-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtgs.hpp
1337 lines (1109 loc) · 58.8 KB
/
tgs.hpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#ifndef TRAVEL_GSEG_H
#define TRAVEL_GSEG_H
//
// Created by Minho Oh & Euigon Jung on 1/31/22.
// We really appreciate Hyungtae Lim and Prof. Hyun Myung! :)
//
#include <iostream>
#include <vector>
#include <random>
#include <mutex>
#include <thread>
#include <chrono>
#include <math.h>
#include <queue>
#include <Eigen/Dense>
#include <pcl/filters/filter.h>
#include <pcl/common/centroid.h>
#include <pcl/filters/voxel_grid.h>
// #include <opencv2/opencv.hpp>
namespace travel {
#define PTCLOUD_SIZE 132000
#define NODEWISE_PTCLOUDSIZE 5000
#define UNKNOWN 1
#define NONGROUND 2
#define GROUND 3
using Eigen::MatrixXf;
using Eigen::JacobiSVD;
using Eigen::VectorXf;
template <typename PointType>
bool point_z_cmp(PointType a, PointType b){
return a.z<b.z;
}
struct TriGridIdx {
int row, col, tri;
};
struct TriGridEdge {
std::pair<TriGridIdx, TriGridIdx> Pair;
bool is_traversable;
};
template <typename PointType>
struct TriGridNode {
int node_type;
pcl::PointCloud<PointType> ptCloud;
bool is_curr_data;
// planar model
Eigen::Vector3f normal;
Eigen::Vector3f mean_pt;
double d;
Eigen::Vector3f singular_values;
Eigen::Matrix3f eigen_vectors;
double weight;
double th_dist_d;
double th_outlier_d;
// graph_searching
bool need_recheck;
bool is_visited;
bool is_rejection;
int check_life;
int depth;
};
struct TriGridCorner {
double x, y;
std::vector<double> zs;
std::vector<double> weights;
};
template <typename PointType>
using GridNode = std::vector<TriGridNode<PointType>>;
template <typename PointType>
using TriGridField = std::vector<std::vector<GridNode<PointType>>>;
template <typename PointType>
class TravelGroundSeg{
private:
bool REFINE_MODE_;
bool VIZ_MDOE_;
double MAX_RANGE_;
double MIN_RANGE_;
double TGF_RESOLUTION_;
int NUM_ITER_;
int NUM_LRP_;
int NUM_MIN_POINTS_;
double TH_SEEDS_;
double TH_DIST_;
double TH_OUTLIER_;
double TH_NORMAL_;
double TH_WEIGHT_;
double TH_LCC_NORMAL_SIMILARITY_;
double TH_LCC_PLANAR_MODEL_DIST_;
double TH_OBSTACLE_HEIGHT_;
TriGridField<PointType> trigrid_field_;
std::vector<TriGridEdge> trigrid_edges_;
std::vector<std::vector<TriGridCorner>> trigrid_corners_;
std::vector<std::vector<TriGridCorner>> trigrid_centers_;
pcl::PointCloud<PointType> empty_cloud_;
TriGridNode<PointType> empty_trigrid_node_;
GridNode<PointType> empty_grid_nodes_;
TriGridCorner empty_trigrid_corner_;
TriGridCorner empty_trigrid_center_;
pcl::PointCloud<PointType> ptCloud_tgfwise_ground_;
pcl::PointCloud<PointType> ptCloud_tgfwise_nonground_;
pcl::PointCloud<PointType> ptCloud_tgfwise_outliers_;
pcl::PointCloud<PointType> ptCloud_tgfwise_obstacle_;
pcl::PointCloud<PointType> ptCloud_nodewise_ground_;
pcl::PointCloud<PointType> ptCloud_nodewise_nonground_;
pcl::PointCloud<PointType> ptCloud_nodewise_outliers_;
pcl::PointCloud<PointType> ptCloud_nodewise_obstacle_;
pcl::PointCloud<PointType> ptCloud_nodewise_nonground_tmp;
pcl::PointCloud<PointType> ptCloud_nodewise_ground_tmp;
pcl::PointCloud<PointType> ptCloud_nodewise_obstacle_tmp;
public:
TravelGroundSeg(){
};
void setParams(const double max_range, const double min_range, const double resolution,
const int num_iter, const int num_lpr, const int num_min_pts, const double th_seeds,
const double th_dist, const double th_outlier, const double th_normal, const double th_weight,
const double th_lcc_normal_similiarity, const double th_lcc_planar_model_dist, const double th_obstacle,
const bool refine_mode, const bool visualization) {
MAX_RANGE_ = max_range;
MIN_RANGE_ = min_range;
TGF_RESOLUTION_ = resolution;
NUM_ITER_ = num_iter;
NUM_LRP_ = num_lpr;
NUM_MIN_POINTS_ = num_min_pts;
TH_SEEDS_ = th_seeds;
TH_DIST_ = th_dist;
TH_OUTLIER_ = th_outlier;
TH_NORMAL_ = th_normal;
TH_WEIGHT_ = th_weight;
TH_LCC_NORMAL_SIMILARITY_ = th_lcc_normal_similiarity;
