Point Cloud Library (PCL)  1.9.1-dev
organized_edge_detection.hpp
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2012-, Open Perception, 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 the copyright holder(s) 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 #ifndef PCL_FEATURES_IMPL_ORGANIZED_EDGE_DETECTION_H_
39 #define PCL_FEATURES_IMPL_ORGANIZED_EDGE_DETECTION_H_
40 
41 #include <pcl/2d/edge.h>
42 #include <pcl/features/organized_edge_detection.h>
43 #include <pcl/console/print.h>
44 
45 /**
46  * Directions: 1 2 3
47  * 0 x 4
48  * 7 6 5
49  * e.g. direction y means we came from pixel with label y to the center pixel x
50  */
51 //////////////////////////////////////////////////////////////////////////////
52 template<typename PointT, typename PointLT> void
53 pcl::OrganizedEdgeBase<PointT, PointLT>::compute (pcl::PointCloud<PointLT>& labels, std::vector<pcl::PointIndices>& label_indices) const
54 {
55  pcl::Label invalid_pt;
56  invalid_pt.label = unsigned (0);
57  labels.points.resize (input_->points.size (), invalid_pt);
58  labels.width = input_->width;
59  labels.height = input_->height;
60 
61  extractEdges (labels);
62 
63  assignLabelIndices (labels, label_indices);
64 }
65 
66 //////////////////////////////////////////////////////////////////////////////
67 template<typename PointT, typename PointLT> void
68 pcl::OrganizedEdgeBase<PointT, PointLT>::assignLabelIndices (pcl::PointCloud<PointLT>& labels, std::vector<pcl::PointIndices>& label_indices) const
69 {
70  const unsigned invalid_label = unsigned (0);
71  label_indices.resize (num_of_edgetype_);
72  for (size_t idx = 0; idx < input_->points.size (); idx++)
73  {
74  if (labels[idx].label != invalid_label)
75  {
76  for (int edge_type = 0; edge_type < num_of_edgetype_; edge_type++)
77  {
78  if ((labels[idx].label >> edge_type) & 1)
79  label_indices[edge_type].indices.push_back (idx);
80  }
81  }
82  }
83 }
84 
85 //////////////////////////////////////////////////////////////////////////////
86 template<typename PointT, typename PointLT> void
88 {
89  if ((detecting_edge_types_ & EDGELABEL_NAN_BOUNDARY) || (detecting_edge_types_ & EDGELABEL_OCCLUDING) || (detecting_edge_types_ & EDGELABEL_OCCLUDED))
90  {
91  // Fill lookup table for next points to visit
92  const int num_of_ngbr = 8;
93  Neighbor directions [num_of_ngbr] = {Neighbor(-1, 0, -1),
94  Neighbor(-1, -1, -labels.width - 1),
95  Neighbor( 0, -1, -labels.width ),
96  Neighbor( 1, -1, -labels.width + 1),
97  Neighbor( 1, 0, 1),
98  Neighbor( 1, 1, labels.width + 1),
99  Neighbor( 0, 1, labels.width ),
100  Neighbor(-1, 1, labels.width - 1)};
101 
102  for (int row = 1; row < int(input_->height) - 1; row++)
103  {
104  for (int col = 1; col < int(input_->width) - 1; col++)
105  {
106  int curr_idx = row*int(input_->width) + col;
107  if (!std::isfinite (input_->points[curr_idx].z))
108  continue;
109 
110  float curr_depth = std::abs (input_->points[curr_idx].z);
111 
112  // Calculate depth distances between current point and neighboring points
113  std::vector<float> nghr_dist;
114  nghr_dist.resize (8);
115  bool found_invalid_neighbor = false;
116  for (int d_idx = 0; d_idx < num_of_ngbr; d_idx++)
117  {
118  int nghr_idx = curr_idx + directions[d_idx].d_index;
119  assert (nghr_idx >= 0 && nghr_idx < input_->points.size ());
120  if (!std::isfinite (input_->points[nghr_idx].z))
121  {
122  found_invalid_neighbor = true;
123  break;
124  }
