forked from rpng/open_vins
-
Notifications
You must be signed in to change notification settings - Fork 8
/
batch_run_euroc.py
184 lines (166 loc) · 7.18 KB
/
batch_run_euroc.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
import os
import sys
import copy
import numpy as np
import matplotlib.pyplot as plt
launch_file = '/home/symao/workspace/catkin_ws_ov/src/open_vins/ov_msckf/launch/pgeneva_eth.launch'
bag_dir = '/home/symao/data/euroc/rosbag'
# bag_name:bag_start in sec
bag_list = {'MH_01_easy':0,
'MH_02_easy':0,
'MH_03_medium':0,
'MH_04_difficult':0,
'MH_05_difficult':3,
'V1_01_easy':5,
'V1_02_medium':10,
'V1_03_difficult':6,
'V2_01_easy':4,
'V2_02_medium':4,
'V2_03_difficult':5}
run_cmd = 'roslaunch ov_msckf pgeneva_eth.launch'
run_res = '/home/symao/temp_rmse.txt'
py_path = os.path.dirname(os.path.abspath(sys.argv[0]))
res_table = os.path.join(py_path,'ov_msckf_result.md')
png_dir = os.path.join(py_path,'figure')
marker_table = ['o','*','^','s','p','+','x','d','h','v','<','>','1','2','3']
png_idx = 0
def run_once():
# return tuple((np.random.rand(2)*3).tolist())
if os.path.exists(run_res):
os.remove(run_res)
os.system(run_cmd)
if os.path.exists(run_res):
return [float(x) for x in open(run_res).readlines()[0].split(' ')]
else:
return [-1,-1]
# [name:(type,value)]
def modify_launch(params):
lines = open(launch_file, "r").readlines()
fp = open(launch_file, "w")
for line in lines:
for name in params.keys():
if name in line:
a, b = params[name]
line = ' <param name="%s" type="%s" value="%s" />\n'%(name,a,b)
break
fp.write(line)
fp.close()
def average_rmse(rmse):
deg, meter = np.mean([(a,b) for a,b in rmse if a>0 and b>0], axis=0)
return deg, meter
def plot_rmse(res_list, save_png):
fig = plt.figure(figsize=(16, 8))
names = [s[:5] for s in bag_list.keys()] + ['avg']
plt.subplot(121)
for i, res in enumerate(res_list):
lege, rmse = res
data = np.array(rmse)
plt.plot(data[:,0], marker_table[i]+'-', label=lege)
plt.legend()
plt.xticks(range(len(names)),names,rotation=60)
plt.ylabel('orientation error[degree]')
plt.ylim([0,5])
plt.subplot(122)
for i, res in enumerate(res_list):
lege, rmse = res
data = np.array(rmse)
plt.plot(data[:,1], marker_table[i]+'-', label=lege)
plt.legend()
plt.xticks(range(len(names)),names,rotation=60)
plt.ylabel('position error[m]')
plt.ylim([0,1])
plt.savefig(save_png)
def loop_rosbag(params):
rmse = []
params = copy.deepcopy(params)
for bag,start_ts in bag_list.items():
fbag = os.path.join(bag_dir,bag+'.bag')
fcsv = "$(find ov_data)/euroc_mav/%s.csv"%bag
if os.path.exists(fbag):
params['path_bag'] = ('string', fbag)
params['path_gt'] = ('string', fcsv)
params['bag_start'] = ('double', '%f'%float(start_ts))
modify_launch(params)
res = run_once()
rmse.append(res)
else:
rmse.append((-1,-1))
rmse.append(average_rmse(rmse))
return rmse
def batch_run_single_change(param_table, default_params, fp):
default_res = loop_rosbag(default_params)
for param_name, param_type, choice in param_table:
fp.write('|%s|%s|avg|\n'%(param_name, '|'.join([s[:5] for s in bag_list.keys()])))
fp.write('|'+'--|'*(len(bag_list.keys())+2)+'\n')
fp.write('|%s|%s|\n'%(choice[0],'|'.join(['%.2f,%.2f'%(a,b) for a,b in default_res])))