TH_LCC_PLANAR_MODEL_DIST_ = th_lcc_planar_model_dist;
TH_OBSTACLE_HEIGHT_ = th_obstacle;
REFINE_MODE_ = refine_mode;
VIZ_MDOE_ = visualization;
initTriGridField(trigrid_field_);
initTriGridCorners(trigrid_corners_, trigrid_centers_);
ptCloud_tgfwise_ground_.clear();
ptCloud_tgfwise_ground_.reserve(PTCLOUD_SIZE);
ptCloud_tgfwise_nonground_.clear();
ptCloud_tgfwise_nonground_.reserve(PTCLOUD_SIZE);
ptCloud_tgfwise_outliers_.clear();
ptCloud_tgfwise_outliers_.reserve(PTCLOUD_SIZE);
ptCloud_tgfwise_obstacle_.clear();
ptCloud_tgfwise_obstacle_.reserve(PTCLOUD_SIZE);
ptCloud_nodewise_ground_.clear();
ptCloud_nodewise_ground_.reserve(NODEWISE_PTCLOUDSIZE);
ptCloud_nodewise_nonground_.clear();
ptCloud_nodewise_nonground_.reserve(NODEWISE_PTCLOUDSIZE);
ptCloud_nodewise_outliers_.clear();
ptCloud_nodewise_outliers_.reserve(NODEWISE_PTCLOUDSIZE);
ptCloud_nodewise_obstacle_.clear();
ptCloud_nodewise_obstacle_.reserve(NODEWISE_PTCLOUDSIZE);
ptCloud_nodewise_nonground_tmp.reserve(NODEWISE_PTCLOUDSIZE);
ptCloud_nodewise_ground_tmp.reserve(NODEWISE_PTCLOUDSIZE);
ptCloud_nodewise_obstacle_tmp.reserve(NODEWISE_PTCLOUDSIZE);
};
// Converted inputs to Eigen::Vector3d so that we can feed Open3D vectors in
void estimateGround(const std::vector<Eigen::Vector3d>& cloud_in_vec,
std::vector<Eigen::Vector3d>& cloudGround_out_vec,
std::vector<Eigen::Vector3d>& cloudNonground_out_vec,
std::vector<Eigen::Vector3d>& ground_inds_vec){
// 0. Init
pcl::PointCloud<PointType> cloud_in;
pcl::PointCloud<PointType> cloudGround_out;
pcl::PointCloud<PointType> cloudNonground_out;
// Convert Eigen to PointCloud
for (uint32_t i=0; i < cloud_in_vec.size(); i++) {
PointType point;
point.x = cloud_in_vec[i][0];
point.y = cloud_in_vec[i][1];
point.z = cloud_in_vec[i][2];
point.id = i;
cloud_in.push_back(point);
}
ptCloud_tgfwise_outliers_.clear();
ptCloud_tgfwise_outliers_.reserve(cloud_in.size());
// 1. Embed PointCloud to TriGridField
clearTriGridField(trigrid_field_);
clearTriGridCorners(trigrid_corners_, trigrid_centers_);
embedCloudToTriGridField(cloud_in, trigrid_field_);
// 2. Node-wise Terrain Modeling
modelNodeWiseTerrain(trigrid_field_);
// 3. Breadth-first Traversable Graph Search
BreadthFirstTraversableGraphSearch(trigrid_field_);
setTGFCornersCenters(trigrid_field_, trigrid_corners_, trigrid_centers_);
// 4. TGF-wise Traversable Terrain Model Fitting
if (REFINE_MODE_){
fitTGFWiseTraversableTerrainModel(trigrid_field_, trigrid_corners_, trigrid_centers_);
}
// 5. Ground Segmentation
segmentTGFGround(trigrid_field_, ptCloud_tgfwise_ground_, ptCloud_tgfwise_nonground_, ptCloud_tgfwise_obstacle_, ptCloud_tgfwise_outliers_);
cloudGround_out = ptCloud_tgfwise_ground_;
cloudNonground_out = ptCloud_tgfwise_nonground_;
// Convert PointCloud back to Eigen::Vector3d
for (auto &point : cloudGround_out.points){
Eigen::Vector3d point_eigen (static_cast<double>(point.x), static_cast<double>(point.y), static_cast<double>(point.z));
cloudGround_out_vec.push_back(point_eigen);
Eigen::Vector3d inds (static_cast<double>(point.id), 0.0, 0.0);
ground_inds_vec.push_back(inds);
}
for (auto &point : cloudNonground_out.points){
Eigen::Vector3d point_eigen (static_cast<double>(point.x), static_cast<double>(point.y), static_cast<double>(point.z));
cloudNonground_out_vec.push_back(point_eigen);
}
return;
};
TriGridIdx getTriGridIdx(const float& x_in, const float& y_in){
TriGridIdx tgf_idx;
int r_i = (x_in - tgf_min_x)/TGF_RESOLUTION_;
int c_i = (y_in - tgf_min_y)/TGF_RESOLUTION_;
int t_i = 0;
double angle = atan2(y_in-(c_i*TGF_RESOLUTION_ + TGF_RESOLUTION_/2 + tgf_min_y), x_in-(r_i*TGF_RESOLUTION_ + TGF_RESOLUTION_/2 + tgf_min_x));
if (angle>=(M_PI/4) && angle <(3*M_PI/4)){
t_i = 1;
} else if (angle>=(-M_PI/4) && angle <(M_PI/4)){
t_i = 0;
} else if (angle>=(-3*M_PI/4) && angle <(-M_PI/4)){
t_i = 3;
} else{
t_i = 2;
}
tgf_idx.row = r_i;
tgf_idx.col = c_i;
tgf_idx.tri = t_i;
return tgf_idx;
}
TriGridNode<PointType> getTriGridNode(const float& x_in, const float& y_in){