125  nghr_dist[d_idx] = curr_depth - std::abs (input_->points[nghr_idx].z);
126  }
127 
128  if (!found_invalid_neighbor)
129  {
130  // Every neighboring points are valid
131  std::vector<float>::iterator min_itr = std::min_element (nghr_dist.begin (), nghr_dist.end ());
132  std::vector<float>::iterator max_itr = std::max_element (nghr_dist.begin (), nghr_dist.end ());
133  float nghr_dist_min = *min_itr;
134  float nghr_dist_max = *max_itr;
135  float dist_dominant = std::abs (nghr_dist_min) > std::abs (nghr_dist_max) ? nghr_dist_min : nghr_dist_max;
136  if (std::abs (dist_dominant) > th_depth_discon_*std::abs (curr_depth))
137  {
138  // Found a depth discontinuity
139  if (dist_dominant > 0.f)
140  {
141  if (detecting_edge_types_ & EDGELABEL_OCCLUDED)
142  labels[curr_idx].label |= EDGELABEL_OCCLUDED;
143  }
144  else
145  {
146  if (detecting_edge_types_ & EDGELABEL_OCCLUDING)
147  labels[curr_idx].label |= EDGELABEL_OCCLUDING;
148  }
149  }
150  }
151  else
152  {
153  // Some neighboring points are not valid (nan points)
154  // Search for corresponding point across invalid points
155  // Search direction is determined by nan point locations with respect to current point
156  int dx = 0;
157  int dy = 0;
158  int num_of_invalid_pt = 0;
159  for (const auto &direction : directions)
160  {
161  int nghr_idx = curr_idx + direction.d_index;
162  assert (nghr_idx >= 0 && nghr_idx < input_->points.size ());
163  if (!std::isfinite (input_->points[nghr_idx].z))
164  {
165  dx += direction.d_x;
166  dy += direction.d_y;
167  num_of_invalid_pt++;
168  }
169  }
170 
171  // Search directions
172  assert (num_of_invalid_pt > 0);
173  float f_dx = static_cast<float> (dx) / static_cast<float> (num_of_invalid_pt);
174  float f_dy = static_cast<float> (dy) / static_cast<float> (num_of_invalid_pt);
175 
176  // Search for corresponding point across invalid points
177  float corr_depth = std::numeric_limits<float>::quiet_NaN ();
178  for (int s_idx = 1; s_idx < max_search_neighbors_; s_idx++)
179  {
180  int s_row = row + static_cast<int> (std::floor (f_dy*static_cast<float> (s_idx)));
181  int s_col = col + static_cast<int> (std::floor (f_dx*static_cast<float> (s_idx)));
182 
183  if (s_row < 0 || s_row >= int(input_->height) || s_col < 0 || s_col >= int(input_->width))
184  break;
185 
186  if (std::isfinite (input_->points[s_row*int(input_->width)+s_col].z))
187  {
188  corr_depth = std::abs (input_->points[s_row*int(input_->width)+s_col].z);
189  break;
190  }
191  }
192 
193  if (!std::isnan (corr_depth))
194  {
195  // Found a corresponding point
196  float dist = curr_depth - corr_depth;
197  if (std::abs (dist) > th_depth_discon_*std::abs (curr_depth))
198  {
199  // Found a depth discontinuity
200  if (dist > 0.f)
201  {
202  if (detecting_edge_types_ & EDGELABEL_OCCLUDED)
203  labels[curr_idx].label |= EDGELABEL_OCCLUDED;
204  }
205  else
206  {
207  if (detecting_edge_types_ & EDGELABEL_OCCLUDING)
208  labels[curr_idx].label |= EDGELABEL_OCCLUDING;
209  }
210  }
211  }
212  else
213  {
214  // Not found a corresponding point, just nan boundary edge
215  if (detecting_edge_types_ & EDGELABEL_NAN_BOUNDARY)
216  labels[curr_idx].label |= EDGELABEL_NAN_BOUNDARY;
217  }
218  }
219  }
220  }
221  }
222 }
223 
224 
225 //////////////////////////////////////////////////////////////////////////////
226 template<typename PointT, typename PointLT> void
227 pcl::OrganizedEdgeFromRGB<PointT, PointLT>::compute (pcl::PointCloud<PointLT>& labels, std::vector<pcl::PointIndices>& label_indices) const
