-
Notifications
You must be signed in to change notification settings - Fork 21
/
augment.py
203 lines (178 loc) · 6.87 KB
/
augment.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
import os
import numpy as np
import cv2
import random
def crop_img(src,top_left_x,top_left_y,crop_w,crop_h):
'''裁剪图像
Args:
src: 源图像
top_left,top_right:裁剪图像左上角坐标
crop_w,crop_h:裁剪图像宽高
return:
crop_img:裁剪后的图像
None:裁剪尺寸错误
'''
rows, cols = src.shape[0: 2]
row_min,col_min = int(top_left_y), int(top_left_x)
row_max,col_max = int(row_min + crop_h), int(col_min + crop_w)
if row_max > rows or col_max > cols:
print("crop size err: src->%dx%d,crop->top_left(%d,%d) %dx%d"%(cols, rows, col_min, row_min,int(crop_w),int(crop_h)))
return None
crop_img = src[row_min:row_max, col_min:col_max]
return crop_img
def crop_imgs(img, label, crop_type='RANDOM_CROP',crop_n=1, dsize=(0, 0), random_wh=False):
'''
Args:
imgs_dir: 待放缩图片
crop_type:裁剪风格 ['RANDOM_CROP','CENTER_CROP','FIVE_CROP']
crop_n: 每原图生成裁剪图个数
dsize:指定crop宽高(w,h),与random_wh==True互斥生效
random_wh:随机选定裁剪宽高
'''
imgh, imgw = img.shape[0: 2]
# fw, fh: 当random_wh == False时为crop比例,否则为随机crop的宽高比例下限
fw = random.uniform(0.2, 0.98)
fh = random.uniform(0.2, 0.98)
crop_imgw, crop_imgh = dsize
if dsize == (0, 0) and not random_wh:
crop_imgw = int(imgw * fw)
crop_imgh = int(imgh * fh)
elif random_wh:
crop_imgw = int(imgw * (fw + random.random() * (1 - fw)))
crop_imgh = int(imgh * (fh + random.random() * (1 - fh)))
if crop_type == 'RANDOM_CROP':
crop_top_left_x, crop_top_left_y = random.randint(0, imgw - crop_imgw - 1), random.randint(0, imgh - crop_imgh - 1)
elif crop_type == 'CENTER_CROP':
crop_top_left_x, crop_top_left_y = int(imgw / 2 - crop_imgw / 2), int(imgh / 2 - crop_imgh / 2)
elif crop_type == 'FIVE_CROP':
crop_top_left_x, crop_top_left_y = 0, 0
else:
print('crop type wrong! expect [RANDOM_CROP,CENTER_CROP,FIVE_CROP]')
croped_img = crop_img(img, crop_top_left_x, crop_top_left_y, crop_imgw, crop_imgh)
croped_label = crop_img(label, crop_top_left_x, crop_top_left_y, crop_imgw, crop_imgh)
#丢弃正样本较少的
tmp = croped_label.copy()
tmp[tmp > 0] = 1
if np.sum(tmp) < 500:
return img, label
else:
return croped_img, croped_label
def rot_img_and_padding(img, rot_angle, scale=1.0):
'''
以图片中心为原点旋转
Args:
img:待旋转图片
rot_angle:旋转角度,逆时针
scale:放缩尺度
return:
imgRotation:旋转后的cv图片
'''
img_rows, img_cols = img.shape[:2]
cterxy = [img_cols//2, img_rows//2]
matRotation = cv2.getRotationMatrix2D((cterxy[0], cterxy[1]), rot_angle, scale)
imgRotation = cv2.warpAffine(img, matRotation, (img_cols, img_rows))
return imgRotation
def rand_rot(img, label):
'''
:param img: [H, W, 3]
:param lable: [H, W, 2]
:return:
'''
angle = random.randint(0, 180)
scale = random.uniform(0.9, 1.5)
res_img = rot_img_and_padding(img, angle, scale)
res_label = rot_img_and_padding(label, angle, scale)
return res_img, res_label
def rand_flip(img, label):
'''图片翻转'''
flag = random.random()
if flag < 0.3333:
res_img = cv2.flip(img, 1)
res_label = cv2.flip(label, 1)
elif (flag >= 0.3333) and (flag < 0.6666):
res_img = cv2.flip(img, -1)
res_label = cv2.flip(label, -1)
else:
res_img = cv2.flip(img, 0)
res_label = cv2.flip(label, 0)
return res_img, res_label
def random_color_distort(img, label, brightness_delta=32, hue_vari=18, sat_vari=0.5, val_vari=0.5):
'''
在图片的HSV空间进行扭曲,还有亮度调整
randomly distort image color. Adjust brightness, hue, saturation, value.
