-
-
Notifications
You must be signed in to change notification settings - Fork 4.6k
/
range_likelihood.cpp
1106 lines (962 loc) · 49.8 KB
/
range_likelihood.cpp
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
#include <GL/glew.h>
#include <time.h>
#include <pcl/pcl_config.h>
#ifdef OPENGL_IS_A_FRAMEWORK
# include <OpenGL/gl.h>
# include <OpenGL/glu.h>
#else
# include <GL/gl.h>
# include <GL/glu.h>
#endif
#include <pcl/common/time.h>
#include <pcl/simulation/range_likelihood.h>
// For adding noise:
static boost::minstd_rand generator (static_cast<unsigned int> (std::time (0))); // seed
//#define SIMULATION_DEBUG 1
#define DO_TIMING_PROFILE 0
using namespace std;
// 301 values, 0.0 uniform 1.0 normal. properly truncated/normalized
float normal_sigma0x5_normal1x0_range0to3_step0x01[] = {1.59576912f, 1.59545000f, 1.59449302f, 1.59289932f, 1.59067083f,
1.58781019f, 1.58432085f, 1.58020697f, 1.57547345f, 1.57012594f,
1.56417078f, 1.55761504f, 1.55046646f, 1.54273348f, 1.53442517f,
1.52555126f, 1.51612211f, 1.50614865f, 1.49564242f, 1.48461552f,
1.47308056f, 1.46105069f, 1.44853953f, 1.43556117f, 1.42213012f,
1.40826131f, 1.39397005f, 1.37927201f, 1.36418316f, 1.34871978f,
1.33289841f, 1.31673585f, 1.30024906f, 1.28345522f, 1.26637163f,
1.24901574f, 1.23140504f, 1.21355714f, 1.19548963f, 1.17722012f,
1.15876621f, 1.14014544f, 1.12137525f, 1.10247299f, 1.08345589f,
1.06434100f, 1.04514521f, 1.02588518f, 1.00657737f, 0.98723796f,
0.96788290f, 0.94852781f, 0.92918802f, 0.90987853f, 0.89061400f,
0.87140871f, 0.85227659f, 0.83323116f, 0.81428555f, 0.79545248f,
0.77674422f, 0.75817263f, 0.73974913f, 0.72148465f, 0.70338972f,
0.68547437f, 0.66774817f, 0.65022022f, 0.63289916f, 0.61579315f,
0.59890986f, 0.58225652f, 0.56583986f, 0.54966616f, 0.53374122f,
0.51807038f, 0.50265855f, 0.48751015f, 0.47262918f, 0.45801920f,
0.44368334f, 0.42962430f, 0.41584438f, 0.40234547f, 0.38912908f,
0.37619631f, 0.36354792f, 0.35118428f, 0.33910545f, 0.32731110f,
0.31580063f, 0.30457310f, 0.29362725f, 0.28296157f, 0.27257426f,
0.26246326f, 0.25262625f, 0.24306068f, 0.23376378f, 0.22473257f,
0.21596387f, 0.20745431f, 0.19920035f, 0.19119830f, 0.18344431f,
0.17593438f, 0.16866443f, 0.16163022f, 0.15482742f, 0.14825164f,
0.14189837f, 0.13576305f, 0.12984106f, 0.12412773f, 0.11861834f,
0.11330815f, 0.10819240f, 0.10326630f, 0.09852508f, 0.09396394f,
0.08957812f, 0.08536286f, 0.08131342f, 0.07742511f, 0.07369324f,
0.07011320f, 0.06668040f, 0.06339032f, 0.06023847f, 0.05722044f,
0.05433188f, 0.05156850f, 0.04892611f, 0.04640054f, 0.04398775f,
0.04168374f, 0.03948462f, 0.03738655f, 0.03538582f, 0.03347876f,
0.03166181f, 0.02993149f, 0.02828442f, 0.02671730f, 0.02522691f,
0.02381013f, 0.02246393f, 0.02118538f, 0.01997160f, 0.01881983f,
0.01772739f, 0.01669169f, 0.01571021f, 0.01478053f, 0.01390031f,
0.01306728f, 0.01227925f, 0.01153414f, 0.01082990f, 0.01016460f,
0.00953635f, 0.00894336f, 0.00838388f, 0.00785626f, 0.00735890f,
0.00689028f, 0.00644891f, 0.00603340f, 0.00564241f, 0.00527464f,
0.00492888f, 0.00460393f, 0.00429869f, 0.00401209f, 0.00374309f,
0.00349073f, 0.00325408f, 0.00303227f, 0.00282444f, 0.00262981f,
0.00244761f, 0.00227712f, 0.00211766f, 0.00196858f, 0.00182926f,
0.00169912f, 0.00157761f, 0.00146420f, 0.00135840f, 0.00125975f,
0.00116779f, 0.00108211f, 0.00100231f, 0.00092803f, 0.00085891f,
0.00079462f, 0.00073485f, 0.00067930f, 0.00062770f, 0.00057979f,
0.00053532f, 0.00049406f, 0.00045581f, 0.00042034f, 0.00038748f,
0.00035705f, 0.00032887f, 0.00030280f, 0.00027868f, 0.00025638f,
0.00023577f, 0.00021673f, 0.00019915f, 0.00018292f, 0.00016795f,
0.00015414f, 0.00014141f, 0.00012968f, 0.00011887f, 0.00010893f,
0.00009977f, 0.00009135f, 0.00008360f, 0.00007648f, 0.00006994f,
0.00006393f, 0.00005842f, 0.00005336f, 0.00004872f, 0.00004446f,
0.00004056f, 0.00003699f, 0.00003372f, 0.00003072f, 0.00002798f,
0.00002548f, 0.00002319f, 0.00002110f, 0.00001918f, 0.00001744f,
0.00001585f, 0.00001439f, 0.00001307f, 0.00001186f, 0.00001076f,
0.00000976f, 0.00000884f, 0.00000801f, 0.00000726f, 0.00000657f,
0.00000595f, 0.00000538f, 0.00000486f, 0.00000440f, 0.00000397f,
0.00000359f, 0.00000324f, 0.00000292f, 0.00000264f, 0.00000238f,
0.00000214f, 0.00000193f, 0.00000174f, 0.00000157f, 0.00000141f,
0.00000127f, 0.00000114f, 0.00000103f, 0.00000092f, 0.00000083f,
0.00000074f, 0.00000067f, 0.00000060f, 0.00000054f, 0.00000048f,
0.00000043f, 0.00000039f, 0.00000035f, 0.00000031f, 0.00000028f,
0.00000025f, 0.00000022f, 0.00000020f, 0.00000018f, 0.00000016f,
0.00000014f, 0.00000013f, 0.00000011f, 0.00000010f, 0.00000009f,
0.00000008f, 0.00000007f, 0.00000006f, 0.00000006f, 0.00000005f,
0.00000004f, 0.00000004f, 0.00000003f, 0.00000003f, 0.00000003f,
0.00000002f};
// Where the above if lhoodf, this a hard coded/optimized version:
//ratio = 0.99; r_min =0; r_max = 3;
//lhood = ratio/(r_max -r_min) + (1-ratio)*lhood ; hard_coded_log_lhood=log(lhood)
