-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathZY3LC_dataset.py
277 lines (218 loc) · 8.66 KB
/
ZY3LC_dataset.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
'''
load ZY3LC dataset filename
'''
from glob import glob
import numpy as np
import os
from os.path import join
import pandas as pd
# load classification
def dataloader_tif(filepath, split=0.95):
# INPUT:
# filepath: files for train
# split: train/val dataset
# RETUREN:
# left_train, right_train, disp_train_L, left_val, right_val, disp_val_L
left = glob(join(filepath,'images', 'image*.tif'))
cls = glob(join(filepath, 'labels', 'label*.tif'))
assert len(left)==len(cls)
left.sort()
cls.sort()
left = np.array(left)
cls = np.array(cls)
num = len(left)
num_train = int(split * num)
# File list
seqpath = join(filepath, 'zy3lcseq.txt')
if not os.path.exists(seqpath): # not exist
# Shuffle the file list
indices = np.arange(num)
np.random.seed(0) # fixed
np.random.shuffle(indices)
np.savetxt(seqpath, indices, fmt='%d', delimiter=',')
else:
indices = np.loadtxt(seqpath, delimiter=',')
indices= indices.astype(np.int32)
train = indices[:num_train] # indices for training
val = indices[num_train:] # indices for validation
left_train = left[train]
cls_train = cls[train]
left_val = left[val]
cls_val = cls[val]
return left_train, cls_train, left_val, cls_val
# load classification in png form
def dataloader(filepath, split=(0.7,0.1,0.2), issave=False):
# INPUT:
# filepath: files for train
# split: train/val dataset
# RETUREN:
# left_train, right_train, disp_train_L, left_val, right_val, disp_val_L
left = glob(join(filepath,'img', 'img*.tif'))
cls = glob(join(filepath, 'lab', 'lab*.png'))
assert len(left)==len(cls)
left.sort()
cls.sort()
left = np.array(left)
cls = np.array(cls)
num = len(left)
num_train = int(split[0] * num)
num_val = int(split[1] * num)
# File list
seqpath = join(filepath, 'seq.txt')
if not os.path.exists(seqpath): # not exist
# Shuffle the file list
indices = np.arange(num)
np.random.seed(0) # fixed
np.random.shuffle(indices)
np.savetxt(seqpath, indices, fmt='%d', delimiter=',')
else:
indices = np.loadtxt(seqpath, delimiter=',')
indices= indices.astype(np.int32)
train = indices[:num_train] # indices for training
val = indices[num_train:(num_train+num_val)] # indices for validation
test = indices[(num_train+num_val):]
left_train = left[train]
cls_train = cls[train]
left_val = left[val]
cls_val = cls[val]
left_test = left[test]
cls_test = cls[test]
if issave==True:
ftrain = pd.DataFrame({'0':left_train, '1':cls_train})
ftrain.to_csv(join(filepath,'trainlist.txt'), header=False, index=False)
fval= pd.DataFrame({'0':left_val, '1':cls_val})
fval.to_csv(join(filepath,'vallist.txt'), header=False, index=False)
ftest = pd.DataFrame({'0':left_test, '1':cls_test})
ftest.to_csv(join(filepath,'testlist.txt'), header=False, index=False)
return left_train, cls_train, left_val, cls_val, left_test, cls_test
# load classification in png form
def dataloader_t1t2(filepath, split=(0.7,0.1,0.2), issave=False, fcode=''):
# INPUT:
# filepath: files for train
# split: train/val dataset
# RETUREN:
# left_train, right_train, disp_train_L, left_val, right_val, disp_val_L
left = glob(join(filepath,'img1', 'img_'+ fcode+ '*.tif'))
cls = glob(join(filepath, 'lab', 'lab_'+fcode+ '*.png'))
assert len(left)==len(cls)
left.sort()
cls.sort()
left = np.array(left)
cls = np.array(cls)
num = len(left)
num_train = int(split[0] * num)
num_val = int(split[1] * num)
# File list
seqpath = join(filepath, 'seq'+fcode+'.txt')
if not os.path.exists(seqpath): # not exist
# Shuffle the file list
indices = np.arange(num)
np.random.seed(0) # fixed
np.random.shuffle(indices)
np.savetxt(seqpath, indices, fmt='%d', delimiter=',')
else:
indices = np.loadtxt(seqpath, delimiter=',')
indices= indices.astype(np.int32)
train = indices[:num_train] # indices for training