param:
img: a BGR uint8 format OpenCV image. HWC format.
'''
def random_hue(img_hsv, hue_vari, p=0.5):
if np.random.uniform(0, 1) > p:
hue_delta = np.random.randint(-hue_vari, hue_vari)
img_hsv[:, :, 0] = (img_hsv[:, :, 0] + hue_delta) % 180
return img_hsv
def random_saturation(img_hsv, sat_vari, p=0.5):
if np.random.uniform(0, 1) > p:
sat_mult = 1 + np.random.uniform(-sat_vari, sat_vari)
img_hsv[:, :, 1] *= sat_mult
return img_hsv
def random_value(img_hsv, val_vari, p=0.5):
if np.random.uniform(0, 1) > p:
val_mult = 1 + np.random.uniform(-val_vari, val_vari)
img_hsv[:, :, 2] *= val_mult
return img_hsv
def random_brightness(img, brightness_delta, p=0.5):
if np.random.uniform(0, 1) > p:
img = img.astype(np.float32)
brightness_delta = int(np.random.uniform(-brightness_delta, brightness_delta))
img = img + brightness_delta
return np.clip(img, 0, 255)
# brightness
img = random_brightness(img, brightness_delta)
img = img.astype(np.uint8)
# color jitter
img_hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV).astype(np.float32)
if np.random.randint(0, 2):
img_hsv = random_value(img_hsv, val_vari)
img_hsv = random_saturation(img_hsv, sat_vari)
img_hsv = random_hue(img_hsv, hue_vari)
else:
img_hsv = random_saturation(img_hsv, sat_vari)
img_hsv = random_hue(img_hsv, hue_vari)
img_hsv = random_value(img_hsv, val_vari)
img_hsv = np.clip(img_hsv, 0, 255)#限幅
img = cv2.cvtColor(img_hsv.astype(np.uint8), cv2.COLOR_HSV2BGR)#转换色彩空间
return img, label
def tranc(img, label):
img = cv2.transpose(img)
label = cv2.transpose(label)
return img, label
def rand_augment(img, label):
'''随机选择一种数据增强'''
# flag = random.random()
# print(flag)
if random.random() < 0.5:
# 随机裁剪
res_img, res_label = crop_imgs(img, label)
if random.random() < 0.5:
res_img, res_label = tranc(res_img, res_label)
elif random.random() < 0.5:
# 随机翻转
res_img, res_label = rand_flip(img, label)
if random.random() < 0.5:
res_img, res_label = tranc(res_img, res_label)
# elif (flag >= 0.5) and (flag < 0.75):
# # 随机旋转
# res_img, res_label = rand_rot(img, label)
elif random.random() < 0.5:
# 随机色度变换
res_img, res_label = random_color_distort(img, label)
if random.random() < 0.5:
res_img, res_label = tranc(res_img, res_label)
else:
res_img, res_label = img, label
return res_img, res_label
if __name__ == '__main__':
img = cv2.imread('./textimg/image.png')
label = cv2.imread('./textimg/weight.png')
label = label[:, :, 0:2]
res_i, res_l = rand_augment(img, label)
cv2.imshow('s', res_i)
cv2.waitKey()