228 {
229  pcl::Label invalid_pt;
230  invalid_pt.label = unsigned (0);
231  labels.points.resize (input_->points.size (), invalid_pt);
232  labels.width = input_->width;
233  labels.height = input_->height;
234 
236  extractEdges (labels);
237 
238  this->assignLabelIndices (labels, label_indices);
239 }
240 
241 //////////////////////////////////////////////////////////////////////////////
242 template<typename PointT, typename PointLT> void
244 {
245  if ((detecting_edge_types_ & EDGELABEL_RGB_CANNY))
246  {
248  gray->width = input_->width;
249  gray->height = input_->height;
250  gray->resize (input_->height*input_->width);
251 
252  for (size_t i = 0; i < input_->size (); ++i)
253  (*gray)[i].intensity = float (((*input_)[i].r + (*input_)[i].g + (*input_)[i].b) / 3);
254 
257  edge.setInputCloud (gray);
258  edge.setHysteresisThresholdLow (th_rgb_canny_low_);
259  edge.setHysteresisThresholdHigh (th_rgb_canny_high_);
260  edge.detectEdgeCanny (img_edge_rgb);
261 
262  for (uint32_t row=0; row<labels.height; row++)
263  {
264  for (uint32_t col=0; col<labels.width; col++)
265  {
266  if (img_edge_rgb (col, row).magnitude == 255.f)
267  labels[row * labels.width + col].label |= EDGELABEL_RGB_CANNY;
268  }
269  }
270  }
271 }
272 
273 //////////////////////////////////////////////////////////////////////////////
274 template<typename PointT, typename PointNT, typename PointLT> void
275 pcl::OrganizedEdgeFromNormals<PointT, PointNT, PointLT>::compute (pcl::PointCloud<PointLT>& labels, std::vector<pcl::PointIndices>& label_indices) const
276 {
277  pcl::Label invalid_pt;
278  invalid_pt.label = unsigned (0);
279  labels.points.resize (input_->points.size (), invalid_pt);
280  labels.width = input_->width;
281  labels.height = input_->height;
282 
284  extractEdges (labels);
285 
286  this->assignLabelIndices (labels, label_indices);
287 }
288 
289 //////////////////////////////////////////////////////////////////////////////
290 template<typename PointT, typename PointNT, typename PointLT> void
292 {
293  if ((detecting_edge_types_ & EDGELABEL_HIGH_CURVATURE))
294  {
295 
297  nx.width = normals_->width;
298  nx.height = normals_->height;
299  nx.resize (normals_->height*normals_->width);
300 
301  ny.width = normals_->width;
302  ny.height = normals_->height;
303  ny.resize (normals_->height*normals_->width);
304 
305  for (uint32_t row=0; row<normals_->height; row++)
306  {
307  for (uint32_t col=0; col<normals_->width; col++)
308  {
309  nx (col, row).intensity = normals_->points[row*normals_->width + col].normal_x;
310  ny (col, row).intensity = normals_->points[row*normals_->width + col].normal_y;
311  }
312  }
313 
316  edge.setHysteresisThresholdLow (th_hc_canny_low_);
317  edge.setHysteresisThresholdHigh (th_hc_canny_high_);
318  edge.canny (nx, ny, img_edge);
319 
320  for (uint32_t row=0; row<labels.height; row++)
321  {
322  for (uint32_t col=0; col<labels.width; col++)
323  {
324  if (img_edge (col, row).magnitude == 255.f)
325  labels[row * labels.width + col].label |= EDGELABEL_HIGH_CURVATURE;
326  }
327  }
328  }
329 }
330 
331 //////////////////////////////////////////////////////////////////////////////
332 template<typename PointT, typename PointNT, typename PointLT> void
333 pcl::OrganizedEdgeFromRGBNormals<PointT, PointNT, PointLT>::compute (pcl::PointCloud<PointLT>& labels, std::vector<pcl::PointIndices>& label_indices) const
334 {
335  pcl::Label invalid_pt;
