Point Cloud Library (PCL)  1.11.1-dev
pyramid.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  * $Id$
37  *
38  */
39 
40 
41 #ifndef PCL_FILTERS_IMPL_PYRAMID_HPP
42 #define PCL_FILTERS_IMPL_PYRAMID_HPP
43 
44 
45 namespace pcl
46 {
47 
48 namespace filters
49 {
50 
51 template <typename PointT> bool
52 Pyramid<PointT>::initCompute ()
53 {
54  if (!input_->isOrganized ())
55  {
56  PCL_ERROR ("[pcl::fileters::%s::initCompute] Number of levels should be at least 2!\n", getClassName ().c_str ());
57  return (false);
58  }
59 
60  if (levels_ < 2)
61  {
62  PCL_ERROR ("[pcl::fileters::%s::initCompute] Number of levels should be at least 2!\n", getClassName ().c_str ());
63  return (false);
64  }
65 
66  // std::size_t ratio (std::pow (2, levels_));
67  // std::size_t last_width = input_->width / ratio;
68  // std::size_t last_height = input_->height / ratio;
69 
70  if (levels_ > 4)
71  {
72  PCL_ERROR ("[pcl::fileters::%s::initCompute] Number of levels should not exceed 4!\n", getClassName ().c_str ());
73  return (false);
74  }
75 
76  if (large_)
77  {
78  Eigen::VectorXf k (5);
79  k << 1.f/16.f, 1.f/4.f, 3.f/8.f, 1.f/4.f, 1.f/16.f;
80  kernel_ = k * k.transpose ();
81  if (threshold_ != std::numeric_limits<float>::infinity ())
82  threshold_ *= 2 * threshold_;
83 
84  }
85  else
86  {
87  Eigen::VectorXf k (3);
88  k << 1.f/4.f, 1.f/2.f, 1.f/4.f;
89  kernel_ = k * k.transpose ();
90  if (threshold_ != std::numeric_limits<float>::infinity ())
91  threshold_ *= threshold_;
92  }
93 
94  return (true);
95 }
96 
97 template <typename PointT> void
98 Pyramid<PointT>::compute (std::vector<PointCloudPtr>& output)
99 {
100  std::cout << "compute" << std::endl;
101  if (!initCompute ())
102  {
103  PCL_ERROR ("[pcl::%s::compute] initCompute failed!\n", getClassName ().c_str ());
104  return;
105  }
106 
107  int kernel_rows = static_cast<int> (kernel_.rows ());
108  int kernel_cols = static_cast<int> (kernel_.cols ());
109  int kernel_center_x = kernel_cols / 2;
110  int kernel_center_y = kernel_rows / 2;
111 
112  output.resize (levels_ + 1);
113  output[0].reset (new pcl::PointCloud<PointT>);
114  *(output[0]) = *input_;
115 
116  if (input_->is_dense)
117  {
118  for (int l = 1; l <= levels_; ++l)
119  {
120  output[l].reset (new pcl::PointCloud<PointT> (output[l-1]->width/2, output[l-1]->height/2));
121  const PointCloud<PointT> &previous = *output[l-1];
122  PointCloud<PointT> &next = *output[l];
123 #pragma omp parallel for \
124  default(none) \
125  shared(next) \
126  num_threads(threads_)
127  for(int i=0; i < next.height; ++i)
128  {
129  for(int j=0; j < next.width; ++j)
130  {
131  for(int m=0; m < kernel_rows; ++m)
132  {
133  int mm = kernel_rows - 1 - m;
134  for(int n=0; n < kernel_cols; ++n)
135  {
136  int nn = kernel_cols - 1 - n;
137 
138  int ii = 2*i + (m - kernel_center_y);
139  int jj = 2*j + (n - kernel_center_x);
140 
141  if (ii < 0) ii = 0;
142  if (ii >= previous.height) ii = previous.height - 1;
143  if (jj < 0) jj = 0;
144  if (jj >= previous.width) jj = previous.width - 1;
145  next.at (j,i) += previous.at (jj,ii) * kernel_ (mm,nn);
