-
Notifications
You must be signed in to change notification settings - Fork 29
/
converter.py
607 lines (505 loc) · 26.8 KB
/
converter.py
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
import os
import numpy as np
import cv2
import matplotlib.pyplot as plt
import tensorflow as tf
import tqdm
from multiprocessing import Pool
from os.path import join, isdir
import argparse
from glob import glob
from waymo_open_dataset.utils.frame_utils import parse_range_image_and_camera_projection
from waymo_open_dataset import dataset_pb2 as open_dataset
from waymo_open_dataset import dataset_pb2
from waymo_open_dataset.utils import range_image_utils
from waymo_open_dataset.utils import transform_utils
# Abbreviations:
# WOD: Waymo Open Dataset
# FOV: field of view
# SDC: self-driving car
# 3dbox: 3D bounding box
# Some 3D bounding boxes do not contain any points
# This switch, when set True, filters these boxes
# It is safe to filter these boxes because they are not counted towards evaluation anyway
filter_empty_3dboxes = False
# There is no bounding box annotations in the No Label Zone (NLZ)
# if set True, points in the NLZ are filtered
filter_no_label_zone_points = True
# Only bounding boxes of certain classes are converted
# Note: Waymo Open Dataset evaluates for ALL_NS, including only 'VEHICLE', 'PEDESTRIAN', 'CYCLIST'
selected_waymo_classes = [
# 'UNKNOWN',
'VEHICLE',
'PEDESTRIAN',
# 'SIGN',
'CYCLIST'
]
# Only data collected in specific locations will be converted
# If set None, this filter is disabled (all data will thus be converted)
# Available options: location_sf (main dataset)
selected_waymo_locations = None
# Save track id
save_track_id = True
# DATA_PATH = '/media/alex/Seagate Expansion Drive/waymo_open_dataset/domain_adaptation_training_labelled(partial)'
# KITTI_PATH = '/home/alex/github/waymo_to_kitti_converter/tools/pose'
class WaymoToKITTI(object):
def __init__(self, load_dir, save_dir, prefix, num_proc):
# turn on eager execution for older tensorflow versions
if int(tf.__version__.split('.')[0]) < 2:
tf.enable_eager_execution()
self.lidar_list = ['_FRONT', '_FRONT_RIGHT', '_FRONT_LEFT', '_SIDE_RIGHT', '_SIDE_LEFT']
self.type_list = ['UNKNOWN', 'VEHICLE', 'PEDESTRIAN', 'SIGN', 'CYCLIST']
self.waymo_to_kitti_class_map = {
'UNKNOWN': 'DontCare',
'PEDESTRIAN': 'Pedestrian',
'VEHICLE': 'Car',
'CYCLIST': 'Cyclist',
'SIGN': 'Sign' # not in kitti
}
self.load_dir = load_dir
self.save_dir = save_dir
self.prefix = prefix
self.num_proc = int(num_proc)
self.tfrecord_pathnames = sorted(glob(join(self.load_dir, '*.tfrecord')))
self.label_save_dir = self.save_dir + '/label_'
self.label_all_save_dir = self.save_dir + '/label_all'
self.image_save_dir = self.save_dir + '/image_'
self.calib_save_dir = self.save_dir + '/calib'
self.point_cloud_save_dir = self.save_dir + '/velodyne'
self.pose_save_dir = self.save_dir + '/pose'
self.create_folder()
def convert(self):
print("start converting ...")
with Pool(self.num_proc) as p:
r = list(tqdm.tqdm(p.imap(self.convert_one, range(len(self))), total=len(self)))
print("\nfinished ...")
