Point Cloud Library (PCL)  1.14.1-dev
esf.hpp
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2010-2012, Willow Garage, Inc.
6  * Copyright (c) 2012-, Open Perception, Inc.
7  *
8  * All rights reserved.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  *
14  * * Redistributions of source code must retain the above copyright
15  * notice, this list of conditions and the following disclaimer.
16  * * Redistributions in binary form must reproduce the above
17  * copyright notice, this list of conditions and the following
18  * disclaimer in the documentation and/or other materials provided
19  * with the distribution.
20  * * Neither the name of the copyright holder(s) nor the names of its
21  * contributors may be used to endorse or promote products derived
22  * from this software without specific prior written permission.
23  *
24  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
25  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
26  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
27  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
28  * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
29  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
30  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
31  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
32  * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
34  * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
35  * POSSIBILITY OF SUCH DAMAGE.
36  *
37  * $Id: pfh.hpp 5027 2012-03-12 03:10:45Z rusu $
38  *
39  */
40 
41 #ifndef PCL_FEATURES_IMPL_ESF_H_
42 #define PCL_FEATURES_IMPL_ESF_H_
43 
44 #include <pcl/features/esf.h>
45 #include <pcl/common/distances.h>
46 #include <pcl/common/transforms.h>
47 #include <vector>
48 #include <ctime> // for time
49 
50 //////////////////////////////////////////////////////////////////////////////////////////////////////////////////
51 template <typename PointInT, typename PointOutT> void
52 pcl::ESFEstimation<PointInT, PointOutT>::computeESF (
53  PointCloudIn &pc, std::vector<float> &hist)
54 {
55  const int binsize = 64;
56  const std::size_t sample_size = 20000;
57  // @TODO: Replace with c++ stdlib uniform_random_generator
58  srand (static_cast<unsigned int> (time (nullptr)));
59  const auto maxindex = pc.size ();
60 
61  std::vector<float> d2v, d1v, d3v, wt_d3;
62  std::vector<int> wt_d2;
63  d1v.reserve (sample_size);
64  d2v.reserve (sample_size * 3);
65  d3v.reserve (sample_size);
66  wt_d2.reserve (sample_size * 3);
67  wt_d3.reserve (sample_size);
68 
69  float h_in[binsize] = {0.0f};
70  float h_out[binsize] = {0.0f};
71  float h_mix[binsize] = {0.0f};
72  float h_mix_ratio[binsize] = {0.0f};
73 
74  float h_a3_in[binsize] = {0.0f};
75  float h_a3_out[binsize] = {0.0f};
76  float h_a3_mix[binsize] = {0.0f};
77 
78  float h_d3_in[binsize] = {0.0f};
79  float h_d3_out[binsize] = {0.0f};
80  float h_d3_mix[binsize] = {0.0f};
81 
82  float ratio=0.0;
83  float pih = static_cast<float>(M_PI) / 2.0f;
84  float a,b,c,s;
85  int th1,th2,th3;
86  int vxlcnt = 0;
87  int pcnt1,pcnt2,pcnt3;
88  for (std::size_t nn_idx = 0; nn_idx < sample_size; ++nn_idx)
89  {
90  // get a new random point
91  int index1 = rand()%maxindex;
