40 #ifndef PCL_FILTERS_CONVOLUTION_3D_IMPL_HPP
41 #define PCL_FILTERS_CONVOLUTION_3D_IMPL_HPP
43 #include <pcl/pcl_config.h>
44 #include <pcl/point_types.h>
45 #include <pcl/common/point_operators.h>
52 template <
typename Po
intT>
58 n.normal_x = n.normal_y = n.normal_z = std::numeric_limits<float>::quiet_NaN ();
62 template <
typename Po
intT>
class
68 p.
x = p.
y = std::numeric_limits<float>::quiet_NaN ();
75 template<
typename Po
intInT,
typename Po
intOutT>
bool
80 PCL_ERROR (
"Sigma is not set or equal to 0!\n", sigma_);
83 sigma_sqr_ = sigma_ * sigma_;
85 if (sigma_coefficient_)
87 if ((*sigma_coefficient_) > 6 || (*sigma_coefficient_) < 3)
89 PCL_ERROR (
"Sigma coefficient (%f) out of [3..6]!\n", (*sigma_coefficient_));
93 threshold_ = (*sigma_coefficient_) * (*sigma_coefficient_) * sigma_sqr_;
100 template<
typename Po
intInT,
typename Po
intOutT> PointOutT
102 const std::vector<float>& distances)
106 float total_weight = 0;
107 std::vector<float>::const_iterator dist_it = distances.begin ();
109 for (std::vector<int>::const_iterator idx_it = indices.begin ();
110 idx_it != indices.end ();
113 if (*dist_it <= threshold_ &&
isFinite ((*input_) [*idx_it]))
115 float weight = expf (-0.5f * (*dist_it) / sigma_sqr_);
116 result += weight * (*input_) [*idx_it];
117 total_weight += weight;
120 if (total_weight != 0)
121 result /= total_weight;
123 makeInfinite (result);
129 template<
typename Po
intInT,
typename Po
intOutT> PointOutT
134 float total_weight = 0;
135 float r = 0, g = 0, b = 0;
136 std::vector<float>::const_iterator dist_it = distances.begin ();
138 for (std::vector<int>::const_iterator idx_it = indices.begin ();
139 idx_it != indices.end ();
142 if (*dist_it <= threshold_ &&
isFinite ((*input_) [*idx_it]))
144 float weight = expf (-0.5f * (*dist_it) / sigma_sqr_);
145 result.x += weight * (*input_) [*idx_it].x;
146 result.y += weight * (*input_) [*idx_it].y;
147 result.z += weight * (*input_) [*idx_it].z;
148 r += weight *
static_cast<float> ((*input_) [*idx_it].r);
149 g += weight *
static_cast<float> ((*input_) [*idx_it].g);
150 b += weight *
static_cast<float> ((*input_) [*idx_it].b);
151 total_weight += weight;
154 if (total_weight != 0)
156 total_weight = 1.f/total_weight;
157 r*= total_weight; g*= total_weight; b*= total_weight;
158 result.x*= total_weight; result.y*= total_weight; result.z*= total_weight;
159 result.r =
static_cast<pcl::uint8_t
> (r);
160 result.g =
static_cast<pcl::uint8_t
> (g);
161 result.b =
static_cast<pcl::uint8_t
> (b);
164 makeInfinite (result);
170 template <
typename Po
intInT,
typename Po
intOutT,
typename KernelT>
179 template <
typename Po
intInT,
typename Po
intOutT,
typename KernelT>
bool
184 PCL_ERROR (
"[pcl::filters::Convlution3D::initCompute] init failed!\n");
190 if (input_->isOrganized ())
199 tree_->setInputCloud (surface_);
201 if (search_radius_ <= 0.0)
203 PCL_ERROR (
"[pcl::filters::Convlution3D::initCompute] search radius (%f) must be > 0",
210 PCL_ERROR (
"[pcl::filters::Convlution3D::initCompute] init failed");
211 PCL_ERROR (
"kernel_ must implement ConvolvingKernel interface\n!");
214 kernel_.setInputCloud (surface_);
216 if (!kernel_.initCompute ())
218 PCL_ERROR (
"[pcl::filters::Convlution3D::initCompute] kernel initialization failed!\n");
225 template <
typename Po
intInT,
typename Po
intOutT,
typename KernelT>
void
230 PCL_ERROR (
"[pcl::filters::Convlution3D::convolve] init failed!\n");
233 output.
resize (surface_->size ());
234 output.
width = surface_->width;
235 output.
height = surface_->height;
236 output.
is_dense = surface_->is_dense;
237 std::vector<int> nn_indices;
238 std::vector<float> nn_distances;
241 #pragma omp parallel for shared (output) private (nn_indices, nn_distances) num_threads (threads_)
243 for (std::size_t point_idx = 0; point_idx < surface_->size (); ++point_idx)
245 const PointInT& point_in = surface_->points [point_idx];
246 PointOutT& point_out = output [point_idx];
248 tree_->radiusSearch (point_in, search_radius_, nn_indices, nn_distances))
250 point_out = kernel_ (nn_indices, nn_distances);
254 kernel_.makeInfinite (point_out);
Convolution3D()
Constructor.
A 2D point structure representing Euclidean xy coordinates.
virtual PointOutT operator()(const std::vector< int > &indices, const std::vector< float > &distances)
Convolve point at the center of this local information.
bool initCompute()
Must call this methode before doing any computation.
void resize(size_t n)
Resize the cloud.
uint32_t width
The point cloud width (if organized as an image-structure).
bool is_dense
True if no points are invalid (e.g., have NaN or Inf values).
void convolve(PointCloudOut &output)
Convolve point cloud.
bool isFinite(const PointT &pt)
Tests if the 3D components of a point are all finite param[in] pt point to be tested.
Class ConvolvingKernel base class for all convolving kernels.
A point structure representing normal coordinates and the surface curvature estimate.
uint32_t height
The point cloud height (if organized as an image-structure).
static void makeInfinite(PointOutT &p)
Utility function that annihilates a point making it fail the pcl::isFinite test.
bool initCompute()
initialize computation
A point structure representing Euclidean xyz coordinates, and the RGB color.
OrganizedNeighbor is a class for optimized nearest neigbhor search in organized point clouds...
PointOutT operator()(const std::vector< int > &indices, const std::vector< float > &distances)
Convolve point at the center of this local information.
PointCloud represents the base class in PCL for storing collections of 3D points. ...
search::KdTree is a wrapper class which inherits the pcl::KdTree class for performing search function...