-
Notifications
You must be signed in to change notification settings - Fork 1
/
Albumentations_data.py
454 lines (384 loc) · 18.5 KB
/
Albumentations_data.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
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
import albumentations as A
import cv2
from matplotlib import pyplot as plt
import os
import numpy as np
import codecs
import json
from glob import glob
import shutil
from sklearn.model_selection import train_test_split
KEYPOINT_COLOR = (0, 255, 0) # Green
def vis_keypoints(image, keypoints, color=KEYPOINT_COLOR, diameter=15):
image = image.copy()
for (x, y) in keypoints:
cv2.circle(image, (int(x), int(y)), diameter, (0, 255, 0), -1)
plt.figure(figsize=(8, 8))
plt.axis('off')
plt.imshow(image)
def data_aug():
# Declare an augmentation pipeline
transform = A.Compose([
# A.RandomCrop(width=450, height=450),
A.HorizontalFlip(p=0.5),
# A.VerticalFlip(p=0.5),
# A.RandomBrightnessContrast(p=0.2),
], bbox_params=A.BboxParams(format='pascal_voc', label_fields=['class_labels']), keypoint_params=A.KeypointParams(format='xy'))
# 3-
#image = cv2.imread("images/labelme/1639726903.jpg")
image = cv2.imread("images/labelme/1639726322.jpg")
#cv2.imwrite("./sss1.jpg", image)
#image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# bboxes = [
# [0, 117, 638, 247],
# ]
# keypoints = [
# (638,119),
# (0, 253),
# (1, 157),
# (637, 124),
# (636, 238),
# (2, 250),
# ]
bboxes = [
[0, 208, 639, 314],
]
keypoints = [
(0, 208),
(639, 314),
(0, 214),
(637, 207),
(639, 307),
(0, 314),
]
image1 = image.copy()
cv2.rectangle(image, (0, 208),(639, 314), (0, 255, 0))
#cv2.circle(image, (0, 208), 5, (0, 255, 0), -1)
#cv2.circle(image, (639, 314), 5, (0, 255, 0), -1)
cv2.circle(image, (0, 214), 5, (0, 255, 255), -1)
cv2.circle(image, (637, 207), 5, (0, 255, 255), -1)
cv2.circle(image, (639, 307), 5, (0, 255, 255), -1)
cv2.circle(image, (0, 314), 5, (0, 255, 255), -1)
cv2.imwrite("./sss1.jpg", image)
class_labels = ['false']
#class_categories = ['animal', 'animal', 'item']
# 4-
transformed = transform(image=image1, bboxes=bboxes, class_labels=class_labels, keypoints=keypoints)
transformed_image = transformed['image']
transformed_bboxes = transformed['bboxes']
transformed_keypoints = transformed['keypoints']
transformed_class_labels = transformed['class_labels']
diameter = 3
cv2.rectangle(transformed_image, (int(transformed_keypoints[0][0]), int(transformed_keypoints[0][1])), (int(transformed_keypoints[1][0]), int(transformed_keypoints[1][1])), (0, 255, 0))
for (x, y) in transformed_keypoints:
print(int(x))
print(int(y))
cv2.circle(transformed_image, (int(x), int(y)), diameter, (0, 255, 0), -1)
#vis_keypoints(transformed_image, transformed_keypoints)
cv2.imwrite("./sss.jpg", transformed_image)
def data_aug_data(p_0, p_1, pic, rename):
# Declare an augmentation pipeline
if rename == '_horizon':
transform = A.Compose([
# A.RandomCrop(width=450, height=450),
A.HorizontalFlip(p=1.0),
# A.VerticalFlip(p=0.5),
#A.RandomBrightnessContrast(p=0.2),