float hard_coded_log_lhood[] = {-1.0614388f, -1.0614480f, -1.0614757f, -1.0615217f, -1.0615862f, -1.0616689f, -1.0617698f, -1.0618887f, -1.0620256f, -1.0621803f, -1.0623526f, -1.0625423f, -1.0627491f, -1.0629730f, -1.0632135f, -1.0634705f, -1.0637437f, -1.0640327f, -1.0643372f, -1.0646569f, -1.0649914f, -1.0653405f, -1.0657036f, -1.0660804f, -1.0664705f, -1.0668735f, -1.0672889f, -1.0677164f, -1.0681554f, -1.0686054f, -1.0690662f, -1.0695370f, -1.0700176f, -1.0705073f, -1.0710057f, -1.0715124f, -1.0720267f, -1.0725482f, -1.0730764f, -1.0736108f, -1.0741509f, -1.0746962f, -1.0752462f, -1.0758003f, -1.0763581f, -1.0769191f, -1.0774827f, -1.0780486f, -1.0786162f, -1.0791851f, -1.0797547f, -1.0803247f, -1.0808945f, -1.0814638f, -1.0820321f, -1.0825989f, -1.0831639f, -1.0837267f, -1.0842868f, -1.0848439f, -1.0853977f, -1.0859476f, -1.0864935f, -1.0870350f, -1.0875718f, -1.0881035f, -1.0886298f, -1.0891506f, -1.0896655f, -1.0901742f, -1.0906766f, -1.0911723f, -1.0916613f, -1.0921433f, -1.0926181f, -1.0930855f, -1.0935454f, -1.0939976f, -1.0944421f, -1.0948787f, -1.0953073f, -1.0957277f, -1.0961400f, -1.0965441f, -1.0969398f, -1.0973272f, -1.0977063f, -1.0980769f, -1.0984391f, -1.0987930f, -1.0991384f, -1.0994755f, -1.0998042f, -1.1001246f, -1.1004367f, -1.1007407f, -1.1010364f, -1.1013241f, -1.1016038f, -1.1018756f, -1.1021396f, -1.1023958f, -1.1026444f, -1.1028855f, -1.1031191f, -1.1033454f, -1.1035646f, -1.1037767f, -1.1039819f, -1.1041802f, -1.1043719f, -1.1045570f, -1.1047358f, -1.1049082f, -1.1050746f, -1.1052349f, -1.1053894f, -1.1055382f, -1.1056815f, -1.1058193f, -1.1059518f, -1.1060792f, -1.1062016f, -1.1063192f, -1.1064320f, -1.1065402f, -1.1066440f, -1.1067435f, -1.1068389f, -1.1069302f, -1.1070176f, -1.1071012f, -1.1071811f, -1.1072575f, -1.1073306f, -1.1074003f, -1.1074668f, -1.1075303f, -1.1075909f, -1.1076486f, -1.1077036f, -1.1077560f, -1.1078059f, -1.1078533f, -1.1078985f, -1.1079414f, -1.1079821f, -1.1080208f, -1.1080576f, -1.1080925f, -1.1081256f, -1.1081569f, -1.1081867f, -1.1082148f, -1.1082415f, -1.1082667f, -1.1082906f, -1.1083132f, -1.1083345f, -1.1083547f, -1.1083737f, -1.1083917f, -1.1084086f, -1.1084246f, -1.1084397f, -1.1084538f, -1.1084672f, -1.1084798f, -1.1084917f, -1.1085028f, -1.1085133f, -1.1085231f, -1.1085324f, -1.1085411f, -1.1085492f, -1.1085569f, -1.1085640f, -1.1085707f, -1.1085770f, -1.1085829f, -1.1085885f, -1.1085936f, -1.1085985f, -1.1086030f, -1.1086072f, -1.1086111f, -1.1086148f, -1.1086183f, -1.1086215f, -1.1086245f, -1.1086272f, -1.1086298f, -1.1086323f, -1.1086345f, -1.1086366f, -1.1086385f, -1.1086404f, -1.1086420f, -1.1086436f, -1.1086451f, -1.1086464f, -1.1086477f, -1.1086488f, -1.1086499f, -1.1086509f, -1.1086518f, -1.1086527f, -1.1086534f, -1.1086542f, -1.1086549f, -1.1086555f, -1.1086561f, -1.1086566f, -1.1086571f, -1.1086575f, -1.1086580f, -1.1086583f, -1.1086587f, -1.1086590f, -1.1086593f, -1.1086596f, -1.1086599f, -1.1086601f, -1.1086603f, -1.1086605f, -1.1086607f, -1.1086609f, -1.1086610f, -1.1086611f, -1.1086613f, -1.1086614f, -1.1086615f, -1.1086616f, -1.1086617f, -1.1086618f, -1.1086619f, -1.1086619f, -1.1086620f, -1.1086620f, -1.1086621f, -1.1086621f, -1.1086622f, -1.1086622f, -1.1086623f, -1.1086623f, -1.1086623f, -1.1086624f, -1.1086624f, -1.1086624f, -1.1086624f, -1.1086624f, -1.1086625f, -1.1086625f, -1.1086625f, -1.1086625f, -1.1086625f, -1.1086625f, -1.1086625f, -1.1086625f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f, -1.1086626f};
// Disparity:
// sigma 0.025
double top_lookup[]={15.9577, 15.8165, 15.4003, 14.7308, 13.8422, 12.7779, 11.5877, 10.3231, 9.0345, 7.7674, 6.5604, 5.4433, 4.4368, 3.5527, 2.7947, 2.1596, 1.6395, 1.2227, 0.89578, 0.64471, 0.45584, 0.31662, 0.21604, 0.14482, 0.095364, 0.061691, 0.039205, 0.024476, 0.015011, 0.0090444, 0.0053532, 0.0031126, 0.001778, 0.0009977, 0.00054999, 0.00029784, 0.00015845, 8.2811e-05, 4.2517e-05, 2.1444e-05, 1.0625e-05, 5.1718e-06, 2.473e-06, 1.1617e-06, 5.361e-07, 2.4304e-07, 1.0824e-07, 4.7354e-08, 2.0353e-08, 8.5933e-09, 3.5644e-09, 1.4524e-09, 5.8138e-10, 2.2862e-10, 8.832e-11, 3.3518e-11, 1.2496e-11, 4.5766e-12, 1.6466e-12, 5.8201e-13, 2.0209e-13, 6.8935e-14, 2.31e-14, 7.6043e-15, 2.4592e-15, 7.8127e-16, 2.4383e-16, 7.4758e-17, 2.2517e-17, 6.6624e-18, 1.9366e-18, 5.5299e-19, 1.5512e-19, 4.2749e-20, 1.1573e-20, 3.0778e-21, 8.0413e-22, 2.0639e-22, 5.2038e-23, 1.289e-23, 3.1365e-24, 7.4975e-25, 1.7606e-25, 4.0617e-26, 9.2049e-27, 2.0493e-27, 4.4821e-28, 9.6302e-29, 2.0327e-29, 4.2148e-30, 8.5855e-31, 1.718e-31, 3.3774e-32, 6.5224e-33, 1.2374e-33, 2.3062e-34, 4.2225e-35, 7.5947e-36, 1.3419e-36, 2.3294e-37, 3.9721e-38, 6.6539e-39, 1.095e-39, 1.7703e-40, 2.8115e-41, 4.3864e-42, 6.7231e-43, 1.0123e-43, 1.4973e-44, 2.1758e-45, 3.1059e-46, 4.3555e-47, 6.0003e-48, 8.1205e-49, 1.0796e-49, 1.4101e-50, 1.8092e-51, 2.2804e-52, 2.8237e-53, 3.4349e-54, 4.1047e-55, 4.8186e-56, 5.5571e-57, 6.2958e-58, 7.007e-59, 7.6611e-60, 8.2287e-61, 8.6827e-62, 9.0002e-63, 9.165e-64, 9.1683e-65, 9.01e-66, 8.6984e-67, 8.2497e-68, 