forked from microsoft/lamar-benchmark
-
Notifications
You must be signed in to change notification settings - Fork 0
/
pipeline_sequence.py
193 lines (163 loc) · 7.81 KB
/
pipeline_sequence.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
from typing import List, Optional
from pathlib import Path
import argparse
import shutil
from scantools import logger
from scantools.capture import Capture, Session
from scantools import (
run_phone_to_capture,
run_sequence_aligner,
run_joint_refinement,
run_combine_sequences,
run_map_query_split,
run_radio_transfer,
)
from scantools.run_joint_refinement import MatchingConf, RefinementConf
conf_matcher = {'output': 'matches-superglue',
'model': {'name': 'superglue', 'weights': 'outdoor', 'sinkhorn_iterations': 5}}
conf_matching = MatchingConf('netvlad', 'superpoint_aachen', conf_matcher)
conf_align = {
'ios': run_sequence_aligner.Conf.from_dict(dict(
**run_sequence_aligner.conf_ios, matching=conf_matching.to_dict())),
'hl': run_sequence_aligner.Conf.from_dict(dict(
**run_sequence_aligner.conf_hololens, matching=conf_matching.to_dict())),
}
conf_align['ios'].matching.local_features['model']['max_keypoints'] = 2048
conf_align['hl'].matching.local_features['model']['max_keypoints'] = 1024
conf_refine = RefinementConf(
conf_matching,
keyframings={
Session.Device.PHONE: conf_align['ios'].localizer.keyframing,
Session.Device.HOLOLENS: conf_align['hl'].localizer.keyframing,
},
)
eval_keyframing = run_combine_sequences.KeyFramingConf()
map_keyframing = run_combine_sequences.KeyFramingConf(max_distance=0.5, max_elapsed=0.4)
def read_sequence_list(path) -> List[str]:
sequences = []
with open(path, 'r') as fid:
for line in fid.read().split('\n'):
line = line.strip()
if not line or line.startswith('#'):
continue
sequences.append(line.split('#')[0].strip())
return sequences
sequence_list_dir = Path(__file__).parent / 'sequences'
def process_sequence(capture, ref_id, input_path, conf, kind):
sequence_id = f'{kind}_{input_path.name}'
logger.info('Working on %s.', sequence_id)
chunk_ids = sorted(filter(lambda i: i.startswith(sequence_id), capture.sessions))
if len(chunk_ids) == 0:
if kind == 'ios':
chunk_ids = run_phone_to_capture.run(input_path, capture, sequence_id)
# run_image_anonymization.run(capture, sequence_id, use_gpu=True, visualize=True)
elif kind.startswith('hl'):
shutil.copytree(input_path, capture.session_path(sequence_id))
capture.sessions[sequence_id] = Session.load(capture.sessions_path() / sequence_id)
chunk_ids = [sequence_id]
else:
raise ValueError(kind)
logger.info('Found %d chunks for sequence %s', len(chunk_ids), sequence_id)
chunk_ids_aligned = []
num_failed = 0
for session_id in chunk_ids:
path_trajectory = capture.registration_path() / session_id / ref_id / 'trajectory_ba.txt'
if not path_trajectory.exists():
logger.info('Aligning session %s.', session_id)
success = run_sequence_aligner.run(
capture, ref_id, session_id, conf,
overwrite=False,
visualize_diff=False,
vis_mesh_id='mesh_simplified')
if not success:
num_failed += 1
continue
chunk_ids_aligned.append(session_id)
if num_failed > 0:
logger.warning('Could not align %d/%d chunks for session %s.',
num_failed, len(chunk_ids), sequence_id)
return chunk_ids_aligned
def run(capture_path: Path,
ref_id: str,
phone_dir: Optional[Path] = None,
hololens_dir: Optional[Path] = None,
phone_sequences: List[str] = ('*',),
hololens_sequences: List[str] = ('*',),
):
capture = Capture.load(capture_path, wireless=False)
select_path = capture.path / 'sequences_select.txt'
if select_path.exists():
sequence_ids = read_sequence_list(select_path)
for i in sequence_ids:
if i not in capture.sessions:
raise ValueError(i, list(capture.sessions.keys()))
logger.info('Read %d sequences from %s', len(sequence_ids), select_path)
else:
sequence_ids = []
if phone_dir is not None:
phone_paths = [p for g in phone_sequences for p in phone_dir.glob(g)]
for path in phone_paths:
sequence_ids += process_sequence(capture, ref_id, path, conf_align['ios'], 'ios')
if hololens_dir is not None:
hololens_paths = [hololens_dir / g for g in hololens_sequences]
for path in hololens_paths:
sequence_ids += process_sequence(capture, ref_id, path, conf_align['hl'], 'hl')
assert len(sequence_ids) > 0
logger.info('Found %d sequences', len(sequence_ids))
with open(capture.path / 'sequences.txt', 'w') as fid:
fid.write("\n".join(sequence_ids))
if not all((capture.registration_path()/i/'trajectory_refined.txt').exists()
for i in sequence_ids):
logger.info('Running the joint refinement.')