], bbox_params=A.BboxParams(format='pascal_voc', label_fields=['class_labels']),
keypoint_params=A.KeypointParams(format='xy'))
if rename == '_Blur':
transform = A.Compose([
# A.RandomCrop(width=450, height=450),
# A.HorizontalFlip(p=0.5),
# A.VerticalFlip(p=0.5),
# A.RandomBrightnessContrast(p=0.2),
A.OneOf([
A.MotionBlur(p=0.5), # 使用随机大小的内核将运动模糊应用于输入图像。
A.MedianBlur(blur_limit=3, p=0.5), # 中值滤波
A.Blur(blur_limit=3, p=0.5), # 使用随机大小的内核模糊输入图像。
], p=1.0),
], bbox_params=A.BboxParams(format='pascal_voc', label_fields=['class_labels']), keypoint_params=A.KeypointParams(format='xy'))
# 3-
image = cv2.imread(pic)
#cv2.imwrite("./sss1.jpg", image)
#image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
bboxes = [
[min(p_0[:, 0]), min(p_0[:, 1]), max(p_0[:, 0]), max(p_0[:, 1])],
]
keypoints = [
(min(p_0[:, 0]), min(p_0[:, 1])),
(max(p_0[:, 0]), max(p_0[:, 1])),
(p_1[0, 0],p_1[0, 1]),
(p_1[1, 0], p_1[1, 1]),
(p_1[2, 0], p_1[2, 1]),
(p_1[3, 0], p_1[3, 1]),
]
class_labels = ['false']
#class_categories = ['animal', 'animal', 'item']
# 4-
transformed = transform(image=image, bboxes=bboxes, class_labels=class_labels, keypoints=keypoints)
transformed_image = transformed['image']
transformed_bboxes = transformed['bboxes']
transformed_keypoints = transformed['keypoints']
transformed_class_labels = transformed['class_labels']
return transformed_keypoints, transformed_image
def data_aug_pic( pic, rename):
# Declare an augmentation pipeline
if rename == '_horizon':
transform = A.Compose([
# A.RandomCrop(width=450, height=450),
A.HorizontalFlip(p=1.0),
# A.VerticalFlip(p=0.5),
#A.RandomBrightnessContrast(p=0.2),
])
if rename == '_Blur':
transform = A.Compose([
# A.RandomCrop(width=450, height=450),
# A.HorizontalFlip(p=0.5),
# A.VerticalFlip(p=0.5),
# A.RandomBrightnessContrast(p=0.2),
A.OneOf([
A.MotionBlur(p=0.5), # 使用随机大小的内核将运动模糊应用于输入图像。
A.MedianBlur(blur_limit=3, p=0.5), # 中值滤波
A.Blur(blur_limit=3, p=0.5), # 使用随机大小的内核模糊输入图像。
], p=1.0),
])
# 3-
image = cv2.imread(pic)
#cv2.imwrite("./sss1.jpg", image)
#image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
#class_categories = ['animal', 'animal', 'item']
# 4-
transformed = transform(image=image)
transformed_image = transformed['image']
return transformed_image
def labelme_json_voc_widerface(rename):
# 1.标签路径
labelme_path = ".\\labelme_to_voc_widerface\\labelme\\" # 原始labelme标注数据路径
saved_path = ".\\labelme_to_voc_widerface\\VOC2007\\" # 保存路径
#labelme_path = ".\\images\\labelme\\" # 原始labelme标注数据路径
#saved_path = ".\\images\\VOC2007\\" # 保存路径
# 2.创建要求文件夹
if not os.path.exists(saved_path + "Annotations"):
os.makedirs(saved_path + "Annotations")
if not os.path.exists(saved_path + "JPEGImages/"):
os.makedirs(saved_path + "JPEGImages/")
if not os.path.exists(saved_path + "ImageSets/Main/"):
os.makedirs(saved_path + "ImageSets/Main/")
# 3.获取待处理文件
files = glob(labelme_path + "*.json")
files = [i.split("\\")[-1].split(".json")[0] for i in files]
#file_handle = open(".\\images\\save_widerface_result_horizon1.txt", mode='w')
if rename == '_horizon':
