features/impl/flare.hpp

/*
* Software License Agreement (BSD License)
*
*  Point Cloud Library (PCL) - www.pointclouds.org
*  Copyright (c) 2016-, Open Perception, Inc.
*
*  All rights reserved.
*
*  Redistribution and use in source and binary forms, with or without
*  modification, are permitted provided that the following conditions
*  are met:
*
*   * Redistributions of source code must retain the above copyright
*     notice, this list of conditions and the following disclaimer.
*   * Redistributions in binary form must reproduce the above
*     copyright notice, this list of conditions and the following
*     disclaimer in the documentation and/or other materials provided
*     with the distribution.
*   * Neither the name of the copyright holder(s) nor the names of its
*     contributors may be used to endorse or promote products derived
*     from this software without specific prior written permission.
*
*  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
*  "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
*  LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
*  FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
*  COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
*  INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
*  BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
*  LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
*  CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
*  LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
*  ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
*  POSSIBILITY OF SUCH DAMAGE.
*
*
*/

#ifndef PCL_FEATURES_IMPL_FLARE_H_
#define PCL_FEATURES_IMPL_FLARE_H_

#include <pcl/features/flare.h>
#include <pcl/common/geometry.h>

//////////////////////////////////////////////////////////////////////////////////////////////
template<typename PointInT, typename PointNT, typename PointOutT, typename SignedDistanceT> bool
  pcl::FLARELocalReferenceFrameEstimation<PointInT, PointNT, PointOutT, SignedDistanceT>::initCompute ()
{
  if (!FeatureFromNormals<PointInT, PointNT, PointOutT>::initCompute ())
  {
    PCL_ERROR ("[pcl::%s::initCompute] Init failed.\n", getClassName ().c_str ());
    return (false);
  }

  if (tangent_radius_ == 0.0f)
  {
    PCL_ERROR ("[pcl::%s::initCompute] tangent_radius_ not set.\n", getClassName ().c_str ());
    return (false);
  }

  // If no search sampled_surface_ has been defined, use the surface_ dataset as the search sampled_surface_ itself
  if (!sampled_surface_)
  {
    fake_sampled_surface_ = true;
    sampled_surface_ = surface_;

    if (sampled_tree_)
    {
      PCL_WARN ("[pcl::%s::initCompute] sampled_surface_ is not set even if sampled_tree_ is already set.", getClassName ().c_str ());
      PCL_WARN ("sampled_tree_ will be rebuilt from surface_. Use sampled_surface_.\n");
    }
  }

  // Check if a space search locator was given for sampled_surface_
  if (!sampled_tree_)
  {
    if (sampled_surface_->isOrganized () && surface_->isOrganized () && input_->isOrganized ())
      sampled_tree_.reset (new pcl::search::OrganizedNeighbor<PointInT> ());
    else
      sampled_tree_.reset (new pcl::search::KdTree<PointInT> (false));
  }

  if (sampled_tree_->getInputCloud () != sampled_surface_) // Make sure the tree searches the sampled surface
    sampled_tree_->setInputCloud (sampled_surface_);

  return (true);
}

//////////////////////////////////////////////////////////////////////////////////////////////
template<typename PointInT, typename PointNT, typename PointOutT, typename SignedDistanceT> bool
  pcl::FLARELocalReferenceFrameEstimation<PointInT, PointNT, PointOutT, SignedDistanceT>::deinitCompute ()
{
  // Reset the surface
  if (fake_surface_)
  {
    surface_.reset ();
    fake_surface_ = false;
  }
  // Reset the sampled surface
  if (fake_sampled_surface_)
  {
    sampled_surface_.reset ();
    fake_sampled_surface_ = false;
  }
  return (true);
}

//////////////////////////////////////////////////////////////////////////////////////////////
template<typename PointInT, typename PointNT, typename PointOutT, typename SignedDistanceT> SignedDistanceT
  pcl::FLARELocalReferenceFrameEstimation<PointInT, PointNT, PointOutT, SignedDistanceT>::computePointLRF (const int index,
  Eigen::Matrix3f &lrf)
{
  Eigen::Vector3f x_axis, y_axis;
  Eigen::Vector3f fitted_normal; //z_axis

  //find Z axis

  //extract support points for the computation of Z axis
  pcl::Indices neighbours_indices;
  std::vector<float> neighbours_distances;

  const std::size_t n_normal_neighbours =
      this->searchForNeighbors (index, search_parameter_, neighbours_indices, neighbours_distances);
  if (n_normal_neighbours < static_cast<std::size_t>(min_neighbors_for_normal_axis_))
  {
    if (!pcl::isFinite ((*normals_)[index]))
    {
      //normal is invalid
      //setting lrf to NaN
      lrf.setConstant (std::numeric_limits<float>::quiet_NaN ());
      return (std::numeric_limits<SignedDistanceT>::max ());
    }

    //set z_axis as the normal of index point
    fitted_normal = (*normals_)[index].getNormalVector3fMap ();
  }
  else
  {
    float plane_curvature;
    normal_estimation_.computePointNormal (*surface_, neighbours_indices, fitted_normal (0), fitted_normal (1), fitted_normal (2), plane_curvature);