TriGridNode<PointType> node;
TriGridIdx node_idx = getTriGridIdx(x_in, y_in);
node = trigrid_field_[node_idx.row][node_idx.col][node_idx.tri];
return node;
};
TriGridNode<PointType> getTriGridNode(const TriGridIdx& tgf_idx){
TriGridNode<PointType> node;
node = trigrid_field_[tgf_idx.row][tgf_idx.col][tgf_idx.tri];
return node;
};
bool is_traversable(const float& x_in, const float& y_in){
TriGridNode<PointType> node = getTriGridNode(x_in, y_in);
if (node.node_type == GROUND){
return true;
} else{
return false;
}
};
pcl::PointCloud<PointType> getObstaclePC(){
pcl::PointCloud<PointType> cloud_obstacle;
cloud_obstacle = ptCloud_tgfwise_obstacle_;
return cloud_obstacle;
};
private:
double tgf_max_x, tgf_max_y, tgf_min_x, tgf_min_y;
double rows_, cols_;
void initTriGridField(TriGridField<PointType>& tgf_in){
// print("Initializing TriGridField...");
tgf_max_x = MAX_RANGE_;
tgf_max_y = MAX_RANGE_;
tgf_min_x = -MAX_RANGE_;
tgf_min_y = -MAX_RANGE_;
rows_ = (int)(tgf_max_x - tgf_min_x) / TGF_RESOLUTION_;
cols_ = (int)(tgf_max_y - tgf_min_y) / TGF_RESOLUTION_;
empty_cloud_.clear();
empty_cloud_.reserve(PTCLOUD_SIZE);
// Set Node structure
empty_trigrid_node_.node_type = UNKNOWN;
empty_trigrid_node_.ptCloud.clear();
empty_trigrid_node_.ptCloud.reserve(NODEWISE_PTCLOUDSIZE);
empty_trigrid_node_.is_curr_data = false;
empty_trigrid_node_.need_recheck = false;
empty_trigrid_node_.is_visited = false;
empty_trigrid_node_.is_rejection = false;
empty_trigrid_node_.check_life = 10;
empty_trigrid_node_.depth = -1;
empty_trigrid_node_.normal;
empty_trigrid_node_.mean_pt;
empty_trigrid_node_.d = 0;
empty_trigrid_node_.singular_values;
empty_trigrid_node_.eigen_vectors;
empty_trigrid_node_.weight = 0;
empty_trigrid_node_.th_dist_d = 0;
empty_trigrid_node_.th_outlier_d = 0;
// Set TriGridField
tgf_in.clear();
std::vector<GridNode<PointType>> vec_gridnode;
for (int i = 0; i < 4 ; i ++)
empty_grid_nodes_.emplace_back(empty_trigrid_node_);
for (int i=0; i< cols_; i++){ vec_gridnode.emplace_back(empty_grid_nodes_);}
for (int j=0; j< rows_; j++){ tgf_in.emplace_back(vec_gridnode);}
return;
};
void initTriGridCorners(std::vector<std::vector<TriGridCorner>>& trigrid_corners_in,
std::vector<std::vector<TriGridCorner>>& trigrid_centers_in){
// print("Initializing TriGridCorners...");
// Set TriGridCorner
empty_trigrid_corner_.x = empty_trigrid_corner_.y = 0.0;
empty_trigrid_corner_.zs.clear();
empty_trigrid_corner_.zs.reserve(8);
empty_trigrid_corner_.weights.clear();
empty_trigrid_corner_.weights.reserve(8);
empty_trigrid_center_.x = empty_trigrid_center_.y = 0.0;
empty_trigrid_center_.zs.clear();
empty_trigrid_center_.zs.reserve(4);
empty_trigrid_center_.weights.clear();
empty_trigrid_center_.weights.reserve(4);
trigrid_corners_in.clear();
trigrid_centers_in.clear();
// Set Corners and Centers
std::vector<TriGridCorner> col_corners;
std::vector<TriGridCorner> col_centers;
for (int i=0; i< cols_; i++){
col_corners.emplace_back(empty_trigrid_corner_);
col_centers.emplace_back(empty_trigrid_center_);
}
col_corners.emplace_back(empty_trigrid_corner_);
for (int j=0; j< rows_; j++){
trigrid_corners_in.emplace_back(col_corners);
trigrid_centers_in.emplace_back(col_centers);
}
trigrid_corners_in.emplace_back(col_corners);
return;
};
void clearTriGridField(TriGridField<PointType> &tgf_in){
// print("Clearing TriGridField...");
for (int r_i = 0; r_i < rows_; r_i++){
for (int c_i = 0; c_i < cols_; c_i++){
tgf_in[r_i][c_i] = empty_grid_nodes_;
}
}
return;
};
void clearTriGridCorners(std::vector<std::vector<TriGridCorner>>& trigrid_corners_in,
std::vector<std::vector<TriGridCorner>>& trigrid_centers_in){
// print("Clearing TriGridCorners...");
TriGridCorner tmp_corner = empty_trigrid_corner_;
TriGridCorner tmp_center = empty_trigrid_center_;
for (int r_i = 0; r_i < rows_+1; r_i++){
for (int c_i = 0; c_i < cols_+1; c_i++){