92  int index2 = rand()%maxindex;
93  int index3 = rand()%maxindex;
94 
95  if (index1==index2 || index1 == index3 || index2 == index3)
96  {
97  nn_idx--;
98  continue;
99  }
100 
101  Eigen::Vector4f p1 = pc[index1].getVector4fMap ();
102  Eigen::Vector4f p2 = pc[index2].getVector4fMap ();
103  Eigen::Vector4f p3 = pc[index3].getVector4fMap ();
104 
105  // A3
106  Eigen::Vector4f v21 (p2 - p1);
107  Eigen::Vector4f v31 (p3 - p1);
108  Eigen::Vector4f v23 (p2 - p3);
109  a = v21.norm (); b = v31.norm (); c = v23.norm (); s = (a+b+c) * 0.5f;
110  if (s * (s-a) * (s-b) * (s-c) <= 0.001f)
111  {
112  nn_idx--;
113  continue;
114  }
115 
116  v21.normalize ();
117  v31.normalize ();
118  v23.normalize ();
119 
120  //TODO: .dot gives nan's
121  th1 = static_cast<int> (pcl_round (std::acos (std::abs (v21.dot (v31))) / pih * (binsize-1)));
122  th2 = static_cast<int> (pcl_round (std::acos (std::abs (v23.dot (v31))) / pih * (binsize-1)));
123  th3 = static_cast<int> (pcl_round (std::acos (std::abs (v23.dot (v21))) / pih * (binsize-1)));
124  if (th1 < 0 || th1 >= binsize)
125  {
126  nn_idx--;
127  continue;
128  }
129  if (th2 < 0 || th2 >= binsize)
130  {
131  nn_idx--;
132  continue;
133  }
134  if (th3 < 0 || th3 >= binsize)
135  {
136  nn_idx--;
137  continue;
138  }
139 
140  // D2
141  d2v.push_back (pcl::euclideanDistance (pc[index1], pc[index2]));
142  d2v.push_back (pcl::euclideanDistance (pc[index1], pc[index3]));
143  d2v.push_back (pcl::euclideanDistance (pc[index2], pc[index3]));
144 
145  int vxlcnt_sum = 0;
146  int p_cnt = 0;
147  // IN, OUT, MIXED, Ratio line tracing, index1->index2
148  {
149  const int xs = p1[0] < 0.0? static_cast<int>(std::floor(p1[0])+GRIDSIZE_H): static_cast<int>(std::ceil(p1[0])+GRIDSIZE_H-1);
150  const int ys = p1[1] < 0.0? static_cast<int>(std::floor(p1[1])+GRIDSIZE_H): static_cast<int>(std::ceil(p1[1])+GRIDSIZE_H-1);
151  const int zs = p1[2] < 0.0? static_cast<int>(std::floor(p1[2])+GRIDSIZE_H): static_cast<int>(std::ceil(p1[2])+GRIDSIZE_H-1);
152  const int xt = p2[0] < 0.0? static_cast<int>(std::floor(p2[0])+GRIDSIZE_H): static_cast<int>(std::ceil(p2[0])+GRIDSIZE_H-1);
153  const int yt = p2[1] < 0.0? static_cast<int>(std::floor(p2[1])+GRIDSIZE_H): static_cast<int>(std::ceil(p2[1])+GRIDSIZE_H-1);
154  const int zt = p2[2] < 0.0? static_cast<int>(std::floor(p2[2])+GRIDSIZE_H): static_cast<int>(std::ceil(p2[2])+GRIDSIZE_H-1);
155  wt_d2.push_back (this->lci (xs, ys, zs, xt, yt, zt, ratio, vxlcnt, pcnt1));
156  if (wt_d2.back () == 2)
157  h_mix_ratio[static_cast<int> (pcl_round (ratio * (binsize-1)))]++;
158  vxlcnt_sum += vxlcnt;
159  p_cnt += pcnt1;
160  }
161  // IN, OUT, MIXED, Ratio line tracing, index1->index3
162  {
163  const int xs = p1[0] < 0.0? static_cast<int>(std::floor(p1[0])+GRIDSIZE_H): static_cast<int>(std::ceil(p1[0])+GRIDSIZE_H-1);
164  const int ys = p1[1] < 0.0? static_cast<int>(std::floor(p1[1])+GRIDSIZE_H): static_cast<int>(std::ceil(p1[1])+GRIDSIZE_H-1);
165  const int zs = p1[2] < 0.0? static_cast<int>(std::floor(p1[2])+GRIDSIZE_H): static_cast<int>(std::ceil(p1[2])+GRIDSIZE_H-1);
166  const int xt = p3[0] < 0.0? static_cast<int>(std::floor(p3[0])+GRIDSIZE_H): static_cast<int>(std::ceil(p3[0])+GRIDSIZE_H-1);