file_handle = open("labelme_to_voc_widerface/save_widerface_result_horizon.txt", mode='w')
if rename == '_Blur':
file_handle = open("labelme_to_voc_widerface/save_widerface_result_blur.txt", mode='w')
# 4.读取标注信息并写入 xml
for json_file_ in files:
json_filename = labelme_path + json_file_ + ".json"
json_file = json.load(open(json_filename, "r", encoding="utf-8"))
height, width, channels = cv2.imread(labelme_path + json_file_ + ".jpg").shape
with codecs.open(saved_path + "Annotations/" + json_file_ + rename+".xml", "w", "utf-8") as xml:
xml.write('<annotation>\n')
xml.write('\t<folder>' + 'STOP_LINE' + '</folder>\n')
xml.write('\t<filename>' + json_file_ +rename+ ".jpg" + '</filename>\n')
xml.write('\t<source>\n')
xml.write('\t\t<database>The STOP LINE Database</database>\n')
xml.write('\t\t<annotation>PASCAL VOC</annotation>\n')
xml.write('\t\t<image>flickr</image>\n')
xml.write('\t\t<flickrid>NULL</flickrid>\n')
xml.write('\t</source>\n')
xml.write('\t<owner>\n')
xml.write('\t\t<flickrid>NULL</flickrid>\n')
xml.write('\t\t<name>Line</name>\n')
xml.write('\t</owner>\n')
xml.write('\t<size>\n')
xml.write('\t\t<width>' + str(width) + '</width>\n')
xml.write('\t\t<height>' + str(height) + '</height>\n')
xml.write('\t\t<depth>' + str(channels) + '</depth>\n')
xml.write('\t</size>\n')
xml.write('\t<segmented>0</segmented>\n')
# print(json_file["shapes"][0]["points"])
points_0 = np.array(json_file["shapes"][0]["points"])
points_1 = np.array(json_file["shapes"][1]["points"])
transform_data, new_img = data_aug_data(points_0, points_1, labelme_path + json_file_ + ".jpg",rename)
cv2.imwrite(saved_path + "JPEGImages\\"+json_file_+ rename+".jpg", new_img)
xmin = min(transform_data[0][0],transform_data[1][0])
xmax = max(transform_data[0][0],transform_data[1][0])
ymin = min(transform_data[0][1],transform_data[1][1])
ymax = max(transform_data[0][1],transform_data[1][1])
label = json_file["shapes"][0]["label"]
if xmax <= xmin:
pass
elif ymax <= ymin:
pass
else:
xml.write('\t<object>\n')
xml.write('\t\t<name>' + str(label) + '</name>\n')
xml.write('\t\t<pose>Unspecified</pose>\n')
xml.write('\t\t<truncated>1</truncated>\n')
xml.write('\t\t<difficult>0</difficult>\n')
xml.write('\t\t<bndbox>\n')
xml.write('\t\t\t<xmin>' + str(int(xmin)) + '</xmin>\n')
xml.write('\t\t\t<ymin>' + str(int(ymin)) + '</ymin>\n')
xml.write('\t\t\t<xmax>' + str(int(xmax)) + '</xmax>\n')
xml.write('\t\t\t<ymax>' + str(int(ymax)) + '</ymax>\n')
xml.write('\t\t</bndbox>\n')
xml.write('\t</object>\n')
label_1 = json_file["shapes"][1]["label"]
x1 = transform_data[2][0]
y1 = transform_data[2][1]
x2 = transform_data[3][0]
y2 = transform_data[3][1]
x3 = transform_data[4][0]
y3 = transform_data[4][1]
x4 = transform_data[5][0]
y4 = transform_data[5][1]
'''
xml.write('\t<object>\n')
xml.write('\t\t<name>'+str(label_1)+'</name>\n')
xml.write('\t\t<pose>Unspecified</pose>\n')
xml.write('\t\t<truncated>1</truncated>\n')
xml.write('\t\t<difficult>0</difficult>\n')
xml.write('\t\t<point>\n')
xml.write('\t\t\t<x1>' + str(int(x1)) + '</x1>\n')