tmp_corner.x = (r_i)*TGF_RESOLUTION_+tgf_min_x; tmp_corner.y = (c_i)*TGF_RESOLUTION_+tgf_min_y;
tmp_corner.zs.clear();
tmp_corner.weights.clear();
trigrid_corners_in[r_i][c_i] = tmp_corner;
if (r_i < rows_ && c_i < cols_) {
tmp_center.x = (r_i+0.5)*TGF_RESOLUTION_+tgf_min_x; tmp_center.y = (c_i+0.5)*TGF_RESOLUTION_+tgf_min_y;
tmp_center.zs.clear();
tmp_center.weights.clear();
trigrid_centers_in[r_i][c_i] = tmp_center;
}
}
}
return;
};
double xy_2Dradius(double x, double y){
return sqrt(x*x + y*y);
};
bool filterPoint(const PointType &pt_in){
double xy_range = xy_2Dradius(pt_in.x, pt_in.y);
if (xy_range >= MAX_RANGE_ || xy_range <= MIN_RANGE_) return true;
return false;
}
void embedCloudToTriGridField(const pcl::PointCloud<PointType>& cloud_in, TriGridField<PointType>& tgf_out) {
// print("Embedding PointCloud to TriGridField...");
for (auto const &pt: cloud_in.points){
if (filterPoint(pt)){
ptCloud_tgfwise_outliers_.points.push_back(pt);
continue;
}
int r_i = (pt.x - tgf_min_x)/TGF_RESOLUTION_;
int c_i = (pt.y - tgf_min_y)/TGF_RESOLUTION_;
if (r_i < 0 || r_i >= rows_ || c_i < 0 || c_i >= cols_) {
ptCloud_tgfwise_outliers_.points.push_back(pt);
continue;
}
double angle = atan2(pt.y-(c_i*TGF_RESOLUTION_ + TGF_RESOLUTION_/2 + tgf_min_y), pt.x-(r_i*TGF_RESOLUTION_ + TGF_RESOLUTION_/2 + tgf_min_x));
if (angle>=(M_PI/4) && angle <(3*M_PI/4)){
// left side
tgf_out[r_i][c_i][1].ptCloud.push_back(pt);
if(!tgf_out[r_i][c_i][1].is_curr_data) {tgf_out[r_i][c_i][1].is_curr_data = true;}
} else if (angle>=(-M_PI/4) && angle <(M_PI/4)){
// upper side
tgf_out[r_i][c_i][0].ptCloud.push_back(pt);
if (!tgf_out[r_i][c_i][0].is_curr_data){tgf_out[r_i][c_i][0].is_curr_data = true;}
} else if (angle>=(-3*M_PI/4) && angle <(-M_PI/4)){
// right side
tgf_out[r_i][c_i][3].ptCloud.push_back(pt);
if (!tgf_out[r_i][c_i][3].is_curr_data) {tgf_out[r_i][c_i][3].is_curr_data = true;}
} else{
// lower side
tgf_out[r_i][c_i][2].ptCloud.push_back(pt);
if (!tgf_out[r_i][c_i][2].is_curr_data) {tgf_out[r_i][c_i][2].is_curr_data = true;}
}
}
return;
};
void extractInitialSeeds(const pcl::PointCloud<PointType>& p_sorted, pcl::PointCloud<PointType>& init_seeds){
//function for uniform mode
init_seeds.points.clear();
// LPR is the mean of Low Point Representative
double sum = 0;
int cnt = 0;
// Calculate the mean height value.
for (int i=0; i< (int) p_sorted.points.size() && cnt<NUM_LRP_; i++){
sum += p_sorted.points[i].z;
cnt++;
}
double lpr_height = cnt!=0?sum/cnt:0;
for(int i=0 ; i< (int) p_sorted.points.size() ; i++){
if(p_sorted.points[i].z < lpr_height + TH_SEEDS_){
if (p_sorted.points[i].z < lpr_height-TH_OUTLIER_) continue;
init_seeds.points.push_back(p_sorted.points[i]);
}
}
return;
}
void estimatePlanarModel(const pcl::PointCloud<PointType>& ground_in, TriGridNode<PointType>& node_out) {
// function for uniform mode
Eigen::Matrix3f cov_;
Eigen::Vector4f pc_mean_;
pcl::computeMeanAndCovarianceMatrix(ground_in, cov_, pc_mean_);
// Singular Value Decomposition: SVD
Eigen::JacobiSVD<Eigen::MatrixXf> svd(cov_, Eigen::DecompositionOptions::ComputeFullU);
// Use the least singular vector as normal
node_out.eigen_vectors = svd.matrixU();
if (node_out.eigen_vectors.col(2)(2,0)<0) {
node_out.eigen_vectors.col(0)*= -1;
node_out.eigen_vectors.col(2)*= -1;
}
node_out.normal = node_out.eigen_vectors.col(2);
node_out.singular_values = svd.singularValues();
// mean ground seeds value
node_out.mean_pt = pc_mean_.head<3>();
// according to noraml.T*[x,y,z] = -d
node_out.d = -(node_out.normal.transpose()*node_out.mean_pt)(0,0);
// set distance theshold to 'th_dist - d'
node_out.th_dist_d = TH_DIST_ - node_out.d;
node_out.th_outlier_d = -node_out.d - TH_OUTLIER_;
return;
}
void modelPCAbasedTerrain(TriGridNode<PointType>& node_in) {
// Initailization
if (!ptCloud_nodewise_ground_.empty()) ptCloud_nodewise_ground_.clear();
// Tri Grid Initialization
// When to initialize the planar model, we don't have prior. so outlier is removed in heuristic parameter.
