forked from Abyssaledge/ImmortalTracker-for-CTRL
-
Notifications
You must be signed in to change notification settings - Fork 0
/
main_waymo.py
177 lines (150 loc) · 7.59 KB
/
main_waymo.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
import os, numpy as np, argparse, json, sys, numba, yaml, shutil
import multiprocessing
# import torch.multiprocessing as multiprocessing
import mot_3d.visualization as visualization, mot_3d.utils as utils
from mot_3d.data_protos import BBox, Validity
from mot_3d.mot import MOTModel
from mot_3d.frame_data import FrameData
from data_loader import WaymoLoader
from ipdb import set_trace
from mot_3d.utils import Timer
import time
timer = Timer(10)
parser = argparse.ArgumentParser()
# running configurations
parser.add_argument('--name', type=str, default='immortal')
parser.add_argument('--det_name', type=str, default='cp')
parser.add_argument('--process', type=int, default=1)
parser.add_argument('--visualize', action='store_true', default=False)
# parser.add_argument('--test', action='store_true', default=False)
parser.add_argument('--skip', action='store_true', default=False)
parser.add_argument('--start_frame', type=int, default=0, help='start at a middle frame for debug')
parser.add_argument('--obj_type', type=str, default='vehicle', choices=['vehicle', 'pedestrian', 'cyclist'])
parser.add_argument('--split', type=str, default='validation')
# paths
parser.add_argument('--config_path', type=str, default='configs/waymo_configs/immortal.yaml')
parser.add_argument('--result_folder', type=str, default='./mot_results/waymo')
parser.add_argument('--data_folder', type=str, default='./data/waymo')
parser.add_argument('--det_data_folder', type=str, default='./data/waymo')
args = parser.parse_args()
if not os.path.exists(args.result_folder):
os.makedirs(args.result_folder)
def load_gt_bboxes(gt_folder, data_folder, segment_name, type_token):
gt_info = np.load(os.path.join(gt_folder, '{:}.npz'.format(segment_name)),
allow_pickle=True)
ego_info = np.load(os.path.join(data_folder, 'ego_info', '{:}.npz'.format(segment_name)),
allow_pickle=True)
bboxes, ids, inst_types = gt_info['bboxes'], gt_info['ids'], gt_info['types']
gt_ids, gt_bboxes = utils.inst_filter(ids, bboxes, inst_types, type_field=[type_token], id_trans=True)
ego_keys = sorted(utils.str2int(ego_info.keys()))
egos = [ego_info[str(key)] for key in ego_keys]
gt_bboxes = gt_bbox2world(gt_bboxes, egos)
return gt_bboxes, gt_ids
def gt_bbox2world(bboxes, egos):
frame_num = len(egos)
for i in range(frame_num):
ego = egos[i]
bbox_num = len(bboxes[i])
for j in range(bbox_num):
bboxes[i][j] = BBox.bbox2world(ego, bboxes[i][j])
return bboxes
def frame_visualization(bboxes, ids, states, gt_bboxes=None, gt_ids=None, pc=None, dets=None, name=''):
visualizer = visualization.Visualizer2D(name=name, figsize=(12, 12))
if pc is not None:
visualizer.handler_pc(pc)
for _, bbox in enumerate(gt_bboxes):
visualizer.handler_box(bbox, message='', color='black')
dets = [d for d in dets if d.s >= 0.1]
for det in dets:
visualizer.handler_box(det, message='%.2f' % det.s, color='green', linestyle='dashed')
for _, (bbox, id, state) in enumerate(zip(bboxes, ids, states)):
if Validity.agein1(state):
visualizer.handler_box(bbox, message=str(id), color='red')
else:
visualizer.handler_box(bbox, message=str(id), color='light_blue')
visualizer.show()
visualizer.save('temp.jpg')
visualizer.close()
import pdb
pdb.set_trace()
def sequence_mot(configs, data_loader: WaymoLoader, sequence_id, gt_bboxes=None, gt_ids=None, visualize=False, return_tracklets=False):
tracker = MOTModel(configs)