336  invalid_pt.label = unsigned (0);
337  labels.points.resize (input_->points.size (), invalid_pt);
338  labels.width = input_->width;
339  labels.height = input_->height;
340 
344 
345  this->assignLabelIndices (labels, label_indices);
346 }
347 
348 #define PCL_INSTANTIATE_OrganizedEdgeBase(T,LT) template class PCL_EXPORTS pcl::OrganizedEdgeBase<T,LT>;
349 #define PCL_INSTANTIATE_OrganizedEdgeFromRGB(T,LT) template class PCL_EXPORTS pcl::OrganizedEdgeFromRGB<T,LT>;
350 #define PCL_INSTANTIATE_OrganizedEdgeFromNormals(T,NT,LT) template class PCL_EXPORTS pcl::OrganizedEdgeFromNormals<T,NT,LT>;
351 #define PCL_INSTANTIATE_OrganizedEdgeFromRGBNormals(T,NT,LT) template class PCL_EXPORTS pcl::OrganizedEdgeFromRGBNormals<T,NT,LT>;
352 
353 #endif //#ifndef PCL_FEATURES_IMPL_ORGANIZED_EDGE_DETECTION_H_
std::vector< PointT, Eigen::aligned_allocator< PointT > > points
The point data.
Definition: point_cloud.h:423
void extractEdges(pcl::PointCloud< PointLT > &labels) const
Perform the 3D edge detection (edges from depth discontinuities) and assign point indices for each ed...
void compute(pcl::PointCloud< PointLT > &labels, std::vector< pcl::PointIndices > &label_indices) const
Perform the 3D edge detection (edges from depth discontinuities, RGB Canny edge, and high curvature r...
void setHysteresisThresholdHigh(float threshold)
Definition: edge.h:154
Definition: edge.h:48
uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:428
void compute(pcl::PointCloud< PointLT > &labels, std::vector< pcl::PointIndices > &label_indices) const
Perform the 3D edge detection (edges from depth discontinuities)
void setInputCloud(PointCloudInPtr input)
Set the input point cloud pointer.
Definition: edge.h:303
uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:426
void extractEdges(pcl::PointCloud< PointLT > &labels) const
Perform the 3D edge detection (edges from depth discontinuities and high curvature regions) ...
uint32_t label
void setHysteresisThresholdLow(float threshold)
Definition: edge.h:148
boost::shared_ptr< PointCloud< PointT > > Ptr
Definition: point_cloud.h:441
void canny(const pcl::PointCloud< PointInT > &input_x, const pcl::PointCloud< PointInT > &input_y, pcl::PointCloud< PointOutT > &output)
Perform Canny edge detection with two separated input images for horizontal and vertical derivatives...
Definition: edge.hpp:383
PointCloud represents the base class in PCL for storing collections of 3D points. ...
OrganizedEdgeBase, OrganizedEdgeFromRGB, OrganizedEdgeFromNormals, and OrganizedEdgeFromRGBNormals fi...
void compute(pcl::PointCloud< PointLT > &labels, std::vector< pcl::PointIndices > &label_indices) const
Perform the 3D edge detection (edges from depth discontinuities and high curvature regions) and assig...
void extractEdges(pcl::PointCloud< PointLT > &labels) const
Perform the 3D edge detection (edges from depth discontinuities and RGB Canny edge) ...
void compute(pcl::PointCloud< PointLT > &labels, std::vector< pcl::PointIndices > &label_indices) const
Perform the 3D edge detection (edges from depth discontinuities and RGB Canny edge) and assign point ...
void resize(size_t n)
Resize the cloud.
Definition: point_cloud.h:468
void assignLabelIndices(pcl::PointCloud< PointLT > &labels, std::vector< pcl::PointIndices > &label_indices) const
Assign point indices for each edge label.
void detectEdgeCanny(pcl::PointCloud< PointOutT > &output)
All edges of magnitude above t_high are always classified as edges.
Definition: edge.hpp:318