7.6862e-69, 7.035e-70, 6.3255e-71, 5.5874e-72, 4.8484e-73, 4.133e-74, 3.4611e-75, 2.8474e-76, 2.3012e-77, 1.827e-78, 1.425e-79, 1.0918e-80, 8.2183e-82, 6.077e-83, 4.4144e-84, 3.1502e-85, 2.2084e-86, 1.5209e-87, 1.029e-88, 6.8387e-90, 4.4651e-91, 2.864e-92, 1.8046e-93, 1.1171e-94, 6.793e-96, 4.058e-97, 2.3815e-98, 1.373e-99, 7.7759e-101, 4.3264e-102, 2.3647e-103, 1.2697e-104, 6.6975e-106, 3.4706e-107, 1.7667e-108, 8.8352e-110, 4.3405e-111, 2.0948e-112, 9.9319e-114, 4.6259e-115, 2.1166e-116, 9.514e-118, 4.2011e-119, 1.8224e-120, 7.7661e-122, 3.2512e-123, 1.3371e-124, 5.402e-126, 2.144e-127, 8.3597e-129, 3.202e-130, 1.2049e-131, 4.4538e-133, 1.6173e-134, 5.7697e-136, 2.022e-137, 6.9614e-139, 2.3544e-140, 7.8227e-142, 2.5533e-143, 8.1871e-145, 2.5789e-146, 7.9803e-148, 2.426e-149, 7.2448e-151, 2.1255e-152, 6.1257e-154, 1.7343e-155, 4.8239e-157, 1.3181e-158, 3.538e-160, 9.3294e-162, 2.4167e-163, 6.1502e-165, 1.5375e-166, 3.7761e-168, 9.1103e-170, 2.1593e-171, 5.0276e-173, 1.15e-174, 2.5841e-176, 5.7042e-178, 1.237e-179, 2.6352e-181, 5.5149e-183, 1.1338e-184, 2.29e-186, 4.5436e-188, 8.8561e-190, 1.6958e-191, 3.1899e-193, 5.8946e-195, 1.0701e-196, 1.9083e-198, 3.3433e-200, 5.7541e-202, 9.7287e-204, 1.6159e-205, 2.6366e-207, 4.2263e-209, 6.6552e-211, 1.0295e-212, 1.5645e-214, 2.3357e-216, 3.4256e-218, 4.9354e-220, 6.9855e-222, 9.7128e-224, 1.3267e-225, 1.7803e-227, 2.3468e-229, 3.039e-231, 3.8662e-233, 4.8318e-235, 5.9321e-237, 7.1548e-239, 8.4773e-241, 9.8673e-243, 1.1283e-244, 1.2674e-246, 1.3986e-248, 1.5162e-250, 1.6147e-252, 1.6893e-254, 1.7362e-256, 1.753e-258, 1.7388e-260, 1.6942e-262, 1.6218e-264, 1.525e-266, 1.4088e-268, 1.2785e-270, 1.1398e-272, 9.9826e-275, 8.5888e-277, 7.2594e-279, 6.0276e-281, 4.9167e-283, 3.9398e-285, 3.1014e-287, 2.3984e-289, 1.8221e-291, 1.3598e-293, 9.9699e-296, 7.1808e-298, 5.0808e-300, 3.5316e-302, 2.4115e-304, 1.6177e-306, 1.066e-308, 6.9011e-311, 4.3889e-313, 2.742e-315, 1.6829e-317, 1.0147e-319, 6.324e-322, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0};
float bottom_lookup[]={0.5f, 0.55304f, 0.60514f, 0.65542f, 0.7031f, 0.74751f, 0.78814f, 0.82468f, 0.85694f, 0.88493f, 0.90879f, 0.92877f, 0.9452f, 0.95848f, 0.96903f, 0.97725f, 0.98355f, 0.98829f, 0.9918f, 0.99435f, 0.99617f, 0.99744f, 0.99832f, 0.99892f, 0.99931f, 0.99957f, 0.99974f, 0.99984f, 0.99991f, 0.99994f, 0.99997f, 0.99998f, 0.99999f, 0.99999f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 0.99999f, 0.99999f, 0.99998f, 0.99997f, 0.99994f, 0.99991f, 0.99984f, 0.99974f, 0.99957f, 0.99931f, 0.99892f, 0.99832f, 0.99744f, 0.99617f, 0.99435f, 0.9918f, 0.98829f, 0.98355f, 0.97725f, 0.96903f, 0.95848f, 0.9452f, 0.92877f, 0.90879f, 0.88493f, 0.85694f, 0.82468f, 0.78814f, 0.74751f, 0.7031f, 0.65542f, 0.60514f, 0.55304f, 0.5f};
using namespace pcl::simulation;
// Finds the maximum level n so a and b are still
// divisible by 2^n
int
max_level (int a, int b)
{
int level = 0;
while (true)
{
if (a%2 || b%2) return level;
a = a / 2;
b = b / 2;
level++;
}
}
// display_tic_toc: a helper function which accepts a set of
// timestamps and displays the elapsed time between them as
// a fraction and time used [for profiling]
void
display_tic_toc (vector<double> &tic_toc,const string &fun_name)
{
size_t tic_toc_size = tic_toc.size ();
double percent_tic_toc_last = 0;
double dtime = tic_toc[tic_toc_size-1] - tic_toc[0];
cout << "fraction_" << fun_name << ",";
for (size_t i = 0; i < tic_toc_size; i++)
{
double percent_tic_toc = (tic_toc[i] - tic_toc[0])/(tic_toc[tic_toc_size-1] - tic_toc[0]);
cout << percent_tic_toc - percent_tic_toc_last << ", ";
percent_tic_toc_last = percent_tic_toc;
}
cout << "\ntime_" << fun_name << ",";
double time_tic_toc_last = 0;
for (size_t i = 0; i < tic_toc_size; i++)
{
double percent_tic_toc = (tic_toc[i] - tic_toc[0])/(tic_toc[tic_toc_size-1] - tic_toc[0]);
cout << percent_tic_toc*dtime - time_tic_toc_last << ", ";
time_tic_toc_last = percent_tic_toc*dtime;
}
cout << "\ntotal_time_" << fun_name << " " << dtime << "\n";
}
pcl::simulation::RangeLikelihood::RangeLikelihood (int rows, int cols, int row_height, int col_width, Scene::Ptr scene) :
scene_(scene), rows_(rows), cols_(cols), row_height_(row_height), col_width_(col_width),
depth_buffer_dirty_(true),
color_buffer_dirty_(true),
score_buffer_dirty_(true),
fbo_ (0),
depth_render_buffer_ (0),
color_render_buffer_ (0),
depth_texture_ (0),
score_texture_ (0),
score_summarized_texture_ (0),
sensor_texture_ (0),
likelihood_texture_ (0),
compute_likelihood_on_cpu_ (false),
aggregate_on_cpu_ (false),
use_instancing_ (false),
use_color_ (true),
sum_reduce_ (cols * col_width, rows * row_height, max_level (col_width, row_height))
{
height_ = rows_ * row_height;
width_ = cols_ * col_width;
depth_buffer_ = new float[width_*height_];
color_buffer_ = new uint8_t[width_*height_*3];
// Set Default Camera Intrinstic Parameters. techquad
// Correspond closely to those stated here:
// http://www.ros.org/wiki/kinect_calibration/technical