def convert_one(self, file_idx):
pathname = self.tfrecord_pathnames[file_idx]
dataset = tf.data.TFRecordDataset(pathname, compression_type='')
for frame_idx, data in enumerate(dataset):
frame = open_dataset.Frame()
frame.ParseFromString(bytearray(data.numpy()))
if selected_waymo_locations is not None and frame.context.stats.location not in selected_waymo_locations:
continue
# save images
self.save_image(frame, file_idx, frame_idx)
# parse calibration files
self.save_calib(frame, file_idx, frame_idx)
# parse point clouds
self.save_lidar(frame, file_idx, frame_idx)
# parse label files
self.save_label(frame, file_idx, frame_idx)
# parse pose files
self.save_pose(frame, file_idx, frame_idx)
def __len__(self):
return len(self.tfrecord_pathnames)
def save_image(self, frame, file_idx, frame_idx):
""" parse and save the images in png format
:param frame: open dataset frame proto
:param file_idx: the current file number
:param frame_idx: the current frame number
:return:
"""
for img in frame.images:
img_path = self.image_save_dir + str(img.name - 1) + '/' + self.prefix + str(file_idx).zfill(3) + str(frame_idx).zfill(3) + '.png'
img = cv2.imdecode(np.frombuffer(img.image, np.uint8), cv2.IMREAD_COLOR)
rgb_img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
plt.imsave(img_path, rgb_img, format='png')
def save_calib(self, frame, file_idx, frame_idx):
""" parse and save the calibration data
:param frame: open dataset frame proto
:param file_idx: the current file number
:param frame_idx: the current frame number
:return:
"""
# kitti:
# bbox in reference camera frame (right-down-front)
# image_x_coord = Px * R0_rect * R0_rot * bbox_coord
# lidar points in lidar frame (front-right-up)
# image_x_coord = Px * R0_rect * Tr_velo_to_cam * lidar_coord
# note: R0_rot is caused by bbox rotation
# Tr_velo_to_cam projects lidar points to cam_0 frame
# waymo:
# bbox in vehicle frame, hence, use a virtual reference frame
# since waymo camera uses frame front-left-up, the virtual reference frame (right-down-front) is
# built on a transformed front camera frame, name this transform T_front_cam_to_ref
# and there is no rectified camera frame
# image_x_coord = intrinsics_x * Tr_front_cam_to_cam_x * inv(T_front_cam_to_ref) * R0_rot * bbox_coord(now in ref frame)
# lidar points in vehicle frame
# image_x_coord = intrinsics_x * Tr_front_cam_to_cam_x * inv(T_front_cam_to_ref) * T_front_cam_to_ref * Tr_velo_to_front_cam * lidar_coord
# hence, waymo -> kitti:
# set Tr_velo_to_cam = T_front_cam_to_ref * Tr_vehicle_to_front_cam = T_front_cam_to_ref * inv(Tr_front_cam_to_vehicle)
# as vehicle and lidar use the same frame after fusion
# set R0_rect = identity
# set P2 = front_cam_intrinsics * Tr_waymo_to_conv * Tr_front_cam_to_front_cam * inv(T_front_cam_to_ref)
# note: front cam is cam_0 in kitti, whereas has name = 1 in waymo
# note: waymo camera has a front-left-up frame,
# instead of the conventional right-down-front frame
# Tr_waymo_to_conv is used to offset this difference. However, Tr_waymo_to_conv is the same as
# T_front_cam_to_ref, hence,
# set P2 = front_cam_intrinsics
calib_context = ''
# front-left-up -> right-down-front
# T_front_cam_to_ref = np.array([
# [0.0, -1.0, 0.0],
# [-1.0, 0.0, 0.0],
# [0.0, 0.0, 1.0]
# ])
T_front_cam_to_ref = np.array([
[0.0, -1.0, 0.0],
[0.0, 0.0, -1.0],
[1.0, 0.0, 0.0]
])
# T_ref_to_front_cam = np.array([
# [0.0, 0.0, 1.0],
# [-1.0, 0.0, 0.0],
# [0.0, -1.0, 0.0]
# ])
# print('context\n',frame.context)
for camera in frame.context.camera_calibrations:
if camera.name == 1: # FRONT = 1, see dataset.proto for details
T_front_cam_to_vehicle = np.array(camera.extrinsic.transform).reshape(4, 4)
# print('T_front_cam_to_vehicle\n', T_front_cam_to_vehicle)
T_vehicle_to_front_cam = np.linalg.inv(T_front_cam_to_vehicle)
front_cam_intrinsic = np.zeros((3, 4))
front_cam_intrinsic[0, 0] = camera.intrinsic[0]
front_cam_intrinsic[1, 1] = camera.intrinsic[1]
front_cam_intrinsic[0, 2] = camera.intrinsic[2]
front_cam_intrinsic[1, 2] = camera.intrinsic[3]
front_cam_intrinsic[2, 2] = 1
break
# print('front_cam_intrinsic\n', front_cam_intrinsic)
self.T_front_cam_to_ref = T_front_cam_to_ref.copy()
self.T_vehicle_to_front_cam = T_vehicle_to_front_cam.copy()
identity_3x4 = np.eye(4)[:3, :]
# although waymo has 5 cameras, for compatibility, we produces 4 P
for i in range(4):
if i == 2:
# note: front camera is labeled camera 2 (kitti) or camera 0 (waymo)
# other Px are given dummy values. this is to ensure compatibility. They are seldom used anyway.
# tmp = cart_to_homo(np.linalg.inv(T_front_cam_to_ref))
# print(front_cam_intrinsic.shape, tmp.shape)
# P2 = np.matmul(front_cam_intrinsic, tmp).reshape(12)
P2 = front_cam_intrinsic.reshape(12)
calib_context += "P2: " + " ".join(['{}'.format(i) for i in P2]) + '\n'
else:
calib_context += "P" + str(i) + ": " + " ".join(['{}'.format(i) for i in identity_3x4.reshape(12)]) + '\n'
calib_context += "R0_rect" + ": " + " ".join(['{}'.format(i) for i in np.eye(3).astype(np.float32).flatten()]) + '\n'
Tr_velo_to_cam = self.cart_to_homo(T_front_cam_to_ref) @ np.linalg.inv(T_front_cam_to_vehicle)
# print('T_front_cam_to_vehicle\n', T_front_cam_to_vehicle)
# print('np.linalg.inv(T_front_cam_to_vehicle)\n', np.linalg.inv(T_front_cam_to_vehicle))
# print('cart_to_homo(T_front_cam_to_ref)\n', cart_to_homo(T_front_cam_to_ref))
# print('Tr_velo_to_cam\n',Tr_velo_to_cam)
calib_context += "Tr_velo_to_cam" + ": " + " ".join(['{}'.format(i) for i in Tr_velo_to_cam[:3, :].reshape(12)]) + '\n'
with open(self.calib_save_dir + '/' + self.prefix + str(file_idx).zfill(3) + str(frame_idx).zfill(3) + '.txt', 'w+') as fp_calib:
fp_calib.write(calib_context)
def save_lidar(self, frame, file_idx, frame_idx):
""" parse and save the lidar data in psd format
:param frame: open dataset frame proto
:param file_idx: the current file number
:param frame_idx: the current frame number
:return:
"""
range_images, camera_projections, range_image_top_pose = parse_range_image_and_camera_projection(frame)
points_0, cp_points_0, intensity_0 = self.convert_range_image_to_point_cloud(
frame,
range_images,
camera_projections,
range_image_top_pose,
ri_index=0
)
points_0 = np.concatenate(points_0, axis=0)
intensity_0 = np.concatenate(intensity_0, axis=0)
points_1, cp_points_1, intensity_1 = self.convert_range_image_to_point_cloud(
frame,
range_images,
camera_projections,
range_image_top_pose,
ri_index=1
)
points_1 = np.concatenate(points_1, axis=0)
intensity_1 = np.concatenate(intensity_1, axis=0)
points = np.concatenate([points_0, points_1], axis=0)
# print('points_0', points_0.shape, 'points_1', points_1.shape, 'points', points.shape)
intensity = np.concatenate([intensity_0, intensity_1], axis=0)
# points = points_1
# intensity = intensity_1
# reference frame:
# front-left-up (waymo) -> right-down-front(kitti)
# lidar frame:
# ?-?-up (waymo) -> front-right-up (kitti)
# print('bef\n', points)
# print('bef\n', points.dtype)
# points = np.transpose(points) # (n, 3) -> (3, n)
# tf = np.array([
# [0.0, -1.0, 0.0],
# [0.0, 0.0, -1.0],
# [1.0, 0.0, 0.0]
# ])
# points = np.matmul(tf, points)
# points = np.transpose(points) # (3, n) -> (n, 3)
# print('aft\n', points)
# print('aft\n', points.dtype)
# concatenate x,y,z and intensity
point_cloud = np.column_stack((points, intensity))
# print(point_cloud.shape)
# save
pc_path = self.point_cloud_save_dir + '/' + self.prefix + str(file_idx).zfill(3) + str(frame_idx).zfill(3) + '.bin'