146  }
147  }
148  }
149  }
150  }
151  }
152  else
153  {
154  for (int l = 1; l <= levels_; ++l)
155  {
156  output[l].reset (new pcl::PointCloud<PointT> (output[l-1]->width/2, output[l-1]->height/2));
157  const PointCloud<PointT> &previous = *output[l-1];
158  PointCloud<PointT> &next = *output[l];
159 #pragma omp parallel for \
160  default(none) \
161  shared(next) \
162  num_threads(threads_)
163  for(int i=0; i < next.height; ++i)
164  {
165  for(int j=0; j < next.width; ++j)
166  {
167  float weight = 0;
168  for(int m=0; m < kernel_rows; ++m)
169  {
170  int mm = kernel_rows - 1 - m;
171  for(int n=0; n < kernel_cols; ++n)
172  {
173  int nn = kernel_cols - 1 - n;
174  int ii = 2*i + (m - kernel_center_y);
175  int jj = 2*j + (n - kernel_center_x);
176  if (ii < 0) ii = 0;
177  if (ii >= previous.height) ii = previous.height - 1;
178  if (jj < 0) jj = 0;
179  if (jj >= previous.width) jj = previous.width - 1;
180  if (!isFinite (previous.at (jj,ii)))
181  continue;
182  if (pcl::squaredEuclideanDistance (previous.at (2*j,2*i), previous.at (jj,ii)) < threshold_)
183  {
184  next.at (j,i) += previous.at (jj,ii).x * kernel_ (mm,nn);
185  weight+= kernel_ (mm,nn);
186  }
187  }
188  }
189  if (weight == 0)
190  nullify (next.at (j,i));
191  else
192  {
193  weight = 1.f/weight;
194  next.at (j,i)*= weight;
195  }
196  }
197  }
198  }
199  }
200 }
201 
202 
203 template <> void
204 Pyramid<pcl::PointXYZRGB>::compute (std::vector<Pyramid<pcl::PointXYZRGB>::PointCloudPtr> &output)
205 {
206  std::cout << "PointXYZRGB" << std::endl;
207  if (!initCompute ())
208  {
209  PCL_ERROR ("[pcl::%s::compute] initCompute failed!\n", getClassName ().c_str ());
210  return;
211  }
212 
213  int kernel_rows = static_cast<int> (kernel_.rows ());
214  int kernel_cols = static_cast<int> (kernel_.cols ());
215  int kernel_center_x = kernel_cols / 2;
216  int kernel_center_y = kernel_rows / 2;
217 
218  output.resize (levels_ + 1);
219  output[0].reset (new pcl::PointCloud<pcl::PointXYZRGB>);
220  *(output[0]) = *input_;
221 
222  if (input_->is_dense)
223  {
224  for (int l = 1; l <= levels_; ++l)
225  {
226  output[l].reset (new pcl::PointCloud<pcl::PointXYZRGB> (output[l-1]->width/2, output[l-1]->height/2));
227  const PointCloud<pcl::PointXYZRGB> &previous = *output[l-1];
228  PointCloud<pcl::PointXYZRGB> &next = *output[l];
229 #pragma omp parallel for \
230  default(none) \
231  shared(next) \
232  num_threads(threads_)
233  for(int i=0; i < next.height; ++i) // rows
234  {
235  for(int j=0; j < next.width; ++j) // columns
236  {
237  float r = 0, g = 0, b = 0;
238  for(int m=0; m < kernel_rows; ++m) // kernel rows
239  {
240  int mm = kernel_rows - 1 - m; // row index of flipped kernel
241  for(int n=0; n < kernel_cols; ++n) // kernel columns
242  {
243  int nn = kernel_cols - 1 - n; // column index of flipped kernel
244  // index of input signal, used for checking boundary
245  int ii = 2*i + (m - kernel_center_y);
246  int jj = 2*j + (n - kernel_center_x);
247 
248  // ignore input samples which are out of bound