    //disambiguate Z axis with normal mean
    if (!pcl::flipNormalTowardsNormalsMean<PointNT> (*normals_, neighbours_indices, fitted_normal))
    {
      //all normals in the neighbourood are invalid
      //setting lrf to NaN
      lrf.setConstant (std::numeric_limits<float>::quiet_NaN ());
      return (std::numeric_limits<SignedDistanceT>::max ());
    }
  }

  //setting LRF Z axis
  lrf.row (2).matrix () = fitted_normal;

  //find X axis

  //extract support points for Rx radius
  const std::size_t n_tangent_neighbours =
      sampled_tree_->radiusSearch ((*input_)[index], tangent_radius_, neighbours_indices, neighbours_distances);

  if (n_tangent_neighbours < static_cast<std::size_t>(min_neighbors_for_tangent_axis_))
  {
    //set X axis as a random axis
    x_axis = pcl::geometry::randomOrthogonalAxis (fitted_normal);
    y_axis = fitted_normal.cross (x_axis);

    lrf.row (0).matrix () = x_axis;
    lrf.row (1).matrix () = y_axis;

    return (std::numeric_limits<SignedDistanceT>::max ());
  }

  //find point with the largest signed distance from tangent plane

  SignedDistanceT shape_score;
  SignedDistanceT best_shape_score = -std::numeric_limits<SignedDistanceT>::max ();
  int best_shape_index = -1;

  Eigen::Vector3f best_margin_point;

  const float radius2 = tangent_radius_ * tangent_radius_;
  const float margin_distance2 = margin_thresh_ * margin_thresh_ * radius2;

  Vector3fMapConst feature_point = (*input_)[index].getVector3fMap ();

  for (std::size_t curr_neigh = 0; curr_neigh < n_tangent_neighbours; ++curr_neigh)
  {
    const int& curr_neigh_idx = neighbours_indices[curr_neigh];
    const float& neigh_distance_sqr = neighbours_distances[curr_neigh];

    if (neigh_distance_sqr <= margin_distance2)
    {
      continue;
    }

    //point curr_neigh_idx is inside the ring between marginThresh and Radius

    shape_score = fitted_normal.dot ((*sampled_surface_)[curr_neigh_idx].getVector3fMap ());

    if (shape_score > best_shape_score)
    {
      best_shape_index = curr_neigh_idx;
      best_shape_score = shape_score;
    }
  } //for each neighbor

  if (best_shape_index == -1)
  {
    x_axis = pcl::geometry::randomOrthogonalAxis (fitted_normal);
    y_axis = fitted_normal.cross (x_axis);

    lrf.row (0).matrix () = x_axis;
    lrf.row (1).matrix () = y_axis;

    return (std::numeric_limits<SignedDistanceT>::max ());
  }

  //find orthogonal axis directed to best_shape_index point projection on plane with fittedNormal as axis
  x_axis = pcl::geometry::projectedAsUnitVector (sampled_surface_->at (best_shape_index).getVector3fMap (), feature_point, fitted_normal);

  y_axis = fitted_normal.cross (x_axis);

  lrf.row (0).matrix () = x_axis;
  lrf.row (1).matrix () = y_axis;
  //z axis already set

  best_shape_score -= fitted_normal.dot (feature_point);
  return (best_shape_score);
}

//////////////////////////////////////////////////////////////////////////////////////////////
template<typename PointInT, typename PointNT, typename PointOutT, typename SignedDistanceT> void
  pcl::FLARELocalReferenceFrameEstimation<PointInT, PointNT, PointOutT, SignedDistanceT>::computeFeature (PointCloudOut &output)
{
  //check whether used with search radius or search k-neighbors
  if (this->getKSearch () != 0)
  {
    PCL_ERROR (
      "[pcl::%s::computeFeature] Error! Search method set to k-neighborhood. Call setKSearch (0) and setRadiusSearch (radius) to use this class.\n",
      getClassName ().c_str ());
    return;
  }

  signed_distances_from_highest_points_.resize (indices_->size ());

  for (std::size_t point_idx = 0; point_idx < indices_->size (); ++point_idx)
  {
    Eigen::Matrix3f currentLrf;
    PointOutT &rf = output[point_idx];

    signed_distances_from_highest_points_[point_idx] = computePointLRF ((*indices_)[point_idx], currentLrf);
    if (signed_distances_from_highest_points_[point_idx] == std::numeric_limits<SignedDistanceT>::max ())
    {
      output.is_dense = false;
    }

    rf.getXAxisVector3fMap () = currentLrf.row (0);
    rf.getYAxisVector3fMap () = currentLrf.row (1);
    rf.getZAxisVector3fMap () = currentLrf.row (2);
  }
}

#define PCL_INSTANTIATE_FLARELocalReferenceFrameEstimation(T,NT,OutT,SdT) template class PCL_EXPORTS pcl::FLARELocalReferenceFrameEstimation<T,NT,OutT,SdT>;

#endif // PCL_FEATURES_IMPL_FLARE_H_