Point Cloud Library (PCL)  1.14.0-dev
bilateral_upsampling.hpp
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2009-2012, Willow Garage, Inc.
6  *
7  * All rights reserved.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  *
13  * * Redistributions of source code must retain the above copyright
14  * notice, this list of conditions and the following disclaimer.
15  * * Redistributions in binary form must reproduce the above
16  * copyright notice, this list of conditions and the following
17  * disclaimer in the documentation and/or other materials provided
18  * with the distribution.
19  * * Neither the name of Willow Garage, Inc. nor the names of its
20  * contributors may be used to endorse or promote products derived
21  * from this software without specific prior written permission.
22  *
23  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
24  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
25  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
26  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
27  * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
28  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
29  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
30  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
31  * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
33  * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
34  * POSSIBILITY OF SUCH DAMAGE.
35  *
36  */
37 
38 
39 #ifndef PCL_SURFACE_IMPL_BILATERAL_UPSAMPLING_H_
40 #define PCL_SURFACE_IMPL_BILATERAL_UPSAMPLING_H_
41 
42 #include <pcl/surface/bilateral_upsampling.h>
43 #include <algorithm>
44 #include <pcl/console/print.h>
45 
46 #include <Eigen/LU> // for inverse
47 
48 //////////////////////////////////////////////////////////////////////////////////////////////
49 template <typename PointInT, typename PointOutT> void
51 {
52  // Copy the header
53  output.header = input_->header;
54 
55  if (!initCompute ())
56  {
57  output.width = output.height = 0;
58  output.clear ();
59  return;
60  }
61 
62  if (input_->isOrganized () == false)
63  {
64  PCL_ERROR ("Input cloud is not organized.\n");
65  return;
66  }
67 
68  // Invert projection matrix
69  unprojection_matrix_ = projection_matrix_.inverse ();
70 
71  for (int i = 0; i < 3; ++i)
72  {
73  for (int j = 0; j < 3; ++j)
74  printf ("%f ", unprojection_matrix_(i, j));
75 
76  printf ("\n");
77  }
78 
79 
80  // Perform the actual surface reconstruction
81  performProcessing (output);
82 
83  deinitCompute ();
84 }
85 
86 //////////////////////////////////////////////////////////////////////////////////////////////
87 template <typename PointInT, typename PointOutT> void
89 {
90  output.resize (input_->size ());
91  float nan = std::numeric_limits<float>::quiet_NaN ();
92 
93  Eigen::MatrixXf val_exp_depth_matrix;
94  Eigen::VectorXf val_exp_rgb_vector;
95  computeDistances (val_exp_depth_matrix, val_exp_rgb_vector);
96 
97  for (int x = 0; x < static_cast<int> (input_->width); ++x)
98  for (int y = 0; y < static_cast<int> (input_->height); ++y)
99  {
100  int start_window_x = std::max (x - window_size_, 0),
101  start_window_y = std::max (y - window_size_, 0),
102  end_window_x = std::min (x + window_size_, static_cast<int> (input_->width)),
103  end_window_y = std::min (y + window_size_, static_cast<int> (input_->height));
104 
105  float sum = 0.0f,
106  norm_sum = 0.0f;
107 
108  for (int x_w = start_window_x; x_w < end_window_x; ++ x_w)
109  for (int y_w = start_window_y; y_w < end_window_y; ++ y_w)
110  {
111  float val_exp_depth = val_exp_depth_matrix (static_cast<Eigen::MatrixXf::Index> (x - x_w + window_size_),
112  static_cast<Eigen::MatrixXf::Index> (y - y_w + window_size_));
113 
114  auto d_color = static_cast<Eigen::VectorXf::Index> (
115  std::abs ((*input_)[y_w * input_->width + x_w].r - (*input_)[y * input_->width + x].r) +
116  std::abs ((*input_)[y_w * input_->width + x_w].g - (*input_)[y * input_->width + x].g) +
117  std::abs ((*input_)[y_w * input_->width + x_w].b - (*input_)[y * input_->width + x].b));
118 
119  float val_exp_rgb = val_exp_rgb_vector (d_color);
120 
121  if (std::isfinite ((*input_)[y_w*input_->width + x_w].z))
122  {
123  sum += val_exp_depth * val_exp_rgb * (*input_)[y_w*input_->width + x_w].z;
124  norm_sum += val_exp_depth * val_exp_rgb;
125  }
126  }
127 
128  output[y*input_->width + x].r = (*input_)[y*input_->width + x].r;
129  output[y*input_->width + x].g = (*input_)[y*input_->width + x].g;
130  output[y*input_->width + x].b = (*input_)[y*input_->width + x].b;
131 
132  if (norm_sum != 0.0f)
133  {
134  float depth = sum / norm_sum;
135  Eigen::Vector3f pc (static_cast<float> (x) * depth, static_cast<float> (y) * depth, depth);
136  Eigen::Vector3f pw (unprojection_matrix_ * pc);
137  output[y*input_->width + x].x = pw[0];
138  output[y*input_->width + x].y = pw[1];
139  output[y*input_->width + x].z = pw[2];
140  }
141  else
142  {
143  output[y*input_->width + x].x = nan;
144  output[y*input_->width + x].y = nan;
145  output[y*input_->width + x].z = nan;
146  }
147  }
148 
149  output.header = input_->header;
150  output.width = input_->width;
151  output.height = input_->height;
152 }
153 
154 
155 template <typename PointInT, typename PointOutT> void
156 pcl::BilateralUpsampling<PointInT, PointOutT>::computeDistances (Eigen::MatrixXf &val_exp_depth, Eigen::VectorXf &val_exp_rgb)
157 {
158  val_exp_depth.resize (2*window_size_+1,2*window_size_+1);
159  val_exp_rgb.resize (3*255+1);
160 
161  int j = 0;
162  for (int dx = -window_size_; dx < window_size_+1; ++dx)
163  {
164  int i = 0;
165  for (int dy = -window_size_; dy < window_size_+1; ++dy)
166  {
167  float val_exp = std::exp (- (dx*dx + dy*dy) / (2.0f * static_cast<float> (sigma_depth_ * sigma_depth_)));
168  val_exp_depth(i,j) = val_exp;
169  i++;
170  }
171  j++;
172  }
173 
174  for (int d_color = 0; d_color < 3*255+1; d_color++)
175  {
176  float val_exp = std::exp (- d_color * d_color / (2.0f * sigma_color_ * sigma_color_));
177  val_exp_rgb(d_color) = val_exp;
178  }
179 }
180 
181 
182 #define PCL_INSTANTIATE_BilateralUpsampling(T,OutT) template class PCL_EXPORTS pcl::BilateralUpsampling<T,OutT>;
183 
184 
185 #endif /* PCL_SURFACE_IMPL_BILATERAL_UPSAMPLING_H_ */
void computeDistances(Eigen::MatrixXf &val_exp_depth, Eigen::VectorXf &val_exp_rgb)
Computes the distance for depth and RGB.
void process(pcl::PointCloud< PointOutT > &output) override
Method that does the actual processing on the input cloud.
void performProcessing(pcl::PointCloud< PointOutT > &output) override
Abstract cloud processing method.
void resize(std::size_t count)
Resizes the container to contain count elements.
Definition: point_cloud.h:462
std::uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:398
pcl::PCLHeader header
The point cloud header.
Definition: point_cloud.h:392
std::uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:400
void clear()
Removes all points in a cloud and sets the width and height to 0.
Definition: point_cloud.h:885