camera_width_ = 640;
camera_height_ = 480;
camera_fx_ = 576.09757860f;
camera_fy_ = 576.09757860f;
camera_cx_ = 321.06398107f;
camera_cy_ = 242.97676897f;
z_near_ = 0.7f;
z_far_ = 20.0f;
which_cost_function_ = 2; // default to commonly used meter based function
// default lhood parameters - these should always be set by the user
// so might want to add to constructor eventually:
sigma_ = 0.1;
floor_proportion_ = 0.9;
int height = rows * row_height;
int width = cols * col_width;
// For now we only support a limited size texture
assert (height >0 && height <= 8192 && width > 0 && width <= 8192);
// throw std::runtime_error "
// Allocate framebuffer
glGenRenderbuffers (1, &depth_render_buffer_);
glBindRenderbuffer (GL_RENDERBUFFER, depth_render_buffer_);
glRenderbufferStorage (GL_RENDERBUFFER, GL_DEPTH_COMPONENT32, width, height);
glBindRenderbuffer (GL_RENDERBUFFER, 0);
glGenRenderbuffers (1, &color_render_buffer_);
glBindRenderbuffer (GL_RENDERBUFFER, color_render_buffer_);
glRenderbufferStorage (GL_RENDERBUFFER, GL_RGB8UI, width, height);
glBindRenderbuffer (GL_RENDERBUFFER, 0);
// Setup texture to store depth image
glGenTextures (1, &depth_texture_);
glBindTexture (GL_TEXTURE_2D, depth_texture_);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_MODE, GL_NONE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_FUNC, GL_LEQUAL);
glTexImage2D (GL_TEXTURE_2D, 0, GL_DEPTH_COMPONENT32, width, height, 0, GL_DEPTH_COMPONENT, GL_FLOAT, NULL);
glBindTexture (GL_TEXTURE_2D, 0);
glGenTextures (1, &color_texture_);
glBindTexture (GL_TEXTURE_2D, color_texture_);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_MODE, GL_NONE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_FUNC, GL_LEQUAL);
glTexImage2D (GL_TEXTURE_2D, 0, GL_RGB, width, height, 0, GL_RGB, GL_UNSIGNED_BYTE, NULL);
glBindTexture (GL_TEXTURE_2D, 0);
// Setup texture for incoming image
glGenTextures (1, &sensor_texture_);
glBindTexture (GL_TEXTURE_2D, sensor_texture_);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_REPEAT);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_MODE, GL_NONE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_FUNC, GL_LEQUAL);
glTexImage2D (GL_TEXTURE_2D, 0, GL_R32F, col_width, row_height, 0, GL_RED, GL_FLOAT, NULL);
glBindTexture (GL_TEXTURE_2D, 0);
// Texture for to score on each pixel
glGenTextures (1, &score_texture_);
glBindTexture (GL_TEXTURE_2D, score_texture_);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_MODE, GL_NONE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_FUNC, GL_LEQUAL);
glTexImage2D (GL_TEXTURE_2D, 0, GL_R32F, width_, height_, 0, GL_RED, GL_FLOAT, NULL);
glBindTexture (GL_TEXTURE_2D, 0);
// Setup texture for likelihood function
// size of likelihood texture
int likelihood_size = 301;
glActiveTexture (GL_TEXTURE2);
glGenTextures (1, &likelihood_texture_);
glBindTexture (GL_TEXTURE_2D, likelihood_texture_);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_WRAP_R, GL_CLAMP_TO_EDGE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_MODE, GL_NONE);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_COMPARE_FUNC, GL_LEQUAL);
glTexParameteri (GL_TEXTURE_2D, GL_TEXTURE_SWIZZLE_R, GL_RED);
glTexImage2D (GL_TEXTURE_2D, 0, GL_R32F, likelihood_size, 1, 0, GL_RED, GL_FLOAT, normal_sigma0x5_normal1x0_range0to3_step0x01);
glBindTexture (GL_TEXTURE_2D, 0);
// Setup the framebuffer object for rendering
glGenFramebuffers (1, &fbo_);
glBindFramebuffer (GL_FRAMEBUFFER, fbo_);
glFramebufferTexture2D (GL_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D, depth_texture_, 0);
glFramebufferTexture2D (GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, color_texture_, 0);
glBindFramebuffer (GL_FRAMEBUFFER, 0);
if (gllib::getGLError() != GL_NO_ERROR)
{
std::cerr << "RangeLikelihoodGLSL::RangeLikelihoodGLSL: Failed initializing OpenGL buffers" << std::endl;
exit(-1);
}
glGenFramebuffers (1, &score_fbo_);
glBindFramebuffer (GL_FRAMEBUFFER, score_fbo_);
glFramebufferRenderbuffer (GL_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_RENDERBUFFER, 0);
glFramebufferTexture2D (GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, score_texture_, 0);
glBindFramebuffer (GL_FRAMEBUFFER, 0);
// Load shader
likelihood_program_ = gllib::Program::Ptr (new gllib::Program ());
// TODO: to remove file dependency include the shader source in the binary
if (!likelihood_program_->addShaderFile ("compute_score.vert", gllib::VERTEX))
{
std::cout << "Failed loading vertex shader" << std::endl;
exit (-1);
}
if (!likelihood_program_->addShaderFile ("compute_score.frag", gllib::FRAGMENT))
{
std::cout << "Failed loading fragment shader" << std::endl;
exit (-1);
}
likelihood_program_->link ();
vertices_.push_back (Eigen::Vector3f (-1.0, 1.0, 0.0));
vertices_.push_back (Eigen::Vector3f ( 1.0, 1.0, 0.0));
vertices_.push_back (Eigen::Vector3f ( 1.0, -1.0, 0.0));
vertices_.push_back (Eigen::Vector3f (-1.0, -1.0, 0.0));
glGenBuffers (1, &quad_vbo_);