167  const int yt = p3[1] < 0.0? static_cast<int>(std::floor(p3[1])+GRIDSIZE_H): static_cast<int>(std::ceil(p3[1])+GRIDSIZE_H-1);
168  const int zt = p3[2] < 0.0? static_cast<int>(std::floor(p3[2])+GRIDSIZE_H): static_cast<int>(std::ceil(p3[2])+GRIDSIZE_H-1);
169  wt_d2.push_back (this->lci (xs, ys, zs, xt, yt, zt, ratio, vxlcnt, pcnt2));
170  if (wt_d2.back () == 2)
171  h_mix_ratio[static_cast<int>(pcl_round (ratio * (binsize-1)))]++;
172  vxlcnt_sum += vxlcnt;
173  p_cnt += pcnt2;
174  }
175  // IN, OUT, MIXED, Ratio line tracing, index2->index3
176  {
177  const int xs = p2[0] < 0.0? static_cast<int>(std::floor(p2[0])+GRIDSIZE_H): static_cast<int>(std::ceil(p2[0])+GRIDSIZE_H-1);
178  const int ys = p2[1] < 0.0? static_cast<int>(std::floor(p2[1])+GRIDSIZE_H): static_cast<int>(std::ceil(p2[1])+GRIDSIZE_H-1);
179  const int zs = p2[2] < 0.0? static_cast<int>(std::floor(p2[2])+GRIDSIZE_H): static_cast<int>(std::ceil(p2[2])+GRIDSIZE_H-1);
180  const int xt = p3[0] < 0.0? static_cast<int>(std::floor(p3[0])+GRIDSIZE_H): static_cast<int>(std::ceil(p3[0])+GRIDSIZE_H-1);
181  const int yt = p3[1] < 0.0? static_cast<int>(std::floor(p3[1])+GRIDSIZE_H): static_cast<int>(std::ceil(p3[1])+GRIDSIZE_H-1);
182  const int zt = p3[2] < 0.0? static_cast<int>(std::floor(p3[2])+GRIDSIZE_H): static_cast<int>(std::ceil(p3[2])+GRIDSIZE_H-1);
183  wt_d2.push_back (this->lci (xs,ys,zs,xt,yt,zt,ratio,vxlcnt,pcnt3));
184  if (wt_d2.back () == 2)
185  h_mix_ratio[static_cast<int>(pcl_round(ratio * (binsize-1)))]++;
186  vxlcnt_sum += vxlcnt;
187  p_cnt += pcnt3;
188  }
189 
190  // D3 ( herons formula )
191  d3v.push_back (std::sqrt (std::sqrt (s * (s-a) * (s-b) * (s-c))));
192  if (vxlcnt_sum <= 21)
193  {
194  wt_d3.push_back (0);
195  h_a3_out[th1] += static_cast<float> (pcnt3) / 32.0f;
196  h_a3_out[th2] += static_cast<float> (pcnt1) / 32.0f;
197  h_a3_out[th3] += static_cast<float> (pcnt2) / 32.0f;
198  }
199  else
200  if (p_cnt - vxlcnt_sum < 4)
201  {
202  h_a3_in[th1] += static_cast<float> (pcnt3) / 32.0f;
203  h_a3_in[th2] += static_cast<float> (pcnt1) / 32.0f;
204  h_a3_in[th3] += static_cast<float> (pcnt2) / 32.0f;
205  wt_d3.push_back (1);
206  }
207  else
208  {
209  h_a3_mix[th1] += static_cast<float> (pcnt3) / 32.0f;
210  h_a3_mix[th2] += static_cast<float> (pcnt1) / 32.0f;
211  h_a3_mix[th3] += static_cast<float> (pcnt2) / 32.0f;
212  wt_d3.push_back (static_cast<float> (vxlcnt_sum) / static_cast<float> (p_cnt));
213  }
214  }
215  // Normalizing, get max
216  float maxd2 = 0;
217  float maxd3 = 0;
218 
219  for (std::size_t nn_idx = 0; nn_idx < sample_size; ++nn_idx)
220  {
221  // get max of Dx
222  if (d2v[nn_idx] > maxd2)
223  maxd2 = d2v[nn_idx];
224  if (d2v[sample_size + nn_idx] > maxd2)
225  maxd2 = d2v[sample_size + nn_idx];
226  if (d2v[sample_size*2 +nn_idx] > maxd2)
227  maxd2 = d2v[sample_size*2 +nn_idx];
228  if (d3v[nn_idx] > maxd3)
229  maxd3 = d3v[nn_idx];
230  }
231 
232  // Normalize and create histogram
233  int index;
234  for (std::size_t nn_idx = 0; nn_idx < sample_size; ++nn_idx)
235  {
236  if (wt_d3[nn_idx] >= 0.999) // IN
237  {
238  index = static_cast<int>(pcl_round (d3v[nn_idx] / maxd3 * (binsize-1)));
239  if (index >= 0 && index < binsize)
240  h_d3_in[index]++;
