读取海量图像文件和Label标签，并保存数据

Halcom

读取海量图像文件和Label标签，并保存数据：
（1）准备我们的图像数据，Train文件夹图像+Labels、Test文件夹图像+Labels
（2）读取Train文件夹图像+Labels、Test文件夹图像+Labels存储为Tensorflow需要的数据格式
（3）写入.pkl文件，方便下次加载直接使用；

1. # -*- coding: utf-8 -*-
   
2. """
   
3. Created on Sat Aug 26 21:52:50 2017
   
4. 
   
5. @author: Administrator
   
6. """
   
7. import pickle
   
8. from PIL import Image
   
9. import numpy as np
   
10. import os
    
11. 
    
12. def load_Img(imgDir,imgFoldName):
    
13.     imgs = os.listdir(imgDir+imgFoldName)
    
14.     imgNum = len(imgs)
    
15.     data = np.empty((imgNum,28,28,1),dtype="uint8")
    
16.     for i in range (imgNum):
    
17.         # img = Image.open(imgDir+imgFoldName+"/"+imgs[i])
    
18.         img = Image.open(imgDir+imgFoldName+"/"+ str(i) + '.png' )
    
19.         arr = np.asarray(img,dtype="uint8")
    
20.         arr = np.reshape(arr, [28,28,1] )
    
21.         data[i,:,:,:] = arr
    
22.     return data
    
23. 
    
24. def dense_to_one_hot(labels_dense, num_classes):
    
25.   """Convert class labels from scalars to one-hot vectors."""
    
26.   num_labels = labels_dense.shape[0]
    
27.   index_offset = np.arange(num_labels) * num_classes
    
28.   labels_one_hot = np.zeros((num_labels, num_classes))
    
29.   labels_one_hot.flat[index_offset.astype(int) + labels_dense.ravel().astype(int)] = 1
    
30.   return labels_one_hot
    
31. 
    
32. def load_Label(LabelDir,LabelFoldName, one_hot, num_classes):
    
33.     Labs = os.listdir(LabelDir+LabelFoldName)
    
34.     File = open(LabelDir+LabelFoldName + Labs[1])
    
35.     File.seek(0,0)
    
36.     Dstr = File.read();
    
37.     data = str.split(Dstr,',')
    
38.     Label = np.array(data, dtype=np.float64)
    
39.     File.close()
    
40.     if one_hot:
    
41.         return dense_to_one_hot(Label, num_classes);
    
42.     return Label
    
43.    
    
44. imgDir = "./Test/"
    
45. imgFoldName = "test"
    
46. test_images = load_Img("./Test/","test")
    
47. train_images = load_Img("./Train/","train")
    
48. LabelDir =  "D:/Deep_learning_TensorFlow_python/Learning_DayByDay/Mnist/Test/"
    
49. LabelFoldName = ""
    
50. train_labels = load_Label("./Train/","", True, 10)
    
51. test_labels = load_Label("./Test/","", True, 10)
    
52. 
    
53. # 保存数据
    
54. pickle_data = open('./PKL_mnist/train_images.pkl','wb');
    
55. pickle.dump(train_images, pickle_data);
    
56. pickle_data.close();
    
57. 
    
58. pickle_data = open('./PKL_mnist/test_images.pkl','wb');
    
59. pickle.dump(test_images, pickle_data);
    
60. pickle_data.close();
    
61. 
    
62. pickle_data = open('./PKL_mnist/train_labels.pkl','wb');
    
63. pickle.dump(train_labels, pickle_data);
    
64. pickle_data.close();
    
65. 
    
66. pickle_data = open('./PKL_mnist/test_labels.pkl','wb');
    
67. pickle.dump(test_labels, pickle_data);
    
68. pickle_data.close();
    
69. 
    
70. print('写入PKL文件完成！！！')

复制代码

百度网盘链接：
链接：http://pan.baidu.com/s/1i4BbOlJ
具体链接在halcom.cn论坛，联系人QQ：3283892722
该论坛是一个学习交流平台，我会逐一的和大家分享学习。
欢迎大家录制视频，你可在论坛进行打赏分享。
视频专用播放器：http://halcom.cn/forum.php?mod=viewthread&tid=258&extra=page%3D1