point_cloud.astype(np.float32).tofile(pc_path) # note: must save as float32, otherwise loading errors
def save_label(self, frame, file_idx, frame_idx):
""" parse and save the label data in .txt format
:param frame: open dataset frame proto
:param file_idx: the current file number
:param frame_idx: the current frame number
:return:
"""
fp_label_all = open(self.label_all_save_dir + '/' + self.prefix + str(file_idx).zfill(3) + str(frame_idx).zfill(3) + '.txt', 'w+')
# preprocess bounding box data
id_to_bbox = dict()
id_to_name = dict()
for labels in frame.projected_lidar_labels:
name = labels.name
for label in labels.labels:
# waymo: bounding box origin is at the center
# TODO: need a workaround as bbox may not belong to front cam
bbox = [label.box.center_x - label.box.length / 2, label.box.center_y - label.box.width / 2,
label.box.center_x + label.box.length / 2, label.box.center_y + label.box.width / 2]
id_to_bbox[label.id] = bbox
id_to_name[label.id] = name - 1
# print([i.type for i in frame.laser_labels])
for obj in frame.laser_labels:
# calculate bounding box
bounding_box = None
name = None
id = obj.id
for lidar in self.lidar_list:
if id + lidar in id_to_bbox:
bounding_box = id_to_bbox.get(id + lidar)
name = str(id_to_name.get(id + lidar))
break
# TODO: temp fix
if bounding_box == None or name == None:
name = '0'
bounding_box = (0, 0, 0, 0)
my_type = self.type_list[obj.type]
if my_type not in selected_waymo_classes:
continue
if filter_empty_3dboxes and obj.num_lidar_points_in_box < 1:
continue
# from waymo_open_dataset.utils.box_utils import compute_num_points_in_box_3d
# print('annot:', obj.num_lidar_points_in_box)
# num_points_in_gt_waymo = compute_num_points_in_box_3d(
# tf.convert_to_tensor(self.pc.astype(np.float32), dtype=tf.float32),
# tf.convert_to_tensor(np.array([[obj.box.center_x, obj.box.center_y, obj.box.center_z, obj.box.length,obj.box.width, obj.box.height,obj.box.heading]]).astype(np.float32), dtype=tf.float32))
# print('actual:', num_points_in_gt_waymo.numpy())
# visualizer
# [261 56 24 15 46 254 24 824 146 26 5 13 30 45
# 60 184 347 222 1774 2 46]
# converter
# 264, 59, 24, 16, 51, 268, 24, 847, 149, 28, 6, 13, 30, 45, \
# 64, 192, 353, 229, 1848, 2, 48
my_type = self.waymo_to_kitti_class_map[my_type]
# length: along the longer axis that is perpendicular to gravity direction
# width: along the shorter axis that is perpendicular to gravity direction
# height: along the gravity direction
# the same for waymo and kitti
height = obj.box.height # up/down
width = obj.box.width # left/right
length = obj.box.length # front/back
# waymo: bbox label in lidar/vehicle frame. kitti: bbox label in reference image frame
# however, kitti uses bottom center as the box origin, whereas waymo uses the true center
x = obj.box.center_x
y = obj.box.center_y
z = obj.box.center_z - height / 2
# print('bef', x,y,z)
# project bounding box to the virtual reference frame
pt_ref = self.cart_to_homo(self.T_front_cam_to_ref) @ self.T_vehicle_to_front_cam @ np.array([x,y,z,1]).reshape((4,1))
