-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathmaskCreator.py
39 lines (28 loc) · 1.03 KB
/
maskCreator.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
import os
import glob
import cv2
import PIL.Image as Image
import torchvision.transforms as transforms
from skimage.color import rgb2yuv
import numpy as np
if __name__ == '__main__':
root1 = "E:/RoboCup/YOLO/Train/"
root2 = "E:/RoboCup/FinetuneHorizon/train/"
imgSize = (256,192)
imSize = (192,256)
resize = transforms.Resize(imSize)
labResize = transforms.Resize(imSize,Image.NEAREST)
imgs = sorted(glob.glob1(root1,"*.png"))
labels = sorted(glob.glob1(root2,"*.png"))
if len(labels) != len(imgs):
for i in imgs:
img = cv2.resize(cv2.imread(root1 + i), imgSize)
cv2.imwrite(root1+i,img)
else:
for i,l in zip(imgs,labels):
img = cv2.cvtColor(cv2.cvtColor(cv2.resize(cv2.imread(root1+i),(160,120)),cv2.COLOR_BGR2YUV),cv2.COLOR_BGR2RGB)
#img = cv2.resize(cv2.imread(root1+i),imgSize)
label = Image.open(root2+l).convert('I')
label = labResize(label)
label.save(root2+l)
cv2.imwrite(root1+i,img)