241  }
242  else
243  {
244  if (wt_d3[nn_idx] <= 0.001) // OUT
245  {
246  index = static_cast<int>(pcl_round (d3v[nn_idx] / maxd3 * (binsize-1)));
247  if (index >= 0 && index < binsize)
248  h_d3_out[index]++ ;
249  }
250  else
251  {
252  index = static_cast<int>(pcl_round (d3v[nn_idx] / maxd3 * (binsize-1)));
253  if (index >= 0 && index < binsize)
254  h_d3_mix[index]++;
255  }
256  }
257  }
258  //normalize and create histogram
259  for (std::size_t nn_idx = 0; nn_idx < d2v.size(); ++nn_idx )
260  {
261  if (wt_d2[nn_idx] == 0)
262  h_in[static_cast<int>(pcl_round (d2v[nn_idx] / maxd2 * (binsize-1)))]++ ;
263  if (wt_d2[nn_idx] == 1)
264  h_out[static_cast<int>(pcl_round (d2v[nn_idx] / maxd2 * (binsize-1)))]++;
265  if (wt_d2[nn_idx] == 2)
266  h_mix[static_cast<int>(pcl_round (d2v[nn_idx] / maxd2 * (binsize-1)))]++ ;
267  }
268 
269  //float weights[10] = {1, 1, 1, 1, 1, 1, 1, 1 , 1 , 1};
270  float weights[10] = {0.5f, 0.5f, 0.5f, 0.5f, 0.5f, 1.0f, 1.0f, 2.0f, 2.0f, 2.0f};
271 
272  hist.reserve (static_cast<std::size_t>(binsize) * 10);
273  for (const float &i : h_a3_in)
274  hist.push_back (i * weights[0]);
275  for (const float &i : h_a3_out)
276  hist.push_back (i * weights[1]);
277  for (const float &i : h_a3_mix)
278  hist.push_back (i * weights[2]);
279 
280  for (const float &i : h_d3_in)
281  hist.push_back (i * weights[3]);
282  for (const float &i : h_d3_out)
283  hist.push_back (i * weights[4]);
284  for (const float &i : h_d3_mix)
285  hist.push_back (i * weights[5]);
286 
287  for (const float &i : h_in)
288  hist.push_back (i*0.5f * weights[6]);
289  for (const float &i : h_out)
290  hist.push_back (i * weights[7]);
291  for (const float &i : h_mix)
292  hist.push_back (i * weights[8]);
293  for (const float &i : h_mix_ratio)
294  hist.push_back (i*0.5f * weights[9]);
295 
296  float sm = 0;
297  for (const float &i : hist)
298  sm += i;
299 
300  for (float &i : hist)
301  i /= sm;
302 }
303 
304 //////////////////////////////////////////////////////////////////////////////////////////////////////////////////
305 template <typename PointInT, typename PointOutT> int
306 pcl::ESFEstimation<PointInT, PointOutT>::lci (
307  const int x1, const int y1, const int z1,
308  const int x2, const int y2, const int z2,
309  float &ratio, int &incnt, int &pointcount)
310 {
311  int voxelcount = 0;
312  int voxel_in = 0;
313  int act_voxel[3];
314  act_voxel[0] = x1;
315  act_voxel[1] = y1;
316  act_voxel[2] = z1;
317  int x_inc, y_inc, z_inc;
318  int dx = x2 - x1;
319  int dy = y2 - y1;
320  int dz = z2 - z1;
321  if (dx < 0)
322  x_inc = -1;
323  else
324  x_inc = 1;
325  int l = std::abs (dx);
326  if (dy < 0)
327  y_inc = -1 ;
328  else
329  y_inc = 1;
330  int m = std::abs (dy);
331  if (dz < 0)
332  z_inc = -1 ;
333  else
334  z_inc = 1;
335  int n = std::abs (dz);
336  int dx2 = 2 * l;
337  int dy2 = 2 * m;
338  int dz2 = 2 * n;
339  if ((l >= m) & (l >= n))
340  {
341  int err_1 = dy2 - l;
342  int err_2 = dz2 - l;
343  for (int i = 1; i<l; i++)
344  {
345  voxelcount++;
346  voxel_in += static_cast<int>(lut_[act_voxel[0]][act_voxel[1]][act_voxel[2]] == 1);
347  if (err_1 > 0)
348  {
349  act_voxel[1] += y_inc;
350  err_1 -= dx2;
351  }