x, y, z, _ = pt_ref.flatten().tolist()
# print('aft', x,y,z)
# x, y, z correspond to l, w, h (waymo) -> l, h, w (kitti)
# length, width, height = length, height, width
# front-left-up (waymo) -> right-down-front(kitti)
# bbox origin at volumetric center (waymo) -> bottom center (kitti)
# x, y, z = -waymo_y, -waymo_z + height / 2, waymo_x
# rotation: +x around y-axis (kitti) -> +x around y-axis (waymo)
# right-down-front front-left-up
# note: the "rotation_y" is kept as the name of the rotation variable for compatibility
# it is, in fact, rotation around positive z
rotation_y = -obj.box.heading - np.pi / 2
# track id
track_id = obj.id
# not available
truncated = 0
occluded = 0
# alpha:
# we set alpha to the default -10, the same as nuscenes to kitti tool
# contribution is welcome
alpha = -10
# save the labels
line = my_type + ' {} {} {} {} {} {} {} {} {} {} {} {} {} {}\n'.format(round(truncated, 2),
occluded,
round(alpha, 2),
round(bounding_box[0], 2),
round(bounding_box[1], 2),
round(bounding_box[2], 2),
round(bounding_box[3], 2),
round(height, 2),
round(width, 2),
round(length, 2),
round(x, 2),
round(y, 2),
round(z, 2),
round(rotation_y, 2))
if save_track_id:
line_all = line[:-1] + ' ' + name + ' ' + track_id + '\n'
else:
line_all = line[:-1] + ' ' + name + '\n'
# store the label
fp_label = open(self.label_save_dir + name + '/' + self.prefix + str(file_idx).zfill(3) + str(frame_idx).zfill(3) + '.txt', 'a')
fp_label.write(line)
fp_label.close()
fp_label_all.write(line_all)
fp_label_all.close()
# print(file_idx, frame_idx)
def save_pose(self, frame, file_idx, frame_idx):
""" Save self driving car (SDC)'s own pose
Note that SDC's own pose is not included in the regular training of KITTI dataset
KITTI raw dataset contains ego motion files but are not often used
Pose is important for algorithms that takes advantage of the temporal information
"""
pose = np.array(frame.pose.transform).reshape(4,4)
np.savetxt(join(self.pose_save_dir, self.prefix + str(file_idx).zfill(3) + str(frame_idx).zfill(3) + '.txt'), pose)
def create_folder(self):
for d in [self.label_all_save_dir, self.calib_save_dir, self.point_cloud_save_dir, self.pose_save_dir]:
if not isdir(d):
os.makedirs(d)
for d in [self.label_save_dir, self.image_save_dir]:
for i in range(5):
if not isdir(d + str(i)):
os.makedirs(d + str(i))
def convert_range_image_to_point_cloud(self,
frame,
range_images,
camera_projections,
range_image_top_pose,
ri_index=0):
"""Convert range images to point cloud.
Args:
frame: open dataset frame
range_images: A dict of {laser_name, [range_image_first_return,
range_image_second_return]}.
camera_projections: A dict of {laser_name,
[camera_projection_from_first_return,
camera_projection_from_second_return]}.
range_image_top_pose: range image pixel pose for top lidar.
ri_index: 0 for the first return, 1 for the second return.
Returns:
points: {[N, 3]} list of 3d lidar points of length 5 (number of lidars).
cp_points: {[N, 6]} list of camera projections of length 5
(number of lidars).