glBindBuffer (GL_ARRAY_BUFFER, quad_vbo_);
glBufferData (GL_ARRAY_BUFFER, sizeof (Eigen::Vector3f) * vertices_.size (), &(vertices_[0]), GL_STATIC_DRAW);
glBindBuffer(GL_ARRAY_BUFFER, 0);
gllib::getGLError ();
// Go back to the default pipeline
glUseProgram (0);
score_buffer_ = new float[width_*height_];
}
pcl::simulation::RangeLikelihood::~RangeLikelihood ()
{
glDeleteBuffers (1, &quad_vbo_);
glDeleteTextures (1, &depth_texture_);
glDeleteTextures (1, &color_texture_);
glDeleteTextures (1, &score_texture_);
glDeleteTextures (1, &score_summarized_texture_);
glDeleteTextures (1, &sensor_texture_);
glDeleteTextures (1, &likelihood_texture_);
glDeleteFramebuffers (1, &fbo_);
glDeleteFramebuffers (1, &score_fbo_);
glDeleteRenderbuffers (1, &depth_render_buffer_);
glDeleteRenderbuffers (1, &color_render_buffer_);
delete [] depth_buffer_;
delete [] color_buffer_;
delete [] score_buffer_;
}
double
pcl::simulation::RangeLikelihood::sampleNormal (double sigma)
{
typedef boost::normal_distribution<double> Normal;
Normal dist (0.0, sigma);
boost::variate_generator<boost::minstd_rand&, Normal> norm (generator, dist);
return (norm ());
}
void
pcl::simulation::RangeLikelihood::setupProjectionMatrix ()
{
glMatrixMode (GL_PROJECTION);
glLoadIdentity ();
// Prepare scaled simulated camera projection matrix
float sx = static_cast<float> (camera_width_) / static_cast<float> (col_width_);
float sy = static_cast<float> (camera_height_) / static_cast<float> (row_height_);
float width = static_cast<float> (col_width_);
float height = static_cast<float> (row_height_);
float fx = camera_fx_/sx;
float fy = camera_fy_/sy;
float cx = camera_cx_/sx;
float cy = camera_cy_/sy;
float m[16];
float z_nf = (z_near_-z_far_);
m[0] = 2.0f*fx/width; m[4] = 0; m[ 8] = 1.0f-(2*cx/width); m[12] = 0;
m[1] = 0; m[5] = 2.0f*fy/height; m[ 9] = 1.0f-(2*cy/height); m[13] = 0;
m[2] = 0; m[6] = 0; m[10] = (z_far_+z_near_)/z_nf; m[14] = 2.0f*z_near_*z_far_/z_nf;
m[3] = 0; m[7] = 0; m[11] = -1.0f; m[15] = 0;
glMultMatrixf (m);
}
void
pcl::simulation::RangeLikelihood::applyCameraTransform (const Eigen::Isometry3d & pose)
{
float T[16];
Eigen::Matrix4f m = (pose.matrix ().inverse ()).cast<float> ();
T[0] = m(0,0); T[4] = m(0,1); T[8] = m(0,2); T[12] = m(0,3);
T[1] = m(1,0); T[5] = m(1,1); T[9] = m(1,2); T[13] = m(1,3);
T[2] = m(2,0); T[6] = m(2,1); T[10] = m(2,2); T[14] = m(2,3);
T[3] = m(3,0); T[7] = m(3,1); T[11] = m(3,2); T[15] = m(3,3);
glMultMatrixf(T);
}
void
pcl::simulation::RangeLikelihood::drawParticles (std::vector<Eigen::Isometry3d, Eigen::aligned_allocator<Eigen::Isometry3d> > poses)
{
int n = 0;
for (int i=0; i<rows_; ++i)
{
for (int j=0; j<cols_; ++j)
{
glMatrixMode (GL_MODELVIEW);
glLoadIdentity ();
glViewport(j*col_width_, i*row_height_, col_width_, row_height_);
// Go from Z-up, X-forward coordinate frame
// to OpenGL Z-out,Y-up [both Right Handed]
float T[16];
T[0] = 0; T[4] = -1.0; T[8] = 0; T[12] = 0;
T[1] = 0; T[5] = 0; T[9] = 1; T[13] = 0;
T[2] = -1.0; T[6] = 0; T[10] = 0; T[14] = 0;
T[3] = 0; T[7] = 0; T[11] = 0; T[15] = 1;
glMultMatrixf (T);
// Apply camera transformation
applyCameraTransform (poses[n++]);
// Draw the planes in each location:
scene_->draw ();
}
}
}
/////////////////////////////////////////////////////////////////
// Below are 4 previously used cost functions:
// 0 original scoring method
float
costFunction0 (float ref_val,float depth_val)
{
return (sqr(ref_val - depth_val));
}
// 1st working cost function:
// Empirical reverse mapping between depthbuffer and true depth:
// Version 0: [25 aug 2011]
// TRUEDEPTH = 1/(1.33 -(DEPTHBUFFER)*1.29)
//float cost = sqr(ref[col%col_width] - 1/(1.33 -(*depth)*1.29));
// Version 1: [29 aug 2011] Exact version using correct mappings:
float
costFunction1 (float ref_val, float depth_val)
{
float cost = sqr (ref_val - 1/(1.4285f - (depth_val)*1.3788f));
//std::cout << " [" << ref_val << "," << 1/(1.4285 -(depth_val)*1.3788) << "] ";
if (ref_val < 0)
{ // all images pixels with no range
cost =1;
}
if (cost > 10)
{ // required to lessen the effect of modelpixel with no range (ie holes in the model)
cost = 10;
}
//return log (cost);
return (cost);
}
// 1st working likelihood function (by far most commonly used)
float
costFunction2 (float ref_val, float depth_val)
{
float min_dist = abs(ref_val - 1/(1.4285f - (depth_val)*1.3788f));
int lup = static_cast<int> (ceil (min_dist*100)); // has resolution of 0.01m
if (lup > 300)
{ // implicitly this caps the cost if there is a hole in the model
lup = 300;
}
double lhood = 1;
if (pcl_isnan (depth_val))
{ // pixels with nan depth - for openNI null points
lhood = 1; // log(1) = 0 ---> has no effect
}
else if(ref_val < 0)
{ // all RGB pixels with no depth - for freenect null points
lhood = 1; // log(1) = 0 ---> has no effect
}
else
{
lhood = normal_sigma0x5_normal1x0_range0to3_step0x01[lup];
// add a ground floor:
// increasing this will mean that the likelihood is less peaked
// but you need more particles to do this...