pcl::PointCloud<PointType> sort_ptCloud = node_in.ptCloud;
// sort in z-coordinate
sort(sort_ptCloud.points.begin(), sort_ptCloud.end(), point_z_cmp<PointType>);
// Set init seeds
extractInitialSeeds(sort_ptCloud, ptCloud_nodewise_ground_);
Eigen::MatrixXf points(sort_ptCloud.points.size(),3);
int j = 0;
for (auto& p:sort_ptCloud.points){
points.row(j++)<<p.x, p.y, p.z;
}
// Extract Ground
for (int i =0; i < NUM_ITER_; i++){
estimatePlanarModel(ptCloud_nodewise_ground_, node_in);
if(ptCloud_nodewise_ground_.size() < 3){
node_in.node_type = NONGROUND;
break;
}
ptCloud_nodewise_ground_.clear();
// threshold filter
Eigen::VectorXf result = points*node_in.normal;
for (int r = 0; r<result.rows(); r++){
if (i < NUM_ITER_-1){
if (result[r]<node_in.th_dist_d){
ptCloud_nodewise_ground_.push_back(sort_ptCloud.points[r]);
}
} else {
// Final interation
if (node_in.normal(2,0) < TH_NORMAL_ ){
node_in.node_type = NONGROUND;
} else {
node_in.node_type = GROUND;
}
}
}
}
return;
}
double calcNodeWeight(const TriGridNode<PointType>& node_in){
double weight = 0;
// weight = (node_in.singular_values[0]/node_in.singular_values[2] + node_in.singular_values[1]/node_in.singular_values[2])/(node_in.singular_values[0]/node_in.singular_values[1]);
weight = (node_in.singular_values[0] + node_in.singular_values[1])*node_in.singular_values[1]/(node_in.singular_values[0]*node_in.singular_values[2]+0.001);
return weight;
}
void modelNodeWiseTerrain(TriGridField<PointType>& tgf_in) {
// print("Node-wise Terrain Modeling...");
for (int r_i = 0; r_i < rows_; r_i++){
for (int c_i = 0; c_i < cols_; c_i++){
for (int s_i = 0; s_i < 4; s_i++){
if (tgf_in[r_i][c_i][s_i].is_curr_data){
if (tgf_in[r_i][c_i][s_i].ptCloud.size() < NUM_MIN_POINTS_){
tgf_in[r_i][c_i][s_i].node_type = UNKNOWN;
continue;
} else {
modelPCAbasedTerrain(tgf_in[r_i][c_i][s_i]);
if (tgf_in[r_i][c_i][s_i].node_type == GROUND){ tgf_in[r_i][c_i][s_i].weight = calcNodeWeight(tgf_in[r_i][c_i][s_i]); }
}
}
}
}
}
return;
};
void findDominantNode(const TriGridField<PointType>& tgf_in, TriGridIdx& node_idx_out) {
// Find the dominant node
// std::cout << "Find the dominant node..." << std::endl;
TriGridIdx max_tri_idx;
TriGridIdx ego_idx;
ego_idx.row = (int)((0-tgf_min_x)/TGF_RESOLUTION_);
ego_idx.col = (int)((0-tgf_min_y)/TGF_RESOLUTION_);
ego_idx.tri = 0;
max_tri_idx = ego_idx;
for (int r_i = ego_idx.row - 2; r_i < ego_idx.row + 2; r_i++){
for (int c_i = ego_idx.col - 2; c_i < ego_idx.col + 2; c_i++){
for (int s_i = 0; s_i < 4; s_i++){
if (tgf_in[r_i][c_i][s_i].is_curr_data){
if (tgf_in[r_i][c_i][s_i].node_type == GROUND){
if (tgf_in[r_i][c_i][s_i].weight > tgf_in[max_tri_idx.row][max_tri_idx.row][max_tri_idx.tri].weight){
max_tri_idx.row = r_i;
max_tri_idx.col = c_i;
max_tri_idx.tri = s_i;
}
}
}
}
}
}
node_idx_out = max_tri_idx;
return;
};
void searchNeighborNodes(const TriGridIdx &cur_idx, std::vector<TriGridIdx> &neighbor_idxs) {
neighbor_idxs.clear();
neighbor_idxs.reserve(14);
int r_i = cur_idx.row;
int c_i = cur_idx.col;
int s_i = cur_idx.tri;
std::vector<TriGridIdx> tmp_neighbors;
tmp_neighbors.clear();
tmp_neighbors.reserve(14);
TriGridIdx neighbor_idx;
for (int s_i = 0; s_i < 4 ; s_i++){
if (s_i == cur_idx.tri) continue;
neighbor_idx = cur_idx;
neighbor_idx.tri = s_i;
tmp_neighbors.push_back(neighbor_idx);
}
switch (s_i) {
case 0:
tmp_neighbors.push_back({r_i+1, c_i+1, 2});
tmp_neighbors.push_back({r_i+1, c_i+1, 3});
tmp_neighbors.push_back({r_i+1, c_i , 1});
tmp_neighbors.push_back({r_i+1, c_i , 2});
tmp_neighbors.push_back({r_i+1, c_i , 3});
tmp_neighbors.push_back({r_i+1, c_i-1, 1});
tmp_neighbors.push_back({r_i+1, c_i-1, 2});
tmp_neighbors.push_back({r_i , c_i+1, 0});
tmp_neighbors.push_back({r_i , c_i+1, 3});
