ceshi.py

import cv2
import os
import numpy as np

###########可以将train中有缺陷图片删除
# a=0
# for i in os.listdir('./dataset/train'):
#     path=os.path.join('./dataset/train',i)
#     img=cv2.imread(path)
#     img1=img[0,0,0]
#     img2 = img[0, 511, 0]
#     img3 = img[0, 1023, 0]
#     img4 = img[0, 1535, 0]
#     img5 = img[0, 2023, 0]
#     img6 = img[511, 0, 0]
#     img7 = img[511, 511, 0]
#     img8 = img[511, 1023, 0]
#     img9 = img[511, 1535, 0]
#     img10 = img[511, 2023, 0]
#
#     if img1>0:
#         if img2<1 or img3<1 or img4<1 or img5<1 or img6<1 or img7<1 or img8<1 or img9<1 or img10<1:
#             print(i)
#         else:
#             cv2.imwrite('./train/{}.jpg'.format(a),img,[int(cv2.IMWRITE_JPEG_QUALITY),100])
#             a=a+1
#     else:
#         if img2<1 and img6<1 and img7<1:
#             cv2.imwrite('./train/{}.jpg'.format(a), img, [int(cv2.IMWRITE_JPEG_QUALITY), 100])
#             a = a + 1

a=0
for i in os.listdir('./test'):
    path=os.path.join('./test',i)
    img1=[]
    img2=[]
    img3=[]
    img4=[]
    for j in os.listdir(path):
        path1=os.path.join(path,j)
        if 'A' in j:
            img1=cv2.imread(path1)
        elif 'B1' in j:
            img2=cv2.imread(path1)[:768,:,:]
        elif 'B2' in j:
            img3=cv2.imread(path1)[:768,:,:]
        elif 'B3' in j:
            img4=cv2.imread(path1)
    img=np.hstack((img1,img2,img3,img4))
    cv2.imwrite('./test/{}.jpg'.format(a),img,[int(cv2.IMWRITE_WEBP_QUALITY),100])
    a=a+1




#
# for i in os.listdir('./org'):
#     path=os.path.join('./org',i)
#     for j in os.listdir(path):
#         if 'B3' in j:
#             path1=os.path.join(path,j)
#             img=cv2.imread(path1)
#             cv2.imwrite('./aaa/{}'.format(j),img,[int(cv2.IMWRITE_JPEG_QUALITY),100])