val = indices[num_train:(num_train+num_val)] # indices for validation
test = indices[(num_train+num_val):]
left_train = left[train]
cls_train = cls[train]
left_val = left[val]
cls_val = cls[val]
left_test = left[test]
cls_test = cls[test]
if issave==True:
ftrain = pd.DataFrame({'0':left_train, '1':cls_train})
ftrain.to_csv(join(filepath,'trainlist.txt'), header=False, index=False)
fval= pd.DataFrame({'0':left_val, '1':cls_val})
fval.to_csv(join(filepath,'vallist.txt'), header=False, index=False)
ftest = pd.DataFrame({'0':left_test, '1':cls_test})
ftest.to_csv(join(filepath,'testlist.txt'), header=False, index=False)
return left_train, cls_train, left_val, cls_val, left_test, cls_test
def dataloader_split(leftp, clsp, seqpath, split=0.9):
# for read imglist
left = pd.read_csv(leftp, header=None)
left = left[0].values.tolist()
# for read lablist
cls = pd.read_csv(clsp, header=None)
cls = cls[0].values.tolist()
left = np.array(left)
cls = np.array(cls)
num = len(left)
num_train = int(split* num)
# File list
if not os.path.exists(seqpath): # not exist
# Shuffle the file list
indices = np.arange(num)
np.random.seed(0) # fixed
np.random.shuffle(indices)
np.savetxt(seqpath, indices, fmt='%d', delimiter=',')
else:
indices = np.loadtxt(seqpath, delimiter=',')
indices= indices.astype(np.int32)
train = indices[:num_train] # indices for training
val = indices[num_train:] # indices for validation
left_train = left[train]
cls_train = cls[train]
left_val = left[val]
cls_val = cls[val]
return left_train, cls_train, left_val, cls_val
# choose specific city to train
def dataloader_city(filepath, cityname, split=(0.7,0.1,0.2), issave=False):
# INPUT:
# filepath: files for train
# split: train/val dataset
# RETUREN:
# left_train, right_train, disp_train_L, left_val, right_val, disp_val_L
left = glob(join(filepath,'img', 'img_'+cityname+'_*.tif'))
cls = glob(join(filepath, 'lab', 'lab_'+cityname+'_*.png'))
assert len(left)==len(cls)
left.sort()
cls.sort()
left = np.array(left)
cls = np.array(cls)
num = len(left)
num_train = int(split[0] * num)
num_val = int(split[1] * num)
# File list
seqpath = join(filepath, 'seq_'+cityname+'.txt')
if not os.path.exists(seqpath): # not exist
# Shuffle the file list
indices = np.arange(num)
np.random.seed(0) # fixed
np.random.shuffle(indices)
np.savetxt(seqpath, indices, fmt='%d', delimiter=',')
else:
indices = np.loadtxt(seqpath, delimiter=',')
indices= indices.astype(np.int32)
train = indices[:num_train] # indices for training
val = indices[num_train:(num_train+num_val)] # indices for validation
test = indices[(num_train+num_val):]
left_train = left[train]
cls_train = cls[train]
left_val = left[val]
cls_val = cls[val]
left_test = left[test]
cls_test = cls[test]
if issave==True:
ftrain = pd.DataFrame({'0':left_train, '1':cls_train})
ftrain.to_csv(join(filepath,'trainlist_'+cityname+'.txt'), header=False, index=False)
fval= pd.DataFrame({'0':left_val, '1':cls_val})
fval.to_csv(join(filepath,'vallist_'+cityname+'.txt'), header=False, index=False)
ftest = pd.DataFrame({'0':left_test, '1':cls_test})
ftest.to_csv(join(filepath,'testlist_'+cityname+'.txt'), header=False, index=False)
return left_train, cls_train, left_val, cls_val, left_test, cls_test
if __name__ == "__main__":
## 20220302 generate valid building patch
# filepath = r'E:\yinxcao\ZY3LC\datanew8bit'
# leftp = join(filepath, 'imglistvalid_train30_imgpath.csv')
# clsp = join(filepath, 'imglistvalid_train30_labpath.csv')
# seqpath = join(filepath, 'seqvalid_train30.txt')
# left_train, cls_train, left_val, cls_val = dataloader_split(leftp, clsp, seqpath, split=0.9)
# print(left_train[:10])
# print(cls_train[:10])
## 20220402 generate test sample patches
filepath = r'E:\yinxcao\ZY3LC\changedata\testdata'
train_img, train_lab, _,_,_,_ = dataloader_t1t2(filepath, split=(1.0, 0, 0), issave=False, fcode='sh')
print(train_img[:5])