tmp_neighbors.push_back({r_i , c_i-1, 0});
tmp_neighbors.push_back({r_i , c_i-1, 1});
break;
case 1:
tmp_neighbors.push_back({r_i+1, c_i+1, 2});
tmp_neighbors.push_back({r_i+1, c_i+1, 3});
tmp_neighbors.push_back({r_i+1, c_i , 1});
tmp_neighbors.push_back({r_i+1, c_i , 2});
tmp_neighbors.push_back({r_i , c_i+1, 0});
tmp_neighbors.push_back({r_i , c_i+1, 2});
tmp_neighbors.push_back({r_i , c_i+1, 3});
tmp_neighbors.push_back({r_i-1, c_i+1, 0});
tmp_neighbors.push_back({r_i-1, c_i+1, 3});
tmp_neighbors.push_back({r_i-1, c_i+1, 0});
tmp_neighbors.push_back({r_i-1, c_i , 1});
break;
case 2:
tmp_neighbors.push_back({r_i , c_i+1, 2});
tmp_neighbors.push_back({r_i , c_i+1, 3});
tmp_neighbors.push_back({r_i , c_i-1, 1});
tmp_neighbors.push_back({r_i , c_i-1, 2});
tmp_neighbors.push_back({r_i-1, c_i+1, 0});
tmp_neighbors.push_back({r_i-1, c_i+1, 3});
tmp_neighbors.push_back({r_i-1, c_i , 0});
tmp_neighbors.push_back({r_i-1, c_i , 1});
tmp_neighbors.push_back({r_i-1, c_i , 3});
tmp_neighbors.push_back({r_i-1, c_i-1, 0});
tmp_neighbors.push_back({r_i-1, c_i-1, 1});
break;
case 3:
tmp_neighbors.push_back({r_i+1, c_i , 2});
tmp_neighbors.push_back({r_i+1, c_i , 3});
tmp_neighbors.push_back({r_i+1, c_i-1, 1});
tmp_neighbors.push_back({r_i+1, c_i-1, 2});
tmp_neighbors.push_back({r_i , c_i-1, 0});
tmp_neighbors.push_back({r_i , c_i-1, 1});
tmp_neighbors.push_back({r_i , c_i-1, 2});
tmp_neighbors.push_back({r_i-1, c_i , 0});
tmp_neighbors.push_back({r_i-1, c_i , 3});
tmp_neighbors.push_back({r_i-1, c_i-1, 0});
tmp_neighbors.push_back({r_i-1, c_i-1, 1});
break;
default:
break;
}
for (int n_i = 0 ; n_i < (int) tmp_neighbors.size() ; n_i++) {
if (tmp_neighbors[n_i].row >= rows_ || tmp_neighbors[n_i].row < 0) {
continue;
}
if (tmp_neighbors[n_i].col >= cols_ || tmp_neighbors[n_i].col < 0) {
continue;
}
neighbor_idxs.push_back(tmp_neighbors[n_i]);
}
}
void searchAdjacentNodes(const TriGridIdx &cur_idx, std::vector<TriGridIdx> &adjacent_idxs) {
adjacent_idxs.clear();
adjacent_idxs.reserve(3);
int r_i = cur_idx.row;
int c_i = cur_idx.col;
int s_i = cur_idx.tri;
std::vector<TriGridIdx> tmp_neighbors;
tmp_neighbors.clear();
tmp_neighbors.reserve(3);
TriGridIdx neighbor_idx;
switch (s_i) {
case 0:
tmp_neighbors.push_back({r_i+1, c_i, 2});
tmp_neighbors.push_back({r_i , c_i, 3});
tmp_neighbors.push_back({r_i , c_i, 1});
break;
case 1:
tmp_neighbors.push_back({r_i, c_i+1, 3});
tmp_neighbors.push_back({r_i, c_i , 0});
tmp_neighbors.push_back({r_i, c_i , 2});
break;
case 2:
tmp_neighbors.push_back({r_i-1, c_i, 0});
tmp_neighbors.push_back({r_i , c_i, 1});
tmp_neighbors.push_back({r_i , c_i, 3});
break;
case 3:
tmp_neighbors.push_back({r_i, c_i-1, 1});
tmp_neighbors.push_back({r_i, c_i , 2});
tmp_neighbors.push_back({r_i, c_i , 0});
break;
default:
break;
}
for (int n_i = 0 ; n_i < (int) tmp_neighbors.size() ; n_i++) {
if (tmp_neighbors[n_i].row >= rows_ || tmp_neighbors[n_i].row < 0) {
continue;
}
if (tmp_neighbors[n_i].col >= cols_ || tmp_neighbors[n_i].col < 0) {
continue;
}
adjacent_idxs.push_back(tmp_neighbors[n_i]);
}
}
bool LocalConvecityConcavity(const TriGridField<PointType> &tgf, const TriGridIdx &cur_node_idx, const TriGridIdx &neighbor_idx,
double & thr_local_normal, double & thr_local_dist) {
TriGridNode<PointType> current_node = tgf[cur_node_idx.row][cur_node_idx.col][cur_node_idx.tri];
TriGridNode<PointType> neighbor_node = tgf[neighbor_idx.row][neighbor_idx.col][neighbor_idx.tri];
Eigen::Vector3f normal_src = current_node.normal;
Eigen::Vector3f normal_tgt = neighbor_node.normal;
Eigen::Vector3f meanPt_diff_s2t = neighbor_node.mean_pt - current_node.mean_pt;
double diff_norm = meanPt_diff_s2t.norm();
double dist_s2t = normal_src.dot(meanPt_diff_s2t);
double dist_t2s = normal_tgt.dot(-meanPt_diff_s2t);