352  if (err_2 > 0)
353  {
354  act_voxel[2] += z_inc;
355  err_2 -= dx2;
356  }
357  err_1 += dy2;
358  err_2 += dz2;
359  act_voxel[0] += x_inc;
360  }
361  }
362  else if ((m >= l) & (m >= n))
363  {
364  int err_1 = dx2 - m;
365  int err_2 = dz2 - m;
366  for (int i=1; i<m; i++)
367  {
368  voxelcount++;
369  voxel_in += static_cast<int>(lut_[act_voxel[0]][act_voxel[1]][act_voxel[2]] == 1);
370  if (err_1 > 0)
371  {
372  act_voxel[0] += x_inc;
373  err_1 -= dy2;
374  }
375  if (err_2 > 0)
376  {
377  act_voxel[2] += z_inc;
378  err_2 -= dy2;
379  }
380  err_1 += dx2;
381  err_2 += dz2;
382  act_voxel[1] += y_inc;
383  }
384  }
385  else
386  {
387  int err_1 = dy2 - n;
388  int err_2 = dx2 - n;
389  for (int i=1; i<n; i++)
390  {
391  voxelcount++;
392  voxel_in += static_cast<int>(lut_[act_voxel[0]][act_voxel[1]][act_voxel[2]] == 1);
393  if (err_1 > 0)
394  {
395  act_voxel[1] += y_inc;
396  err_1 -= dz2;
397  }
398  if (err_2 > 0)
399  {
400  act_voxel[0] += x_inc;
401  err_2 -= dz2;
402  }
403  err_1 += dy2;
404  err_2 += dx2;
405  act_voxel[2] += z_inc;
406  }
407  }
408  voxelcount++;
409  voxel_in += static_cast<int>(lut_[act_voxel[0]][act_voxel[1]][act_voxel[2]] == 1);
410  incnt = voxel_in;
411  pointcount = voxelcount;
412 
413  if (voxel_in >= voxelcount-1)
414  return (0);
415 
416  if (voxel_in <= 7)
417  return (1);
418 
419  ratio = static_cast<float>(voxel_in) / static_cast<float>(voxelcount);
420  return (2);
421 }
422 
423 //////////////////////////////////////////////////////////////////////////////////////////////////////////////////
424 template <typename PointInT, typename PointOutT> void
425 pcl::ESFEstimation<PointInT, PointOutT>::voxelize9 (PointCloudIn &cluster)
426 {
427  for (const auto& point: cluster)
428  {
429  int xx = point.x<0.0? static_cast<int>(std::floor(point.x)+GRIDSIZE_H) : static_cast<int>(std::ceil(point.x)+GRIDSIZE_H-1);
430  int yy = point.y<0.0? static_cast<int>(std::floor(point.y)+GRIDSIZE_H) : static_cast<int>(std::ceil(point.y)+GRIDSIZE_H-1);
431  int zz = point.z<0.0? static_cast<int>(std::floor(point.z)+GRIDSIZE_H) : static_cast<int>(std::ceil(point.z)+GRIDSIZE_H-1);
432 
433  for (int x = -1; x < 2; x++)
434  for (int y = -1; y < 2; y++)
435  for (int z = -1; z < 2; z++)
436  {
437  int xi = xx + x;
438  int yi = yy + y;
439  int zi = zz + z;
440 
441  if (yi >= GRIDSIZE || xi >= GRIDSIZE || zi>=GRIDSIZE || yi < 0 || xi < 0 || zi < 0)
442  {
443  ;
444  }
445  else
446  this->lut_[xi][yi][zi] = 1;
447  }
448  }
449 }
450 
451 //////////////////////////////////////////////////////////////////////////////////////////////////////////////////
452 template <typename PointInT, typename PointOutT> void
453 pcl::ESFEstimation<PointInT, PointOutT>::cleanup9 (PointCloudIn &cluster)
454 {
455  for (const auto& point: cluster)
456  {
457  int xx = point.x<0.0? static_cast<int>(std::floor(point.x)+GRIDSIZE_H) : static_cast<int>(std::ceil(point.x)+GRIDSIZE_H-1);
458  int yy = point.y<0.0? static_cast<int>(std::floor(point.y)+GRIDSIZE_H) : static_cast<int>(std::ceil(point.y)+GRIDSIZE_H-1);
459  int zz = point.z<0.0? static_cast<int>(std::floor(point.z)+GRIDSIZE_H) : static_cast<int>(std::ceil(point.z)+GRIDSIZE_H-1);
460 
461  for (int x = -1; x < 2; x++)