params = copy.deepcopy(default_params)
res_list = [('%s=%s'%(param_name,choice[0]), default_res)]
for c in choice[1:]:
params[param_name] = (param_type, c)
res = loop_rosbag(params)
res_list.append(('%s=%s'%(param_name,c), res))
fp.write('|%s|%s|\n'%(c,'|'.join(['%.2f,%.2f'%(a,b) for a,b in res])))
fp.write('\n')
global png_idx
png_name = os.path.join(png_dir, '%06d.jpg'%png_idx)
plot_rmse(res_list, png_name)
fp.write('\n'%(png_idx))
fp.flush()
png_idx = png_idx + 1
def run_special_case(params, case_name, fej, intrin, extrin, dt, slam, fp):
params = copy.deepcopy(params)
params['use_fej'] = ('bool', 'true' if fej else 'false')
params['calib_cam_intrinsics'] = ('bool', 'true' if intrin else 'false')
params['calib_cam_extrinsics'] = ('bool', 'true' if extrin else 'false')
params['calib_cam_timeoffset'] = ('bool', 'true' if dt else 'false')
params['max_slam'] = ('int', str(slam))
fp.write('|%s|%d|%d|%d|%d|%d|%s|\n'%(case_name,fej,intrin,extrin,dt,slam,
'|'.join(['%.2f,%.2f'%(a,b) for a,b in loop_rosbag(params)])))
if __name__ == '__main__':
# the first choice always be default
param_table = [('use_fej', 'bool', ['true','false']),
('calib_cam_intrinsics', 'bool', ['true','false']),
('calib_cam_extrinsics', 'bool', ['true','false']),
('calib_cam_timeoffset', 'bool', ['true','false']),
('max_clones', 'int', ['11','5','10','20','30']),
('max_slam', 'int', ['50','0','20','100','200']),
('feat_representation', 'string',['GLOBAL_3D','GLOBAL_FULL_INVERSE_DEPTH',
'ANCHORED_3D','ANCHORED_FULL_INVERSE_DEPTH',
'ANCHORED_MSCKF_INVERSE_DEPTH'])]
default_params = {a:(b,c[0]) for a,b,c in param_table}
# clean roslog
os.system('rm -rf ~/.ros/log')
# record result in markdown
if not os.path.exists(png_dir):
os.makedirs(png_dir)
fp = open(res_table,'w')
fp.write('# OpenVINS evaluation on EuROC dataset\n')
fp.write('NOTE: we log the RMSE of orientation and postion which print by ov_msckf. Unit: [deg, m]\n')
# single variable-controlling
fp.write('## 1. Single param comparision\n')
fp.write('### 1.1 Mono Version\n')
params = copy.deepcopy(default_params)
params['max_cameras'] = ('int', '1')
batch_run_single_change(param_table, params, fp)
fp.write('### 1.2 Stereo Version\n')
params = copy.deepcopy(default_params)
params['max_cameras'] = ('int', '2')
batch_run_single_change(param_table, params, fp)
# user define combination
fp.write('## 2. special cases comparision\n')
fp.write('We test several special cases\n\n')
fp.write('### 2.1 Effect from fej, calib, dt, slam\n')
fp.write('We use default param for: sliding window(11), feature representation(GLOBAL_3D).\n\n')
fp.write('|case|fej|intr|extr|dt|slam|%s|average|\n'%('|'.join([s[:5] for s in bag_list.keys()])))
fp.write('|'+'--|'*(len(bag_list.keys())+7)+'\n')
params = copy.deepcopy(default_params)
params['feat_representation'] = ('string','GLOBAL_3D')
params['max_clones'] = ('int','11')
run_special_case(params,'Naive',0,0,0,0,0,fp)
run_special_case(params,'FEJ',1,0,0,0,0,fp)
run_special_case(params,'Extrin',0,0,1,0,0,fp)
run_special_case(params,'Extrin+Intrin',0,1,1,0,0,fp)
run_special_case(params,'Extrin+Intrin+camdt',0,1,1,1,0,fp)
run_special_case(params,'SLAM',0,0,0,0,50,fp)
run_special_case(params,'All Open',1,1,1,1,50,fp)
fp.close()