double normal_similarity = normal_src.dot(normal_tgt);
double TH_NORMAL_cos_similarity = sin(diff_norm*thr_local_normal);
if ((normal_similarity < (1-TH_NORMAL_cos_similarity))) {
return false;
}
double TH_DIST_to_planar = diff_norm*sin(thr_local_dist);
if ( (abs(dist_s2t) > TH_DIST_to_planar || abs(dist_t2s) > TH_DIST_to_planar) ) {
return false;
}
return true;
}
void BreadthFirstTraversableGraphSearch(TriGridField<PointType>& tgf_in) {
// Find the dominant node
std::queue<TriGridIdx> searching_idx_queue;
TriGridIdx dominant_node_idx;
findDominantNode(tgf_in, dominant_node_idx);
tgf_in[dominant_node_idx.row][dominant_node_idx.col][dominant_node_idx.tri].is_visited = true;
tgf_in[dominant_node_idx.row][dominant_node_idx.col][dominant_node_idx.tri].depth = 0;
tgf_in[dominant_node_idx.row][dominant_node_idx.col][dominant_node_idx.tri].node_type = GROUND;
searching_idx_queue.push(dominant_node_idx);
double max_planar_height = 0;
trigrid_edges_.clear();
trigrid_edges_.reserve(rows_*cols_*4);
TriGridEdge cur_edge;
TriGridIdx current_node_idx;
while (!searching_idx_queue.empty()){
// set current node
current_node_idx = searching_idx_queue.front();
searching_idx_queue.pop();
// search the neighbor nodes
std::vector<TriGridIdx> neighbor_idxs;
searchNeighborNodes(current_node_idx, neighbor_idxs);
// set the traversable edges
for (int i = 0; i < (int) neighbor_idxs.size(); i++){
// if the neighbor node is traversable, add it to the queue
TriGridIdx n_i = neighbor_idxs[i];
if (tgf_in[n_i.row][n_i.col][n_i.tri].depth >=0){
continue;
}
if (tgf_in[n_i.row][n_i.col][n_i.tri].is_visited) {
if (!tgf_in[n_i.row][n_i.col][n_i.tri].need_recheck){
continue;
} else {
if (tgf_in[n_i.row][n_i.col][n_i.tri].check_life <= 0){
continue;
}
}
continue;
} else {
if (tgf_in[n_i.row][n_i.col][n_i.tri].node_type != GROUND) {
continue;
}
}
tgf_in[n_i.row][n_i.col][n_i.tri].is_visited =true;
if (!LocalConvecityConcavity(tgf_in, current_node_idx, n_i, TH_LCC_NORMAL_SIMILARITY_, TH_LCC_PLANAR_MODEL_DIST_)) {
tgf_in[n_i.row][n_i.col][n_i.tri].is_rejection = true;
tgf_in[n_i.row][n_i.col][n_i.tri].node_type = NONGROUND;
if(tgf_in[n_i.row][n_i.col][n_i.tri].check_life > 0) {
tgf_in[n_i.row][n_i.col][n_i.tri].check_life -=1;
tgf_in[n_i.row][n_i.col][n_i.tri].need_recheck = true;
} else {
tgf_in[n_i.row][n_i.col][n_i.tri].need_recheck = false;
}
continue;
}
if (max_planar_height < tgf_in[n_i.row][n_i.col][n_i.tri].mean_pt[2]) max_planar_height = tgf_in[n_i.row][n_i.col][n_i.tri].mean_pt[2];
tgf_in[n_i.row][n_i.col][n_i.tri].node_type = GROUND;
tgf_in[n_i.row][n_i.col][n_i.tri].is_rejection = false;
tgf_in[n_i.row][n_i.col][n_i.tri].depth = tgf_in[current_node_idx.row][current_node_idx.col][current_node_idx.tri].depth + 1;
if (VIZ_MDOE_){
cur_edge.Pair.first = current_node_idx;
cur_edge.Pair.second = n_i;
cur_edge.is_traversable = true;
trigrid_edges_.push_back(cur_edge);
}
searching_idx_queue.push(n_i);
}
if (searching_idx_queue.empty()){
// set the new dominant node
for (int r_i = 0; r_i < rows_; r_i++) {
for (int c_i = 0; c_i < cols_; c_i++) {
for (int s_i = 0; s_i < (int) tgf_in[r_i][c_i].size() ; s_i++){
if (tgf_in[r_i][c_i][s_i].is_visited) { continue; }
if (tgf_in[r_i][c_i][s_i].node_type != GROUND ) { continue; }
// if (tgf_in[r_i][c_i][s_i].mean_pt[2] >= max_planar_height+1) { continue; }
if (tgf_in[r_i][c_i][s_i].depth >= 0) { continue; }
// if (tgf_in[r_i][c_i][s_i].weight > 5*TH_WEIGHT_){
tgf_in[r_i][c_i][s_i].depth = 0;
tgf_in[r_i][c_i][s_i].is_visited = true;
TriGridIdx new_dominant_idx = {r_i, c_i, s_i};
searching_idx_queue.push(new_dominant_idx);
// }
}
}
}
}
}
return;
};
double getCornerWeight(const TriGridNode<PointType>& node_in, const pcl::PointXYZ &tgt_corner){
