forked from sunwillz/tianchi_land_monitor
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathutils.py
190 lines (168 loc) · 7.92 KB
/
utils.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
# -*- coding: utf-8 -*-
"""
@author：sunwill
滑动窗口、数据增广
"""
import os
import cv2
import re
import tifffile as tiff
import numpy as np
import matplotlib.image as pltimage
image_size = 256
def scale_percentile(matrix):
w, h, d = matrix.shape
matrix = np.reshape(matrix, [w * h, d]).astype(np.float64)
mins = np.percentile(matrix, 1, axis=0)
maxs = np.percentile(matrix, 99, axis=0) - mins
matrix = (matrix - mins[None, :]) / maxs[None, :]
matrix = np.reshape(matrix, [w, h, d])
matrix = matrix.clip(0, 1)
return matrix
## 数据采样：选择标记占20%-70%的样本作为训练集
def data_sample(images, labels, low_rate=0.2, high_rate=0.7):
new_images = []
new_labels = []
assert np.max(labels) == 1.0
for img, label in zip(images, labels):
if low_rate <= np.mean(label[:, :, 1]) <= high_rate:
new_images.append(img)
new_labels.append(label)
new_images = np.array(new_images)
new_labels = np.array(new_labels)
print 'after sample:', new_images.shape
print 'after sample:', new_labels.shape
return new_images, new_labels
# def data_augmentation():
# training_data = Dataset_reader(dataset_dir=training_dir,
# file_name=train_file,
# image_size=image_size,
# image_channel=image_channel,
# label_channel=label_channel
# )
# train_images, train_annotations = training_data.get_all_data()
#
# data_gen_args = dict(rotation_range=0.2,
# width_shift_range=0.2,
# height_shift_range=0.2,
# shear_range=0.2,
# zoom_range=0.2,
# horizontal_flip=True,
# fill_mode='nearest')
#
# image_datagen = ImageDataGenerator(**data_gen_args)
# mask_datagen = ImageDataGenerator(**data_gen_args)
#
# seed = 10
#
# i = 0
# for _ in image_datagen.flow(
# train_images, seed=seed,
# batch_size=batch_size, save_to_dir='./data_224/images/data_aug/',
# save_prefix='data_aug', save_format='jpg'):
# i += 1
# if i == 100:
# break
# i = 0
# for _ in mask_datagen.flow(
# train_annotations, seed=seed,
# batch_size=batch_size, save_to_dir='./data_224/labels/data_aug/',
# save_prefix='data_aug', save_format='jpg'):
# i += 1
# if i == 100:
# break
## 滑动窗口分割图像
def gen_data_by_slide_window():
base_dir = './data_{}'.format(image_size)
img_2015 = tiff.imread('./original_data/quickbird2015.tif').transpose([1, 2, 0]) ## shape=(5106,15106,4)
img_2017 = tiff.imread('./original_data/quickbird2017.tif').transpose([1, 2, 0])
img_2015 = img_2015[:, :14400, :]
img_2017 = img_2017[:, :14400, :]
label_2015 = pltimage.imread(base_dir+'/2015.jpg') ## shape=(5106,14400,3)
label_2017 = pltimage.imread(base_dir+'/2017.jpg')
if image_size == 160:
x_step = 120
y_step = 120
# 14400-160 = 14240 = 118*120+80
# 5106-160 = 4946 = 41*120+26
for i in range(42):
for j in range(119):
im_name = str(i) + '_' + str(j) + '_' + str(image_size) + '_.jpg'
cv2.imwrite(base_dir+'/slide_window/images/2015/' + im_name, scale_percentile(
img_2015[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/images/2017/' + im_name, scale_percentile(
img_2017[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/labels/2015/' + im_name, scale_percentile(
label_2015[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/labels/2017/' + im_name, scale_percentile(
label_2017[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
elif image_size == 224:
x_step = 194
y_step = 194
# 14400-224 = 14176 = 73*194+14
# 5106-224 = 4882 = 25*194+32
for i in range(26):
for j in range(74):
im_name = str(i) + '_' + str(j) + '_' + str(image_size) + '_.jpg'
cv2.imwrite(base_dir+'/slide_window/images/2015/' + im_name, scale_percentile(
img_2015[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/images/2017/' + im_name, scale_percentile(
img_2017[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/labels/2015/' + im_name, scale_percentile(
label_2015[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/labels/2017/' + im_name, scale_percentile(
label_2017[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
elif image_size == 256:
x_step = 224
y_step = 224
# 14400-256 = 14144 = 63*224+32
# 5106-256 = 4856 = 21*224*152
for i in range(22):
for j in range(64):
im_name = str(i) + '_' + str(j) + '_' + str(image_size) + '_.jpg'
cv2.imwrite(base_dir+'/slide_window/images/2015/' + im_name, scale_percentile(
img_2015[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/images/2017/' + im_name, scale_percentile(
img_2017[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/labels/2015/' + im_name, scale_percentile(
label_2015[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
cv2.imwrite(base_dir+'/slide_window/labels/2017/' + im_name, scale_percentile(
label_2017[i * y_step:i * y_step + image_size, j * x_step:j * x_step + image_size, :3]) * 255)
if __name__ == "__main__":
# data_augmentation()
# x = np.array(os.listdir('./data_224/images/data_aug/'))
# y = np.array(os.listdir('./data_224/labels/data_aug/'))
# print x.shape
# print y.shape
# assert x.shape == y.shape
# with open('./data_224/train.txt', 'a') as f:
# for item in x:
# f.write('data_aug/'+item+'\n')
gen_data_by_slide_window()
to_dir = './data_{}/slide_window/'.format(image_size)
images_list_2015 = np.array(os.listdir(to_dir+'images/2015/'))
label_list_2015 = np.array(os.listdir(to_dir+'labels/2015/'))
images_list_2017 = np.array(os.listdir(to_dir+'images/2017/'))
label_list_2017 = np.array(os.listdir(to_dir+'labels/2017/'))
common_2015 = np.intersect1d(images_list_2015, label_list_2015)
common_2017 = np.intersect1d(images_list_2017, label_list_2017)
_2015 = ['2015/' + i for i in common_2015]
_2017 = ['2017/' + i for i in common_2017]
common = np.hstack([_2015, _2017])
print common.shape[0]
perm = np.arange(common.shape[0])
np.random.shuffle(perm)
common = common[perm]
train = common
print train.shape
valid = common[int(common.shape[0]*0.8):]
print valid.shape
reg = r'([0-9]{4})\/[0-9]{0,3}_[0-9]{0,3}_[0-9]{3}_.jpg'
with open(to_dir+'train.txt', 'w') as f:
for line in train:
if re.match(reg, line):
f.write(line+'\n')
with open(to_dir+'valid.txt', 'w') as f:
for line in valid:
if re.match(reg, line):
f.write(line+'\n')