xml.write('\t\t\t<y1>' + str(int(y1)) + '</y1>\n')
xml.write('\t\t\t<x2>' + str(int(x2)) + '</x2>\n')
xml.write('\t\t\t<y2>' + str(int(y2)) + '</y2>\n')
xml.write('\t\t\t<x3>' + str(int(x3)) + '</x3>\n')
xml.write('\t\t\t<y3>' + str(int(y3)) + '</y3>\n')
xml.write('\t\t\t<x4>' + str(int(x4)) + '</x4>\n')
xml.write('\t\t\t<y4>' + str(int(y4)) + '</y4>\n')
xml.write('\t\t</point>\n')
xml.write('\t</object>\n')
'''
# print(json_file_ + ".jpg",xmin,ymin,xmax,ymax,label,x1,y1,x2,y2,x3,y3,x4,y4)
if rename == '_Blur':
txt_file = "# " + json_file_ + rename+".jpg" + '\n' + str(int(xmin)) + " " + str(int(ymin)) + " " + str(
int(xmax-xmin)) + " " + str(int(ymax-ymin)) + " " + str(int(label)) + " " + str(int(x1)) + " " + str(
int(y1)) + " " + str(int(x2)) + " " + str(int(y2)) + " " + str(int(x3)) + " " + str(
int(y3)) + " " + str(int(x4)) + " " + str(int(y4)) + " " + str(int(label_1))
if rename == '_horizon':
txt_file = "# " + json_file_ + rename + ".jpg" + '\n' + str(int(xmin)) + " " + str(
int(ymin)) + " " + str(int(xmax - xmin)) + " " + str(int(ymax - ymin)) + " " + str(int(label)) + " " + str(
int(x2)) + " " + str(int(y2)) + " " + str(int(x1)) + " " + str(int(y1)) + " " + str(int(x4)) + " " + str(
int(y4)) + " " + str(int(x3)) + " " + str(int(y3)) + " " + str(int(label_1))
file_handle.write(txt_file + '\n')
xml.write('</annotation>')
# 5.复制图片到 VOC2007/JPEGImages/下
#image_files = glob(labelme_path + "*.jpg")
#print("copy image files to VOC007/JPEGImages/")
# for image in image_files:
# new_name = os.path.basename(image).split(".")[0]+"_abl.jpg"
# shutil.copy(image, saved_path + "JPEGImages\\"+new_name)
# 6.split files for txt
txtsavepath = saved_path + "ImageSets\\Main\\"
ftrainval = open(txtsavepath + '\\trainval.txt', 'w')
ftest = open(txtsavepath + '\\test.txt', 'w')
ftrain = open(txtsavepath + '\\train.txt', 'w')
fval = open(txtsavepath + '\\val.txt', 'w')
#total_files = glob(".\\images\\VOC2007\\Annotations\\*.xml")
total_files = glob(".\\labelme_to_voc_widerface\\VOC2007\\Annotations\\*.xml")
total_files = [i.split("\\")[-1].split(".xml")[0] for i in total_files]
# test_filepath = ""
for file in total_files:
#os.path.basename(file).split(".")[0] + "_abl.jpg"
ftrainval.write(file + "\n")
# test
# for file in os.listdir(test_filepath):
# ftest.write(file.split(".jpg")[0] + "\n")
# split
train_files, val_files = train_test_split(total_files, test_size=0.15, random_state=42)
# train
for file in train_files:
ftrain.write(file +"\n")
# val
for file in val_files:
fval.write(file + "\n")
ftrainval.close()
ftrain.close()
fval.close()
def labelme_json_voc_widerface_copy():
# 1.标签路径
labelme_path = "F:\\stop_line_data\\train\\" # 原始labelme标注数据路径
saved_path = "F:\\stop_line_data\\label_1\\" # 保存路径
# 3.获取待处理文件
files = glob(labelme_path + "*.json")
files = [i.split("\\")[-1].split(".json")[0] for i in files]
file_handle = open("labelme_to_voc_widerface/save_widerface_result_horizon1.txt", mode='w')
# file_handle = open("labelme_to_voc_widerface/save_widerface_result_blur1.txt", mode='w')
# 4.读取标注信息并写入 xml
for json_file_ in files:
json_filename = labelme_path + json_file_ + ".json"
json_file = json.load(open(json_filename, "r", encoding="utf-8"))
height, width, channels = cv2.imread(labelme_path + json_file_ + ".jpg").shape
label = json_file["shapes"][0]["label"]
if int(label) > 0:
shutil.copy(labelme_path + json_file_ + ".jpg", saved_path)
shutil.copy(json_filename, saved_path)
def val_horizon_data():
#txt_path = ".\\labelme_to_voc_widerface\\save_widerface_result_horizon.txt"
txt_path = ".\\labelme_to_voc_widerface\\save_widerface_result_blur.txt"
f = open(txt_path, 'r')
lines = f.readlines()
isFirst = True
imgs_path = []
words = []
labels = []
for line in lines:
line = line.rstrip()
if line.startswith('#'):
if isFirst is True:
isFirst = False
else:
labels_copy = labels.copy()
words.append(labels_copy)
#labels.clear()
path = line[2:]
#path = txt_path.replace('save_widerface_result_horizon.txt', 'VOC2007\\JPEGImages\\') + path
path = txt_path.replace('save_widerface_result_blur.txt', 'VOC2007\\JPEGImages\\') + path
imgs_path.append(path)
else:
line = line.split(' ')
label = [float(x) for x in line]
labels.append(label)
for index in range(0, len(imgs_path)):
img_raw = cv2.imread(imgs_path[index])
#print(img_raw.shape)
#print(int(labels[index][0]))
cv2.rectangle(img_raw, (int(labels[index][0]), int(labels[index][1])), (int(labels[index][0])+int(labels[index][2]), int(labels[index][1])+int(labels[index][3])), (0, 0, 255), 2)
cv2.circle(img_raw, (int(labels[index][5]), int(labels[index][6])), 2, (255, 255, 0), 8)
cv2.circle(img_raw, (int(labels[index][7]), int(labels[index][8])), 2, (0, 255, 255), 8)
cv2.circle(img_raw, (int(labels[index][9]), int(labels[index][10])), 2, (255, 0, 255), 8)
cv2.circle(img_raw, (int(labels[index][11]), int(labels[index][12])), 2, (0, 255, 0), 8)
#saved_path = os.path.join(os.path.dirname(imgs_path[index]), os.path.basename(imgs_path[index]).split(".")[0]+"_1.jpg")
saved_path = os.path.join(".\\labelme_to_voc_widerface\\save_result", os.path.basename(imgs_path[index]).split(".")[0]+"_1.jpg")
cv2.imwrite(saved_path, img_raw)
#shutil.copy(imgs_path[index], saved_path)
def data_album(dataDir, rename):
list_dirs = os.walk(dataDir)
for root, dirs, files in list_dirs:
# for d in dirs:
# print("@@@@@@@@@@@@@@@@@@@@")
# print(os.path.join(root,d))
for f in files:
#print("********************")
#./images\VOC2007\ImageSets\Main
#print(root)
path_name = root
#./images\VOC2007\ImageSets\Main\val.txt
#print(os.path.join(root,f))
pic_path = os.path.join(root,f)
new_img = data_aug_pic(pic_path,rename)
cv2.imwrite(os.path.join(".\\data_album", os.path.basename(pic_path).split(".")[0]+rename+".jpg"), new_img)
#data_aug()
#将label为1的 jpg和json复制在新的文件夹
#labelme_json_voc_widerface_copy()
#验证水平翻转和模糊后关键点和框的正确性，画出来
#val_horizon_data()
#单纯的对图片做模糊和水平翻转，没有坐标值和数据
# data_album(".\\data_album","_horizon")
data_album(".\\data_album","_Blur")
#对labelme标注数据做增强，并将增强的数据转换成voc和widerface格式
#labelme_json_voc_widerface("_horizon")
# labelme_json_voc_widerface("_Blur")