double xy_dist = sqrt( (node_in.mean_pt[0]-tgt_corner.x)*(node_in.mean_pt[0]-tgt_corner.x)+(node_in.mean_pt[1]-tgt_corner.y)*(node_in.mean_pt[1]-tgt_corner.y) );
return (node_in.weight/xy_dist);
}
void setTGFCornersCenters(const TriGridField<PointType>& tgf_in,
std::vector<std::vector<TriGridCorner>>& trigrid_corners_out,
std::vector<std::vector<TriGridCorner>>& trigrid_centers_out) {
pcl::PointXYZ corner_TL, corner_BL, corner_BR, corner_TR, corner_C;
for (int r_i = 0; r_i<rows_; r_i++){
for (int c_i = 0; c_i<cols_; c_i++){
corner_TL.x = trigrid_corners_out[r_i+1][c_i+1].x; corner_TL.y = trigrid_corners_out[r_i+1][c_i+1].y; // LT
corner_BL.x = trigrid_corners_out[r_i][c_i+1].x; corner_BL.y = trigrid_corners_out[r_i][c_i+1].y; // LL
corner_BR.x = trigrid_corners_out[r_i][c_i].x; corner_BR.y = trigrid_corners_out[r_i][c_i].y; // RL
corner_TR.x = trigrid_corners_out[r_i+1][c_i].x; corner_TR.y = trigrid_corners_out[r_i+1][c_i].y; // RT
corner_C.x = trigrid_centers_out[r_i][c_i].x; corner_C.y = trigrid_centers_out[r_i][c_i].y; // Center
for (int s_i = 0; s_i< (int) tgf_in[r_i][c_i].size();s_i++){
if (tgf_in[r_i][c_i][s_i].node_type != GROUND) { continue; }
if (tgf_in[r_i][c_i][s_i].is_rejection) { continue; }
if (tgf_in[r_i][c_i][s_i].depth == -1) { continue; }
switch(s_i){
case 0: // upper Tri-grid bin
// RT / LT / C
trigrid_corners_out[r_i+1][c_i].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_TR.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_TR.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );
trigrid_corners_out[r_i+1][c_i].weights.push_back(getCornerWeight(tgf_in[r_i][c_i][s_i],corner_TR));
trigrid_corners_out[r_i+1][c_i+1].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_TL.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_TL.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );
trigrid_corners_out[r_i+1][c_i+1].weights.push_back(getCornerWeight(tgf_in[r_i][c_i][s_i],corner_TL));
trigrid_centers_out[r_i][c_i].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_C.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_C.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );
trigrid_centers_out[r_i][c_i].weights.push_back(getCornerWeight(tgf_in[r_i][c_i][s_i],corner_C));
break;
case 1: // left Tri-grid bin
// LT / LL / C
trigrid_corners_out[r_i+1][c_i+1].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_TL.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_TL.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );
trigrid_corners_out[r_i+1][c_i+1].weights.push_back(getCornerWeight(tgf_in[r_i][c_i][s_i],corner_TL));
trigrid_corners_out[r_i][c_i+1].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_BL.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_BL.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );
trigrid_corners_out[r_i][c_i+1].weights.push_back(getCornerWeight(tgf_in[r_i][c_i][s_i],corner_BL));
trigrid_centers_out[r_i][c_i].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_C.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_C.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );
trigrid_centers_out[r_i][c_i].weights.push_back(getCornerWeight(tgf_in[r_i][c_i][s_i],corner_C));
break;
case 2: // lower Tri-grid bin
// LL / RL / C
trigrid_corners_out[r_i][c_i+1].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_BL.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_BL.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );
trigrid_corners_out[r_i][c_i+1].weights.push_back(getCornerWeight(tgf_in[r_i][c_i][s_i],corner_BL));
trigrid_corners_out[r_i][c_i].zs.push_back( (-tgf_in[r_i][c_i][s_i].normal(0,0)*corner_BR.x - tgf_in[r_i][c_i][s_i].normal(1,0)*corner_BR.y-tgf_in[r_i][c_i][s_i].d)/tgf_in[r_i][c_i][s_i].normal(2,0) );