249  if (ii < 0) ii = 0;
250  if (ii >= previous.height) ii = previous.height - 1;
251  if (jj < 0) jj = 0;
252  if (jj >= previous.width) jj = previous.width - 1;
253  next.at (j,i).x += previous.at (jj,ii).x * kernel_ (mm,nn);
254  next.at (j,i).y += previous.at (jj,ii).y * kernel_ (mm,nn);
255  next.at (j,i).z += previous.at (jj,ii).z * kernel_ (mm,nn);
256  b += previous.at (jj,ii).b * kernel_ (mm,nn);
257  g += previous.at (jj,ii).g * kernel_ (mm,nn);
258  r += previous.at (jj,ii).r * kernel_ (mm,nn);
259  }
260  }
261  next.at (j,i).b = static_cast<std::uint8_t> (b);
262  next.at (j,i).g = static_cast<std::uint8_t> (g);
263  next.at (j,i).r = static_cast<std::uint8_t> (r);
264  }
265  }
266  }
267  }
268  else
269  {
270  for (int l = 1; l <= levels_; ++l)
271  {
272  output[l].reset (new pcl::PointCloud<pcl::PointXYZRGB> (output[l-1]->width/2, output[l-1]->height/2));
273  const PointCloud<pcl::PointXYZRGB> &previous = *output[l-1];
274  PointCloud<pcl::PointXYZRGB> &next = *output[l];
275 #pragma omp parallel for \
276  default(none) \
277  shared(next) \
278  num_threads(threads_)
279  for(int i=0; i < next.height; ++i)
280  {
281  for(int j=0; j < next.width; ++j)
282  {
283  float weight = 0;
284  float r = 0, g = 0, b = 0;
285  for(int m=0; m < kernel_rows; ++m)
286  {
287  int mm = kernel_rows - 1 - m;
288  for(int n=0; n < kernel_cols; ++n)
289  {
290  int nn = kernel_cols - 1 - n;
291  int ii = 2*i + (m - kernel_center_y);
292  int jj = 2*j + (n - kernel_center_x);
293  if (ii < 0) ii = 0;
294  if (ii >= previous.height) ii = previous.height - 1;
295  if (jj < 0) jj = 0;
296  if (jj >= previous.width) jj = previous.width - 1;
297  if (!isFinite (previous.at (jj,ii)))
298  continue;
299  if (pcl::squaredEuclideanDistance (previous.at (2*j,2*i), previous.at (jj,ii)) < threshold_)
300  {
301  next.at (j,i).x += previous.at (jj,ii).x * kernel_ (mm,nn);
302  next.at (j,i).y += previous.at (jj,ii).y * kernel_ (mm,nn);
303  next.at (j,i).z += previous.at (jj,ii).z * kernel_ (mm,nn);
304  b += previous.at (jj,ii).b * kernel_ (mm,nn);
305  g += previous.at (jj,ii).g * kernel_ (mm,nn);
306  r += previous.at (jj,ii).r * kernel_ (mm,nn);
307  weight+= kernel_ (mm,nn);
308  }
309  }
310  }
311  if (weight == 0)
312  nullify (next.at (j,i));
313  else
314  {
315  weight = 1.f/weight;
316  r*= weight; g*= weight; b*= weight;
317  next.at (j,i).x*= weight; next.at (j,i).y*= weight; next.at (j,i).z*= weight;
318  next.at (j,i).b = static_cast<std::uint8_t> (b);
319  next.at (j,i).g = static_cast<std::uint8_t> (g);
320  next.at (j,i).r = static_cast<std::uint8_t> (r);
321  }
322  }
323  }
324  }
325  }
326 }
327 
328 template <> void
329 Pyramid<pcl::PointXYZRGBA>::compute (std::vector<Pyramid<pcl::PointXYZRGBA>::PointCloudPtr> &output)
330 {
331  std::cout << "PointXYZRGBA" << std::endl;
332  if (!initCompute ())
333  {
334  PCL_ERROR ("[pcl::%s::compute] initCompute failed!\n", getClassName ().c_str ());
335  return;
336  }
337 
338  int kernel_rows = static_cast<int> (kernel_.rows ());
339  int kernel_cols = static_cast<int> (kernel_.cols ());
