synchronization
typename PointCloudOut::CloudVectorType projected_points (threads);
typename NormalCloud::CloudVectorType projected_points_normals (threads);
std::vector<PointIndices> corresponding_input_indices (threads);
#endif
// For all points
#ifdef _OPENMP
#pragma omp parallel for schedule (dynamic,1000) num_threads (threads)
#endif
for (int cp = 0; cp < static_cast<int> (indices_->size ()); ++cp)
{
// Allocate enough space to hold the results of nearest neighbor searches
// \note resize is irrelevant for a radiusSearch ().
std::vector<int> nn_indices;
std::vector<float> nn_sqr_dists;
// Get the initial estimates of point positions and their neighborhoods
if (searchForNeighbors ((*indices_)[cp], nn_indices, nn_sqr_dists))
{
// Check the number of nearest neighbors for normal estimation (and later for polynomial fit as well)
if (nn_indices.size () >= 3)
{
// This thread's ID (range 0 to threads-1)
#ifdef _OPENMP
const int tn = omp_get_thread_num ();
// Size of projected points before computeMLSPointNormal () adds points
size_t pp_size = projected_points[tn].size ();
#else
PointCloudOut projected_points;
NormalCloud projected_points_normals;
#endif
// Get a plane approximating the local surface's tangent and project point onto it
const int index = (*indices_)[cp];
size_t mls_result_index = 0;
if (cache_mls_results_)
mls_result_index = index; // otherwise we give it a dummy location.
#ifdef _OPENMP
computeMLSPointNormal (index, nn_indices, projected_points[tn], projected_points_normals[tn], corresponding_input_indices[tn], mls_results_[mls_result_index]);
// Copy all information from the input cloud to the output points (not doing any interpolation)
for (size_t pp = pp_size; pp < projected_points[tn].size (); ++pp)
copyMissingFields (input_->points[(*indices_)[cp]], projected_points[tn][pp]);
#else
computeMLSPointNormal (index, nn_indices, projected_points, projected_points_normals, *corresponding_input_indices_, mls_results_[mls_result_index]);
// Append projected points to output
output.insert (output.end (), projected_points.begin (), projected_points.end ());
if (compute_normals_)
normals_->insert (normals_->end (), projected_points_normals.begin (), projected_points_normals.end ());
#endif
}
}
}
#ifdef _OPENMP
// Combine all threads' results into the output vectors
for (unsigned int tn = 0; tn < threads; ++tn)
{
output.insert (output.end (), projected_points[tn].begin (), projected_points[tn].end ());
corresponding_input_indices_->indices.insert (corresponding_input_indices_->indices.end (),
corresponding_input_indices[tn].indices.begin (), corresponding_input_indices[tn].indices.end ());
if (compute_normals_)
normals_->insert (normals_->end (), projected_points_normals[tn].begin (), projected_points_normals[tn].end ());
}
#endif
// Perform the distinct-cloud or voxel-grid upsampling
performUpsampling (output);
}