462  for (int y = -1; y < 2; y++)
463  for (int z = -1; z < 2; z++)
464  {
465  int xi = xx + x;
466  int yi = yy + y;
467  int zi = zz + z;
468 
469  if (yi >= GRIDSIZE || xi >= GRIDSIZE || zi>=GRIDSIZE || yi < 0 || xi < 0 || zi < 0)
470  {
471  ;
472  }
473  else
474  this->lut_[xi][yi][zi] = 0;
475  }
476  }
477 }
478 
479 //////////////////////////////////////////////////////////////////////////////////////////////////////////////////
480 template <typename PointInT, typename PointOutT> void
481 pcl::ESFEstimation<PointInT, PointOutT>::scale_points_unit_sphere (
482  const pcl::PointCloud<PointInT> &pc, float scalefactor, Eigen::Vector4f& centroid)
483 {
484  pcl::compute3DCentroid (pc, centroid);
485  pcl::demeanPointCloud (pc, centroid, local_cloud_);
486 
487  float max_distance = 0;
488  pcl::PointXYZ cog (0, 0, 0);
489 
490  for (const auto& point: local_cloud_)
491  {
492  float d = pcl::euclideanDistance(cog,point);
493  if (d > max_distance)
494  max_distance = d;
495  }
496 
497  float scale_factor = 1.0f / max_distance * scalefactor;
498 
499  Eigen::Affine3f matrix = Eigen::Affine3f::Identity();
500  matrix.scale (scale_factor);
501  pcl::transformPointCloud (local_cloud_, local_cloud_, matrix);
502 }
503 
504 //////////////////////////////////////////////////////////////////////////////////////////////
505 template<typename PointInT, typename PointOutT> void
506 pcl::ESFEstimation<PointInT, PointOutT>::compute (PointCloudOut &output)
507 {
508  if (!Feature<PointInT, PointOutT>::initCompute ())
509  {
510  output.width = output.height = 0;
511  output.clear ();
512  return;
513  }
514  // Copy the header
515  output.header = input_->header;
516 
517  // Resize the output dataset
518  // Important! We should only allocate precisely how many elements we will need, otherwise
519  // we risk at pre-allocating too much memory which could lead to bad_alloc
520  // (see http://dev.pointclouds.org/issues/657)
521  output.width = output.height = 1;
522  output.is_dense = input_->is_dense;
523  output.resize (1);
524 
525  // Perform the actual feature computation
526  computeFeature (output);
527 
528  Feature<PointInT, PointOutT>::deinitCompute ();
529 }
530 
531 
532 //////////////////////////////////////////////////////////////////////////////////////////////
533 template <typename PointInT, typename PointOutT> void
534 pcl::ESFEstimation<PointInT, PointOutT>::computeFeature (PointCloudOut &output)
535 {
536  Eigen::Vector4f xyz_centroid;
537  std::vector<float> hist;
538  scale_points_unit_sphere (*surface_, static_cast<float>(GRIDSIZE_H), xyz_centroid);
539  this->voxelize9 (local_cloud_);
540  this->computeESF (local_cloud_, hist);
541  this->cleanup9 (local_cloud_);
542 
543  // We only output _1_ signature
544  output.resize (1);
545  output.width = 1;
546  output.height = 1;
547 
548  for (std::size_t d = 0; d < hist.size (); ++d)
549  output[0].histogram[d] = hist[d];
550 }
551 
552 #define PCL_INSTANTIATE_ESFEstimation(T,OutT) template class PCL_EXPORTS pcl::ESFEstimation<T,OutT>;
553 
554 #endif // PCL_FEATURES_IMPL_ESF_H_
555 
pcl::ESFEstimation::compute
void compute(PointCloudOut &output)
Overloaded computed method from pcl::Feature.