340  int kernel_center_x = kernel_cols / 2;
341  int kernel_center_y = kernel_rows / 2;
342 
343  output.resize (levels_ + 1);
344  output[0].reset (new pcl::PointCloud<pcl::PointXYZRGBA>);
345  *(output[0]) = *input_;
346 
347  if (input_->is_dense)
348  {
349  for (int l = 1; l <= levels_; ++l)
350  {
351  output[l].reset (new pcl::PointCloud<pcl::PointXYZRGBA> (output[l-1]->width/2, output[l-1]->height/2));
352  const PointCloud<pcl::PointXYZRGBA> &previous = *output[l-1];
353  PointCloud<pcl::PointXYZRGBA> &next = *output[l];
354 #pragma omp parallel for \
355  default(none) \
356  shared(next) \
357  num_threads(threads_)
358  for(int i=0; i < next.height; ++i) // rows
359  {
360  for(int j=0; j < next.width; ++j) // columns
361  {
362  float r = 0, g = 0, b = 0, a = 0;
363  for(int m=0; m < kernel_rows; ++m) // kernel rows
364  {
365  int mm = kernel_rows - 1 - m; // row index of flipped kernel
366  for(int n=0; n < kernel_cols; ++n) // kernel columns
367  {
368  int nn = kernel_cols - 1 - n; // column index of flipped kernel
369  // index of input signal, used for checking boundary
370  int ii = 2*i + (m - kernel_center_y);
371  int jj = 2*j + (n - kernel_center_x);
372 
373  // ignore input samples which are out of bound
374  if (ii < 0) ii = 0;
375  if (ii >= previous.height) ii = previous.height - 1;
376  if (jj < 0) jj = 0;
377  if (jj >= previous.width) jj = previous.width - 1;
378  next.at (j,i).x += previous.at (jj,ii).x * kernel_ (mm,nn);
379  next.at (j,i).y += previous.at (jj,ii).y * kernel_ (mm,nn);
380  next.at (j,i).z += previous.at (jj,ii).z * kernel_ (mm,nn);
381  b += previous.at (jj,ii).b * kernel_ (mm,nn);
382  g += previous.at (jj,ii).g * kernel_ (mm,nn);
383  r += previous.at (jj,ii).r * kernel_ (mm,nn);
384  a += previous.at (jj,ii).a * kernel_ (mm,nn);
385  }
386  }
387  next.at (j,i).b = static_cast<std::uint8_t> (b);
388  next.at (j,i).g = static_cast<std::uint8_t> (g);
389  next.at (j,i).r = static_cast<std::uint8_t> (r);
390  next.at (j,i).a = static_cast<std::uint8_t> (a);
391  }
392  }
393  }
394  }
395  else
396  {
397  for (int l = 1; l <= levels_; ++l)
398  {
399  output[l].reset (new pcl::PointCloud<pcl::PointXYZRGBA> (output[l-1]->width/2, output[l-1]->height/2));
400  const PointCloud<pcl::PointXYZRGBA> &previous = *output[l-1];
401  PointCloud<pcl::PointXYZRGBA> &next = *output[l];
402 #pragma omp parallel for \
403  default(none) \
404  shared(next) \
405  num_threads(threads_)
406  for(int i=0; i < next.height; ++i)
407  {
408  for(int j=0; j < next.width; ++j)
409  {
410  float weight = 0;
411  float r = 0, g = 0, b = 0, a = 0;
412  for(int m=0; m < kernel_rows; ++m)
413  {
414  int mm = kernel_rows - 1 - m;
415  for(int n=0; n < kernel_cols; ++n)
416  {
417  int nn = kernel_cols - 1 - n;
418  int ii = 2*i + (m - kernel_center_y);
419  int jj = 2*j + (n - kernel_center_x);
420  if (ii < 0) ii = 0;
421  if (ii >= previous.height) ii = previous.height - 1;
422  if (jj < 0) jj = 0;
423  if (jj >= previous.width) jj = previous.width - 1;
424  if (!isFinite (previous.at (jj,ii)))
425  continue;