// with ~90particles user 0.999, for example in the quad dataset
// ratio of uniform to normal
double ratio = 0.99;//was always 0.99;
double r_min = 0; // metres
double r_max = 3; // metres
lhood = ratio/(r_max -r_min) + (1-ratio)*lhood ;
}
return static_cast<float> (log (lhood));
}
float
costFunction3 (float ref_val,float depth_val)
{
float log_lhood=0;
// log(1) = 0 ---> has no effect
if (ref_val < 0)
{
// all images pixels with no range
}
else if (ref_val > 7)
{
// ignore long ranges... for now
}
else
{ // working range
float min_dist = abs (ref_val - 0.7253f/(1.0360f - (depth_val)));
int lup = static_cast<int> (ceil (min_dist*100)); // has resulution of 0.01m
if (lup > 300)
{ // implicitly this caps the cost if there is a hole in the model
lup = 300;
}
log_lhood = hard_coded_log_lhood[lup];
}
return log_lhood;
}
float
costFunction4(float ref_val,float depth_val)
{
float disparity_diff = abs( ( -0.7253f/ref_val +1.0360f ) - depth_val );
int top_lup = static_cast<int> (ceil (disparity_diff*300)); // has resulution of 0.001m
if (top_lup > 300)
{
top_lup =300;
}
float top = static_cast<float> (top_lookup[top_lup]);// round( abs(x-mu) *1000+1) );
// bottom:
//bottom = bottom_lookup( round(mu*1000+1));
int bottom_lup = static_cast<int> (ceil( (depth_val) * 300)); // has resulution of 0.001m
if (bottom_lup > 300)
{
bottom_lup =300;
}
float bottom = bottom_lookup[bottom_lup];// round( abs(x-mu) *1000+1) );
float proportion = 0.999f;
float lhood = proportion + (1-proportion)*(top/bottom);
// safety fix thats seems to be required due to opengl ayschronizate
// ask hordur about this
if (bottom == 0)
{
lhood = proportion;
}
if (ref_val< 0)
{ // all images pixels with no range
lhood = 1; // log(1) = 0 ---> has no effect
}
return log(lhood);
}
// TODO: WHEN WE'RE HAPPY THIS SHOULD BE "THE" LIKELIHOOD FUNCTION
// add these global variables into the class
// abd use sigma and floor_proportion directly from class also
using boost::math::normal; // typedef provides default type is double.
normal unit_norm_dist(0,1); // (default mean = zero, and standard deviation = unity)
double costFunction5(double measured_depth, double model_disp, double sigma, double floor_proportion)
{
// NEED TO CONVERT MEASURED TO DISPARITY
double measured_disp = (-0.7253/measured_depth + 1.0360 );
// measured_depth = ref_val [m]
// model_disp = depth_val [0-1]
// upper and lower bound on depth buffer:
double lower_bound =0;
double upper_bound =1;
double gaussian_part = pdf(unit_norm_dist, (measured_disp-model_disp)/sigma)/sigma;
double truncation = 1/cdf(unit_norm_dist,(upper_bound-model_disp)/sigma) - cdf(unit_norm_dist, (lower_bound-model_disp)/sigma);
double trunc_gaussian_part = truncation*gaussian_part;
double lhood= (floor_proportion/(upper_bound-lower_bound) + (1-floor_proportion)*trunc_gaussian_part);
if (measured_depth< 0){ // all images pixels with no range
lhood = 1; // log(1) = 0 ---> has no effect
}
return log (lhood);
}
void
pcl::simulation::RangeLikelihood::computeScores (float* reference,
std::vector<float> & scores)
{
const float* depth = getDepthBuffer();
// Mapping between disparity and range:
// range or depth = 1/disparity
//
// the freenect produces a disparity <here we call this depth_buffer>
// that is mapped between 0->1 to minimize quantization
// near_range = n = 0.7m | far_range = f = 20m
// disparity can be found as a linear function of the depth_buffer (d = [0,1] )
// disparity = 1/n - (f-n)*d / (n*f)
// Below We compare range-versus-range using this mapping
//
// TODO: remove usage of 'depth' and 'depth_buffer_' as variable names as it implies
// that that is was held by these variables
// ref[col%col_width] - z/depth value in metres, 0-> ~20
// depth_val - contents of depth buffer [0->1]
// for row across each image in a row of model images
for (int row = 0; row < rows_*row_height_; row++)
{
float* ref = reference + col_width_*(row % row_height_);
// for each column: across each image in a column of model images
for (int col = 0; col < cols_*col_width_; col++)
{
float depth_val = (*depth++); // added jan 2012 - check this is not a breaking fix later mfallon
float score = 0;
if (which_cost_function_ == 0)
{
score = costFunction0 (ref[col%col_width_],depth_val);
}
else if (which_cost_function_ == 1)
{
score = costFunction1 (ref[col%col_width_],depth_val);
}
else if (which_cost_function_ == 2)
{
score = costFunction2 (ref[col%col_width_],depth_val);
}
else if(which_cost_function_==3)
{
score = costFunction3 (ref[col%col_width_],depth_val);
}
else if (which_cost_function_ == 4)
{
score = costFunction4 (ref[col%col_width_],depth_val);
}
else if (which_cost_function_ == 5)
{
//double sigma = 0.025;
//double floor_proportion_ = 0.999;
score = static_cast<float> (costFunction5 (ref[col%col_width_],depth_val,sigma_,floor_proportion_));
}
scores[row/row_height_ * cols_ + col/col_width_] += score;
