torchvision.transforms的使用


参考链接: torchvision.transforms

参考链接: 透视变换原理实例代码详解
参考链接: 图像几何变换之透视变换

打开并显示图片文件:

1
2
3
4
5
6
7
8
9
10
11
from torchvision import transforms
from matplotlib import pyplot as plt
import numpy as np
import torch
from PIL import Image

img = Image.open("KJ.jpg")
img.show()

plt.imshow(img)
plt.show()

运行结果展示:
在这里插入图片描述

PIL image转为tensor类型:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
from torchvision import transforms
from matplotlib import pyplot as plt
import numpy as np
import torch
from PIL import Image

img = Image.open("KJ.jpg")

to_tensor = transforms.ToTensor()
img_t = to_tensor(img) # 注意：这里会自动调整通道维度的位置，即通道维度放到最开头
print('img_t.shape:',img_t.shape)  # img_t.shape: torch.Size([3, 374, 500])
print('type(img_t):',type(img_t))  # type(img_t): <class 'torch.Tensor'>
print(img_t.shape, img_t.dtype)  # torch.Size([3, 374, 500]) torch.float32

plt.imshow(img_t.permute(1, 2, 0))  # 需要将通道所对应的维度放到最后
plt.show()

运行结果:
在这里插入图片描述
中心裁剪加随机旋转:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
from torchvision import transforms
from matplotlib import pyplot as plt
import numpy as np
import torch
from PIL import Image

img = Image.open("KJ.jpg")
trans4cxq = transforms.Compose([
    transforms.CenterCrop((374,400)),
    transforms.RandomRotation([-180,180]),
    transforms.ToTensor(),
    ])

img_tr = trans4cxq(img)
plt.imshow(img_tr.permute(1, 2, 0))  # 需要将通道所对应的维度放到最后
plt.show()

运行结果展示:
在这里插入图片描述

透视变换:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
from torchvision import transforms
from matplotlib import pyplot as plt
import numpy as np
import torch
from PIL import Image

img = Image.open("KJ.jpg")
scale = 5.0
img = Image.open("poker.png")
img_func = transforms.functional.perspective(
    img,
    [[63,72], [162,30], [270,144], [151,218]],
    [[10*scale,20*scale], [30*scale,20*scale], [30*scale,40*scale], [10*scale,40*scale]],
    interpolation=3)

img_func.show()

plt.imshow(img_func)  
plt.show()

所使用的图片和运行结果展示:
在这里插入图片描述
在这里插入图片描述

使用函数来实现随机旋转,并且添加填充:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
from torchvision import transforms
from matplotlib import pyplot as plt
import numpy as np
import torch
from PIL import Image

img = Image.open("KJ.jpg")
img = Image.open("KJ.jpg")
img_func = transforms.functional.rotate(
    img, 45,
    resample=False,
    expand=True,
    center=None)

img_func.show()
plt.imshow(img_func)  
plt.show()

运行效果:
在这里插入图片描述

完整实验代码:

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
from torchvision import transforms
from matplotlib import pyplot as plt
import numpy as np
import torch
from PIL import Image

img = Image.open("KJ.jpg")
# img.show()

# plt.imshow(img)
# plt.show()



# to_tensor = transforms.ToTensor()
# img_t = to_tensor(img) # 注意：这里会自动调整通道维度的位置，即通道维度放到最开头
# print('img_t.shape:',img_t.shape)  # img_t.shape: torch.Size([3, 374, 500])
# print('type(img_t):',type(img_t))  # type(img_t): <class 'torch.Tensor'>
# print(img_t.shape, img_t.dtype)  # torch.Size([3, 374, 500]) torch.float32

# plt.imshow(img_t.permute(1, 2, 0))  # 需要将通道所对应的维度放到最后
# plt.show()



# trans4cxq = transforms.Compose([
#     transforms.CenterCrop((374,400)),
#     transforms.RandomRotation([-180,180]),
#     transforms.ToTensor(),
#     ])

# img_tr = trans4cxq(img)
# plt.imshow(img_tr.permute(1, 2, 0))  # 需要将通道所对应的维度放到最后
# plt.show()





# scale = 5.0
# img = Image.open("poker.png")
# img_func = transforms.functional.perspective(
#     img,
#     [[63,72], [162,30], [270,144], [151,218]],
#     [[10*scale,20*scale], [30*scale,20*scale], [30*scale,40*scale], [10*scale,40*scale]],
#     interpolation=3)

# img_func.show()

# plt.imshow(img_func)  
# plt.show()






# scale = 100.0
# img = Image.open("cxq.jpg")
# img_func = transforms.functional.perspective(
#     img,
#     [[3097,3015], [1649,3405], [591,1393], [1761,1477]],
#     [[10*scale,20*scale], [30*scale,20*scale], [30*scale,40*scale], [10*scale,40*scale]],
#     interpolation=3)

# img_func.show()





img = Image.open("KJ.jpg")
img_func = transforms.functional.rotate(
    img, 45,
    resample=False,
    expand=True,
    center=None)

img_func.show()
plt.imshow(img_func)  
plt.show()

实验所使用的图片文件:

在这里插入图片描述
在这里插入图片描述