426  if (pcl::squaredEuclideanDistance (previous.at (2*j,2*i), previous.at (jj,ii)) < threshold_)
427  {
428  next.at (j,i).x += previous.at (jj,ii).x * kernel_ (mm,nn);
429  next.at (j,i).y += previous.at (jj,ii).y * kernel_ (mm,nn);
430  next.at (j,i).z += previous.at (jj,ii).z * kernel_ (mm,nn);
431  b += previous.at (jj,ii).b * kernel_ (mm,nn);
432  g += previous.at (jj,ii).g * kernel_ (mm,nn);
433  r += previous.at (jj,ii).r * kernel_ (mm,nn);
434  a += previous.at (jj,ii).a * kernel_ (mm,nn);
435  weight+= kernel_ (mm,nn);
436  }
437  }
438  }
439  if (weight == 0)
440  nullify (next.at (j,i));
441  else
442  {
443  weight = 1.f/weight;
444  r*= weight; g*= weight; b*= weight; a*= weight;
445  next.at (j,i).x*= weight; next.at (j,i).y*= weight; next.at (j,i).z*= weight;
446  next.at (j,i).b = static_cast<std::uint8_t> (b);
447  next.at (j,i).g = static_cast<std::uint8_t> (g);
448  next.at (j,i).r = static_cast<std::uint8_t> (r);
449  next.at (j,i).a = static_cast<std::uint8_t> (a);
450  }
451  }
452  }
453  }
454  }
455 }
456 
457 template<> void
458 Pyramid<pcl::RGB>::nullify (pcl::RGB& p)
459 {
460  p.r = 0; p.g = 0; p.b = 0;
461 }
462 
463 template <> void
464 Pyramid<pcl::RGB>::compute (std::vector<Pyramid<pcl::RGB>::PointCloudPtr> &output)
465 {
466  std::cout << "RGB" << std::endl;
467  if (!initCompute ())
468  {
469  PCL_ERROR ("[pcl::%s::compute] initCompute failed!\n", getClassName ().c_str ());
470  return;
471  }
472 
473  int kernel_rows = static_cast<int> (kernel_.rows ());
474  int kernel_cols = static_cast<int> (kernel_.cols ());
475  int kernel_center_x = kernel_cols / 2;
476  int kernel_center_y = kernel_rows / 2;
477 
478  output.resize (levels_ + 1);
479  output[0].reset (new pcl::PointCloud<pcl::RGB>);
480  *(output[0]) = *input_;
481 
482  if (input_->is_dense)
483  {
484  for (int l = 1; l <= levels_; ++l)
485  {
486  output[l].reset (new pcl::PointCloud<pcl::RGB> (output[l-1]->width/2, output[l-1]->height/2));
487  const PointCloud<pcl::RGB> &previous = *output[l-1];
488  PointCloud<pcl::RGB> &next = *output[l];
489 #pragma omp parallel for \
490  default(none) \
491  shared(next) \
492  num_threads(threads_)
493  for(int i=0; i < next.height; ++i)
494  {
495  for(int j=0; j < next.width; ++j)
496  {
497  float r = 0, g = 0, b = 0;
498  for(int m=0; m < kernel_rows; ++m)
499  {
500  int mm = kernel_rows - 1 - m;
501  for(int n=0; n < kernel_cols; ++n)
502  {
503  int nn = kernel_cols - 1 - n;
504  int ii = 2*i + (m - kernel_center_y);
505  int jj = 2*j + (n - kernel_center_x);
506  if (ii < 0) ii = 0;
507  if (ii >= previous.height) ii = previous.height - 1;
508  if (jj < 0) jj = 0;
509  if (jj >= previous.width) jj = previous.width - 1;
510  b += previous.at (jj,ii).b * kernel_ (mm,nn);
511  g += previous.at (jj,ii).g * kernel_ (mm,nn);
512  r += previous.at (jj,ii).r * kernel_ (mm,nn);
513  }
514  }
515  next.at (j,i).b = static_cast<std::uint8_t> (b);
516  next.at (j,i).g = static_cast<std::uint8_t> (g);
517  next.at (j,i).r = static_cast<std::uint8_t> (r);
518  }
519  }
520  }
521  }
522  else
523  {
524  for (int l = 1; l <= levels_; ++l)