//std::cout << "(" << scores[row/row_height_ * cols_ + col/col_width_] <<"," << score << "," << ref[col%col_width_] << "," << depth_val << ") ";
score_buffer_[row*width_ + col] = score;
}
}
score_buffer_dirty_ = false;
}
void
pcl::simulation::RangeLikelihood::getPointCloud (pcl::PointCloud<pcl::PointXYZRGB>::Ptr pc,
bool make_global,
const Eigen::Isometry3d & pose)
{
// TODO: check if this works for for rows/cols >1 and for width&height != 640x480
// i.e. multiple tiled images
pc->width = col_width_;
pc->height = row_height_;
// Was:
//pc->width = camera_width_;
//pc->height = camera_height_;
pc->is_dense = false;
pc->points.resize (pc->width*pc->height);
int points_added = 0;
float camera_fx_reciprocal_ = 1.0f / camera_fx_;
float camera_fy_reciprocal_ = 1.0f / camera_fy_;
float zn = z_near_;
float zf = z_far_;
const uint8_t* color_buffer = getColorBuffer();
// TODO: support decimation
// Copied the format of RangeImagePlanar::setDepthImage()
// Use this as a template for decimation
for (int y = 0; y < row_height_ ; ++y) //camera_height_
{
for (int x = 0; x < col_width_ ; ++x) // camera_width_
{
// Find XYZ from normalized 0->1 mapped disparity
int idx = points_added; // y*camera_width_ + x;
float d = depth_buffer_[y*camera_width_ + x] ;
if (d < 1.0) // only add points with depth buffer less than max (20m) range
{
float z = zf*zn/((zf-zn)*(d - zf/(zf-zn)));
// TODO: add mode to ignore points with no return i.e. depth_buffer_ ==1
// NB: OpenGL uses a Right Hand system with +X right, +Y up, +Z back out of the screen,
// The Z-buffer is natively -1 (far) to 1 (near)
// But in this class we invert this to be 0 (near, 0.7m) and 1 (far, 20m)
// ... so by negating y we get to a right-hand computer vision system
// which is also used by PCL and OpenNi
pc->points[idx].z = z;
pc->points[idx].x = (static_cast<float> (x)-camera_cx_) * z * (-camera_fx_reciprocal_);
pc->points[idx].y = (static_cast<float> (y)-camera_cy_) * z * (-camera_fy_reciprocal_);
int rgb_idx = y*col_width_ + x; //camera_width_
pc->points[idx].b = color_buffer[rgb_idx*3+2]; // blue
pc->points[idx].g = color_buffer[rgb_idx*3+1]; // green
pc->points[idx].r = color_buffer[rgb_idx*3]; // red
points_added++;
}
}
}
pc->width = 1;
pc->height = points_added;
pc->points.resize (points_added);
if (make_global)
{
// Go from OpenGL to (Z-up, X-forward, Y-left)
Eigen::Matrix4f T;
T << 0, 0, -1, 0,
-1, 0, 0, 0,
0, 1, 0, 0,
0, 0, 0, 1;
Eigen::Matrix4f m = pose.matrix ().cast<float> ();
m = m * T;
pcl::transformPointCloud (*pc, *pc, m);
}
else
{
// Go from OpenGL to Camera (Z-forward, X-right, Y-down)
Eigen::Matrix4f T;
T << 1, 0, 0, 0,
0, -1, 0, 0,
0, 0, -1, 0,
0, 0, 0, 1;
pcl::transformPointCloud (*pc, *pc, T);
// Go from Camera to body (Z-up, X-forward, Y-left)
Eigen::Matrix4f cam_to_body;
cam_to_body << 0, 0, 1, 0,
-1, 0, 0, 0,
0, -1, 0, 0,
0, 0, 0, 1;
Eigen::Matrix4f camera = pose.matrix ().cast<float> ();
camera = camera * cam_to_body;
pc->sensor_origin_ = camera.rightCols (1);
Eigen::Quaternion<float> quat (camera.block<3,3> (0,0));
pc->sensor_orientation_ = quat;
}
}
void
pcl::simulation::RangeLikelihood::getRangeImagePlanar(pcl::RangeImagePlanar &rip)
{
rip.setDepthImage (depth_buffer_,
camera_width_,camera_height_, camera_fx_,camera_fy_,
camera_fx_, camera_fy_);
}
void
pcl::simulation::RangeLikelihood::addNoise ()
{
// Other noises:
// edge noise: look for edges in the range image and add a few pixels here and there
// texture noise: look at the normals and
// Add Gaussian Noise
// TODO: make the variance a parameter
// TODO: might want to add a range-based variance
float variance = 0.0015f;
for (int i = 0; i < camera_width_*camera_height_ ; i++)
{
if (depth_buffer_[i] < 1)
{
depth_buffer_[i] = depth_buffer_[i] + variance * static_cast<float> (sampleNormal ());
if (depth_buffer_[i] > 1)
{
depth_buffer_[i] = 1.0;
}
else if (depth_buffer_[i] < 0)
{
depth_buffer_[i] = 0.0;
}
}
}
// Add Kinect/Primesense Quantisation Noise:
// TODO: better fit this:
// 0.6m = ~600 kinect return
// 20m = ~1070 kinect return - not not well calibrated
// The fitted model stated here cannot work for long ranges:
// http://www.ros.org/wiki/kinect_calibration/technical
// TODO: make a parameter
float bins = 470;
for (int i = 0; i < camera_width_*camera_height_ ; i++)
{
depth_buffer_[i] = ceil (depth_buffer_[i]*bins)/bins;
}
cout << "in add noise\n";
}
void
RangeLikelihood::computeLikelihoods (float* reference,
std::vector<Eigen::Isometry3d, Eigen::aligned_allocator<Eigen::Isometry3d> > poses,
std::vector<float> & scores)
{
#if DO_TIMING_PROFILE
vector<double> tic_toc;
tic_toc.push_back(getTime());
#endif
scores.resize (cols_*rows_);
std::fill (scores.begin (), scores.end (), 0);
// Generate depth image for each particle
render (poses);
#if DO_TIMING_PROFILE
tic_toc.push_back (getTime ());
#endif
#if DO_TIMING_PROFILE
tic_toc.push_back (getTime ());
#endif
// The depth image is now in depth_texture_
// Compute likelihoods