frame_num = len(data_loader)
IDs, bboxes, states, types = list(), list(), list(), list()
for _ in range(data_loader.cur_frame, frame_num):
frame_data = next(data_loader)
frame_data = FrameData(dets=frame_data['dets'], ego=frame_data['ego'], pc=frame_data['pc'],
det_types=frame_data['det_types'], aux_info=frame_data['aux_info'], time_stamp=frame_data['time_stamp'], abs_frame_index=frame_data['abs_frame_index'])
# mot
frame_results = tracker.frame_mot(frame_data)
result_pred_bboxes = [trk['bboxes'] for trk in frame_results]
result_pred_ids = [trk['id'] for trk in frame_results]
result_pred_states = [trk['state'] for trk in frame_results]
result_types = [trk['type'] for trk in frame_results]
# wrap for output
IDs.append(result_pred_ids)
result_pred_bboxes = [BBox.bbox2array(bbox) for bbox in result_pred_bboxes]
bboxes.append(result_pred_bboxes)
states.append(result_pred_states)
types.append(result_types)
if return_tracklets:
return tracker.tracklets
return IDs, bboxes, states, types
def main(name, obj_type, config_path, data_folder, det_data_folder, result_folder, counter_list, start_frame=0, token=0, process=1):
summary_folder = os.path.join(result_folder, 'summary', obj_type)
# simply knowing about all the segments
file_names = sorted(os.listdir(os.path.join(data_folder, 'ego_info')))
if args.skip:
file_names = [fname for fname in file_names if not os.path.exists(os.path.join(summary_folder, fname))]
# load model configs
configs = yaml.load(open(config_path, 'r'))
gpu = configs['running'].get('gpu', False)
if gpu:
import torch
torch.cuda.set_device(token % 8)
if obj_type == 'vehicle':
type_token = 1
elif obj_type == 'pedestrian':
type_token = 2
elif obj_type == 'cyclist':
type_token = 4
for file_index, file_name in enumerate(file_names[:]):
if file_index % process != token:
continue
segment_name = file_name.split('.')[0]
data_loader = WaymoLoader(configs, [type_token], segment_name, data_folder, det_data_folder, start_frame)
ids, bboxes, states, types = sequence_mot(configs, data_loader, file_index)
counter_list.append(file_index)
print('FINISH TYPE {:} SEQ {:} / {:}'.format(obj_type, len(counter_list), len(file_names)))
np.savez_compressed(os.path.join(summary_folder, '{}.npz'.format(segment_name)),
ids=ids, bboxes=bboxes, states=states)
if __name__ == '__main__':
assert args.split in ('training', 'validation', 'testing')
args.data_folder = os.path.join(args.data_folder, args.split)
args.det_data_folder = os.path.join(args.det_data_folder, args.split, 'detection')
args.result_folder = os.path.join(args.result_folder, args.split)
result_folder = os.path.join(args.result_folder, args.name + f'_{args.det_name}')
if not os.path.exists(result_folder):
os.makedirs(result_folder)
summary_folder = os.path.join(result_folder, 'summary')
if not os.path.exists(summary_folder):
os.makedirs(summary_folder)
summary_folder = os.path.join(summary_folder, args.obj_type)
if not os.path.exists(summary_folder):
os.makedirs(summary_folder)
det_data_folder = os.path.join(args.det_data_folder, args.det_name)
manager = multiprocessing.Manager()
counter_list = manager.list()
beg = time.time()
if args.process > 1:
pool = multiprocessing.Pool(args.process)
for token in range(args.process):
result = pool.apply_async(main, args=(args.name, args.obj_type, args.config_path, args.data_folder, det_data_folder,
result_folder, counter_list, 0, token, args.process))
pool.close()
pool.join()
else:
main(args.name, args.obj_type, args.config_path, args.data_folder, det_data_folder, result_folder, counter_list,
args.start_frame, 0, 1)
end = time.time()
print(f'Tracking time cost: {end - beg}s')