run_joint_refinement.run(capture, ref_id, sequence_ids, conf_refine)
logger.info('Splitting sequences into maps and queries.')
map_ids, query_ids = run_map_query_split.run(capture, sequence_ids, ref_id=ref_id)
query_ids_phone = list(filter(lambda i: i.startswith('ios'), query_ids))
query_ids_hololens = list(filter(lambda i: i.startswith('hl'), query_ids))
map_id = 'map'
query_id_phone = 'query_phone'
query_id_hololens = 'query_hololens'
logger.info('Writing map and query sessions')
run_combine_sequences.run(
capture, map_ids, map_id, overwrite_poses=True,
keyframing=map_keyframing, reference_id=ref_id)
for i, ids in [[query_id_phone, query_ids_phone], [query_id_hololens, query_ids_hololens]]:
run_combine_sequences.run(
capture, ids, i, overwrite_poses=False, keyframing=eval_keyframing)
run_radio_transfer.run(capture, [map_id, query_id_phone, query_id_hololens])
def get_data_CAB():
ref_id = '2022-06-21_09.28.22+2022-06-25_11.14.36'
phone_sequences = read_sequence_list(sequence_list_dir / 'CAB_phone.txt')
hololens_sequences = read_sequence_list(sequence_list_dir / 'CAB_hololens.txt')
return ref_id, phone_sequences, hololens_sequences
def get_data_HGE():
ref_id = '2022-02-06_12.55.11+2022-02-26_16.21.10'
phone_sequences = read_sequence_list(sequence_list_dir / 'HGE_phone.txt')
hololens_sequences = read_sequence_list(sequence_list_dir / 'HGE_hololens.txt')
return ref_id, phone_sequences, hololens_sequences
def get_data_LIN():
ref_id = '2022-07-03_08.30.21'
phone_sequences = read_sequence_list(sequence_list_dir / 'LIN_phone.txt')
hololens_sequences = read_sequence_list(sequence_list_dir / 'LIN_hololens.txt')
return ref_id, phone_sequences, hololens_sequences
def main(args):
scene = args.scene
ref_id, phone_sequences, hololens_sequences = eval('get_data_'+scene)()
if args.skip_phone:
phone_sequences = []
if args.skip_hololens:
hololens_sequences = []
logger.info('Found %d phone and %d HoloLens sequences in lists.',
len(phone_sequences), len(hololens_sequences))
run(args.capture_root/scene, ref_id, args.phone_dir/scene, args.hololens_dir/scene,
phone_sequences, hololens_sequences)
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--capture_root', type=Path, default=Path('data/captures/'))
parser.add_argument('--scene', type=str, required=True, choices=['CAB', 'HGE', 'LIN'])
parser.add_argument('--phone_dir', type=Path, default=Path('/media/HD8TB/ios_rec/'))
parser.add_argument('--hololens_dir', type=Path,
default=Path('/media/SSD2/hololens_proc/all-fix'))
parser.add_argument('--skip_phone', action='store_true')
parser.add_argument('--skip_hololens', action='store_true')
main(parser.parse_args())