if (compute_likelihood_on_cpu_)
{
computeScores (reference, scores);
}
else
{
computeScoresShader (reference);
// Aggregate results (we do not use GPU to sum cpu scores)
if (aggregate_on_cpu_)
{
const float* score_buffer = getScoreBuffer();
for (int n = 0, row = 0; row < height_; ++row)
{
for (int col = 0; col < width_; ++col, ++n)
{
scores[row/row_height_ * cols_ + col/col_width_] += score_buffer[n];
}
}
}
else
{
int levels = max_level (row_height_, col_width_);
int reduced_width = width_ >> levels;
int reduced_height = height_ >> levels;
int reduced_col_width = col_width_ >> levels;
int reduced_row_height = row_height_ >> levels;
float* score_sum = new float[reduced_width * reduced_height];
sum_reduce_.sum (score_texture_, score_sum);
for (int n = 0, row = 0; row < reduced_height; ++row)
{
for (int col = 0; col < reduced_width; ++col, ++n)
{
scores[row/reduced_row_height * cols_ + col/reduced_col_width] += score_sum[n];
}
}
delete [] score_sum;
}
}
#if DO_TIMING_PROFILE
tic_toc.push_back (getTime ());
display_tic_toc (tic_toc, "range_likelihood");
#endif
}
// Computes the likelihood scores using a shader
void
pcl::simulation::RangeLikelihood::computeScoresShader (float* reference)
{
if (gllib::getGLError () != GL_NO_ERROR)
{
std::cout << "GL error: RangeLikelihood::compute_scores_shader - enter" << std::endl;
}
#ifdef SIMULATION_DEBUG
std::cout << "DepthSampler location: " << likelihood_program_->getUniformLocation ("DepthSampler") << std::endl;
std::cout << "ReferenceSampler location: " << likelihood_program_->getUniformLocation ("ReferenceSampler") << std::endl;
std::cout << "CostSampler location: " << likelihood_program_->getUniformLocation ("CostSampler") << std::endl;
int depth_tex_id;
int ref_tex_id;
int cost_tex_id;
glGetUniformiv(likelihood_program_->programId (), likelihood_program_->getUniformLocation ("DepthSampler"), &depth_tex_id);
glGetUniformiv(likelihood_program_->programId (), likelihood_program_->getUniformLocation ("ReferenceSampler"), &ref_tex_id);
glGetUniformiv(likelihood_program_->programId (), likelihood_program_->getUniformLocation ("CostSampler"), &cost_tex_id);
std::cout << "depth id: " << depth_tex_id << " " << GL_TEXTURE0 << std::endl;
std::cout << "ref id: " << ref_tex_id << " " << GL_TEXTURE1 << std::endl;
std::cout << "cost id: " << cost_tex_id << " " << GL_TEXTURE2 << std::endl;
#endif
likelihood_program_->use ();
likelihood_program_->setUniform ("DepthSampler", 0);
likelihood_program_->setUniform ("ReferenceSampler", 1);
likelihood_program_->setUniform ("CostSampler", 2);
likelihood_program_->setUniform ("cols", cols_);
likelihood_program_->setUniform ("rows", rows_);
likelihood_program_->setUniform ("near", z_near_);
likelihood_program_->setUniform ("far", z_far_);
glBindFramebuffer (GL_FRAMEBUFFER, score_fbo_);
glDrawBuffer (GL_COLOR_ATTACHMENT0);
glReadBuffer (GL_NONE);
GLboolean enable_depth_test;
glGetBooleanv (GL_DEPTH_TEST, &enable_depth_test);
glDisable (GL_DEPTH_TEST);
glViewport (0, 0, width_, height_);
// Setup textures
glActiveTexture (GL_TEXTURE0);
glBindTexture (GL_TEXTURE_2D, depth_texture_);
glActiveTexture (GL_TEXTURE1);
glBindTexture (GL_TEXTURE_2D, sensor_texture_);
glTexImage2D (GL_TEXTURE_2D, 0, GL_R32F, col_width_, row_height_, 0, GL_RED, GL_FLOAT, reference);
glActiveTexture (GL_TEXTURE2);
glBindTexture (GL_TEXTURE_2D, likelihood_texture_);
quad_.render ();
glUseProgram (0);
glBindFramebuffer (GL_FRAMEBUFFER, 0);
// Unbind all textures that were used
glActiveTexture (GL_TEXTURE0);
glBindTexture (GL_TEXTURE_2D, 0);
glActiveTexture (GL_TEXTURE1);
glBindTexture (GL_TEXTURE_2D, 0);
glActiveTexture (GL_TEXTURE2);
glBindTexture (GL_TEXTURE_2D, 0);
if (gllib::getGLError () != GL_NO_ERROR)
{
std::cout << "GL error: RangeLikelihood::compute_scores_shader - exit" << std::endl;
}
if (enable_depth_test == GL_TRUE) glEnable (GL_DEPTH_TEST);
}
void
RangeLikelihood::render (const std::vector<Eigen::Isometry3d, Eigen::aligned_allocator<Eigen::Isometry3d> > & poses)
{
if (gllib::getGLError () != GL_NO_ERROR)
{
std::cerr << "GL Error: RangeLikelihood::render - enter" << std::endl;
}
GLint old_matrix_mode;
GLint old_draw_buffer;
GLint old_read_buffer;
glGetIntegerv (GL_DRAW_BUFFER, &old_draw_buffer);
glGetIntegerv (GL_READ_BUFFER, &old_read_buffer);
glGetIntegerv (GL_MATRIX_MODE, &old_matrix_mode);
glMatrixMode (GL_PROJECTION);
glPushMatrix ();
glMatrixMode (GL_MODELVIEW);
glPushMatrix ();
glBindFramebuffer (GL_FRAMEBUFFER, fbo_);
if (use_color_)
{
glDrawBuffer (GL_COLOR_ATTACHMENT0);
}
else
{
glDrawBuffer (GL_NONE);
}
glReadBuffer (GL_NONE);
GLenum status;
status = glCheckFramebufferStatus (GL_FRAMEBUFFER);
switch (status)
{
case GL_FRAMEBUFFER_COMPLETE:
{
break;
}
default:
{
std::cout << "RangeLikelihood::render: Framebuffer failed" << std::endl;
exit (-1);
}
}
// Render
glPushAttrib (GL_ALL_ATTRIB_BITS);
glEnable (GL_COLOR_MATERIAL);
glClearColor (0.0f, 0.0f, 0.0f, 0.0f);
glClearDepth (1.0);