525  {
526  output[l].reset (new pcl::PointCloud<pcl::RGB> (output[l-1]->width/2, output[l-1]->height/2));
527  const PointCloud<pcl::RGB> &previous = *output[l-1];
528  PointCloud<pcl::RGB> &next = *output[l];
529 #pragma omp parallel for \
530  default(none) \
531  shared(next) \
532  num_threads(threads_)
533  for(int i=0; i < next.height; ++i)
534  {
535  for(int j=0; j < next.width; ++j)
536  {
537  float weight = 0;
538  float r = 0, g = 0, b = 0;
539  for(int m=0; m < kernel_rows; ++m)
540  {
541  int mm = kernel_rows - 1 - m;
542  for(int n=0; n < kernel_cols; ++n)
543  {
544  int nn = kernel_cols - 1 - n;
545  int ii = 2*i + (m - kernel_center_y);
546  int jj = 2*j + (n - kernel_center_x);
547  if (ii < 0) ii = 0;
548  if (ii >= previous.height) ii = previous.height - 1;
549  if (jj < 0) jj = 0;
550  if (jj >= previous.width) jj = previous.width - 1;
551  if (!isFinite (previous.at (jj,ii)))
552  continue;
553  if (pcl::squaredEuclideanDistance (previous.at (2*j,2*i), previous.at (jj,ii)) < threshold_)
554  {
555  b += previous.at (jj,ii).b * kernel_ (mm,nn);
556  g += previous.at (jj,ii).g * kernel_ (mm,nn);
557  r += previous.at (jj,ii).r * kernel_ (mm,nn);
558  weight+= kernel_ (mm,nn);
559  }
560  }
561  }
562  if (weight == 0)
563  nullify (next.at (j,i));
564  else
565  {
566  weight = 1.f/weight;
567  r*= weight; g*= weight; b*= weight;
568  next.at (j,i).b = static_cast<std::uint8_t> (b);
569  next.at (j,i).g = static_cast<std::uint8_t> (g);
570  next.at (j,i).r = static_cast<std::uint8_t> (r);
571  }
572  }
573  }
574  }
575  }
576 }
577 
578 } // namespace filters
579 } // namespace pcl
580 
581 #endif
582 
pcl::filters::Pyramid::compute
void compute(std::vector< PointCloudPtr > &output)
compute the pyramid levels.
Definition: pyramid.hpp:98
pcl
Definition: convolution.h:46
pcl::PointCloud::height
std::uint32_t height
The point cloud height (if organized as an image-structure).
Definition: point_cloud.h:393
pcl::isFinite
bool isFinite(const PointT &pt)
Tests if the 3D components of a point are all finite param[in] pt point to be tested return true if f...
Definition: point_tests.h:55
pcl::PointCloud
PointCloud represents the base class in PCL for storing collections of 3D points.
Definition: distances.h:55
pcl::PointCloud::at
const PointT & at(int column, int row) const
Obtain the point given by the (column, row) coordinates.
Definition: point_cloud.h:261
pcl::PointCloud::width
std::uint32_t width
The point cloud width (if organized as an image-structure).
Definition: point_cloud.h:391
pcl::squaredEuclideanDistance
float squaredEuclideanDistance(const PointType1 &p1, const PointType2 &p2)
Calculate the squared euclidean distance between the two given points.
Definition: distances.h:182
pcl::RGB
A structure representing RGB color information.
Definition: point_types.hpp:345
pcl::filters::Pyramid
Pyramid constructs a multi-scale representation of an organised point cloud.
Definition: pyramid.h:62
pcl::filters::Pyramid::PointCloudPtr
typename PointCloud< PointT >::Ptr PointCloudPtr
Definition: pyramid.h:65