Definition: esf.hpp:506
pcl::ESFEstimation::computeESF
void computeESF(PointCloudIn &pc, std::vector< float > &hist)
...
Definition: esf.hpp:52
pcl::ESFEstimation::PointCloudOut
typename Feature< PointInT, PointOutT >::PointCloudOut PointCloudOut
Definition: esf.h:74
pcl::ESFEstimation::voxelize9
void voxelize9(PointCloudIn &cluster)
...
Definition: esf.hpp:425
pcl::ESFEstimation::scale_points_unit_sphere
void scale_points_unit_sphere(const pcl::PointCloud< PointInT > &pc, float scalefactor, Eigen::Vector4f &centroid)
...
Definition: esf.hpp:481
pcl::ESFEstimation::lci
int lci(const int x1, const int y1, const int z1, const int x2, const int y2, const int z2, float &ratio, int &incnt, int &pointcount)
...
Definition: esf.hpp:306
pcl::ESFEstimation::cleanup9
void cleanup9(PointCloudIn &cluster)
...
Definition: esf.hpp:453
pcl::ESFEstimation::computeFeature
void computeFeature(PointCloudOut &output) override
Estimate the Ensebmel of Shape Function (ESF) descriptors at a set of points given by <setInputCloud ...
Definition: esf.hpp:534
pcl::Feature
Feature represents the base feature class.
Definition: feature.h:107
pcl::PointCloud::size
std::size_t size() const
Definition: point_cloud.h:443
distances.h
Define standard C methods to do distance calculations.
pcl::demeanPointCloud
void demeanPointCloud(ConstCloudIterator< PointT > &cloud_iterator, const Eigen::Matrix< Scalar, 4, 1 > &centroid, pcl::PointCloud< PointT > &cloud_out, int npts=0)
Subtract a centroid from a point cloud and return the de-meaned representation.
Definition: centroid.hpp:933
pcl::transformPointCloud
void transformPointCloud(const pcl::PointCloud< PointT > &cloud_in, pcl::PointCloud< PointT > &cloud_out, const Eigen::Matrix< Scalar, 4, 4 > &transform, bool copy_all_fields)
Apply a rigid transform defined by a 4x4 matrix.
Definition: transforms.hpp:221
pcl::compute3DCentroid
unsigned int compute3DCentroid(ConstCloudIterator< PointT > &cloud_iterator, Eigen::Matrix< Scalar, 4, 1 > &centroid)
Compute the 3D (X-Y-Z) centroid of a set of points and return it as a 3D vector.
Definition: centroid.hpp:57
pcl::euclideanDistance
float euclideanDistance(const PointType1 &p1, const PointType2 &p2)
Calculate the euclidean distance between the two given points.
Definition: distances.h:204
pcl_round
__inline double pcl_round(double number)
Win32 doesn't seem to have rounding functions.
Definition: pcl_macros.h:239
M_PI
#define M_PI
Definition: pcl_macros.h:201
pcl::PointXYZ
A point structure representing Euclidean xyz coordinates.
Definition: point_types.hpp:322