"""
calibrations = sorted(frame.context.laser_calibrations, key=lambda c: c.name)
points = []
cp_points = []
intensity = []
frame_pose = tf.convert_to_tensor(
value=np.reshape(np.array(frame.pose.transform), [4, 4]))
# [H, W, 6]
range_image_top_pose_tensor = tf.reshape(
tf.convert_to_tensor(value=range_image_top_pose.data),
range_image_top_pose.shape.dims)
# [H, W, 3, 3]
range_image_top_pose_tensor_rotation = transform_utils.get_rotation_matrix(
range_image_top_pose_tensor[..., 0], range_image_top_pose_tensor[..., 1],
range_image_top_pose_tensor[..., 2])
range_image_top_pose_tensor_translation = range_image_top_pose_tensor[..., 3:]
range_image_top_pose_tensor = transform_utils.get_transform(
range_image_top_pose_tensor_rotation,
range_image_top_pose_tensor_translation)
for c in calibrations:
range_image = range_images[c.name][ri_index]
if len(c.beam_inclinations) == 0: # pylint: disable=g-explicit-length-test
beam_inclinations = range_image_utils.compute_inclination(
tf.constant([c.beam_inclination_min, c.beam_inclination_max]),
height=range_image.shape.dims[0])
else:
beam_inclinations = tf.constant(c.beam_inclinations)
beam_inclinations = tf.reverse(beam_inclinations, axis=[-1])
extrinsic = np.reshape(np.array(c.extrinsic.transform), [4, 4])
range_image_tensor = tf.reshape(
tf.convert_to_tensor(value=range_image.data), range_image.shape.dims)
pixel_pose_local = None
frame_pose_local = None
if c.name == dataset_pb2.LaserName.TOP:
pixel_pose_local = range_image_top_pose_tensor
pixel_pose_local = tf.expand_dims(pixel_pose_local, axis=0)
frame_pose_local = tf.expand_dims(frame_pose, axis=0)
range_image_mask = range_image_tensor[..., 0] > 0
# No Label Zone
if filter_no_label_zone_points:
nlz_mask = range_image_tensor[..., 3] != 1.0 # 1.0: in NLZ
# print(range_image_tensor[range_image_tensor[..., 3] == 1.0])
range_image_mask = range_image_mask & nlz_mask
range_image_cartesian = range_image_utils.extract_point_cloud_from_range_image(
tf.expand_dims(range_image_tensor[..., 0], axis=0),
tf.expand_dims(extrinsic, axis=0),
tf.expand_dims(tf.convert_to_tensor(value=beam_inclinations), axis=0),
pixel_pose=pixel_pose_local,
frame_pose=frame_pose_local)
range_image_cartesian = tf.squeeze(range_image_cartesian, axis=0)
points_tensor = tf.gather_nd(range_image_cartesian,
tf.compat.v1.where(range_image_mask))
cp = camera_projections[c.name][ri_index]
cp_tensor = tf.reshape(tf.convert_to_tensor(value=cp.data), cp.shape.dims)
cp_points_tensor = tf.gather_nd(cp_tensor,
tf.compat.v1.where(range_image_mask))
points.append(points_tensor.numpy())
cp_points.append(cp_points_tensor.numpy())
intensity_tensor = tf.gather_nd(range_image_tensor,
tf.where(range_image_mask))
intensity.append(intensity_tensor.numpy()[:, 1])
return points, cp_points, intensity
# def get_intensity(self, frame, range_images, ri_index=0):
# """Convert range images to point cloud.
# Args:
# frame: open dataset frame
# range_images: A dict of {laser_name,
# [range_image_first_return, range_image_second_return]}.
# camera_projections: A dict of {laser_name,
# [camera_projection_from_first_return,
# camera_projection_from_second_return]}.
# range_image_top_pose: range image pixel pose for top lidar.
# ri_index: 0 for the first return, 1 for the second return.
# Returns:
# intensity: {[N, 1]} list of intensity of length 5 (number of lidars).
# """
# calibrations = sorted(frame.context.laser_calibrations, key=lambda c: c.name)
# intensity = []
# for c in calibrations:
# range_image = range_images[c.name][ri_index]
# range_image_tensor = tf.reshape(
# tf.convert_to_tensor(range_image.data), range_image.shape.dims)
# range_image_mask = range_image_tensor[..., 0] > 0
# intensity_tensor = tf.gather_nd(range_image_tensor,
# tf.where(range_image_mask))
# intensity.append(intensity_tensor.numpy()[:, 1])
#
# return intensity
def cart_to_homo(self, mat):
ret = np.eye(4)
if mat.shape == (3, 3):
ret[:3, :3] = mat
elif mat.shape == (3, 4):
ret[:3, :] = mat
else:
raise ValueError(mat.shape)
return ret
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('load_dir', help='Directory to load Waymo Open Dataset tfrecords')
parser.add_argument('save_dir', help='Directory to save converted KITTI-format data')
parser.add_argument('--prefix', default='', help='Prefix to be added to converted file names')
parser.add_argument('--num_proc', default=1, help='Number of processes to spawn')
args = parser.parse_args()
converter = WaymoToKITTI(args.load_dir, args.save_dir, args.prefix, args.num_proc)
converter.convert()