当前位置：网站首页>Image batch processing ｜ necessary skills

Image batch processing ｜ necessary skills

2022-07-28 08:47:00 【51CTO】

WeChat official account ：*AI Algorithm and image processing
Follow to learn more about programming tips and quality resources . A question or suggestion , Please leave a message for the official account ;
If you think it helps , welcome Good looking and sharing

Content catalog

Image batch processing brief introduction Specific matters 1） Analyze the storage of files 2） Try to analyze good , Now it's clear , It's time to prepare something Code implementation Output results ： At this time, the name and address of the picture have been obtained —— You can start working Result display

Image batch processing brief introduction

Image batch operation or data batch processing is often used in this era of data explosion , Except for special circumstances , Otherwise, we don't want to rely on manpower to complete data processing one by one （ Local steelmaking ？）, Therefore, the coolies handed over to us —— The computer .
Clear objectives ： Spend time doing something valuable
Ideas ：
1. Read files in bulk
2. Batch processing files
3. Batch storage of files
abbreviation ——3P operation ？ Vulgar words ...

Specific matters

1） Analyze the storage of files

Take a look at what the storage of files looks like

Image batch processing ｜ Essential skills _ mapping

0 The folder is under 3 Small folder （ More common situation ）

Image batch processing ｜ Essential skills _ data _02

In each secondary folder （ for example 1_1 Folder ） There are pictures stored under

Image batch processing ｜ Essential skills _ data _03

Skull pain after watching ！

Image batch processing ｜ Essential skills _ data _04

2） Try to analyze

explain ： For ease of understanding , I name folders regularly , from 0 Start adding to the next level 1

(1) Read files in bulk Turn into —— Three steps ：

A） Navigate to the first level folder （0） Location
B） Read the secondary folder （1_1,1_2,1_3） name
C） Read the pictures under the secondary folder （ Or data ） name

good , Now it's clear , It's time to prepare something

A） Import library
Os —— Interface of operating system functions
Tqdm —— Add a progress bar , Convenient installation , No , Is to check the implementation
cv2 —— Do something with the picture , Read storage and changes

B） Picture preparation
Here are two in each folder

(2) Batch processing pictures

Here you can add various scanning operations you like , As long as it doesn't break the law , Do whatever you want , For the sake of simplicity , I don't do fancy , Grayscale change operation is used in the code

(3) Batch storage of files

Understand the batch reading of files , In fact, batch storage of files is similar

Code implementation

       
       '''
       
 Read files in bulk （）
       
'''
       
import os
       
from tqdm import tqdm
       
'''
       
1. Set the necessary read path and storage path 
       
'''
       
# path = os.getcwd() #  Get the current path 
       
#  The processed image is not under the current file , You can add absolute paths directly 
       
path = '/media/pzw/0E50196C0E50196C/weixin/MultiProcess/0'
       
## print(path) #  It's customary to make sure every step is correct , You can directly ignore 
       
dst_path = '/media/pzw/0E50196C0E50196C/weixin/MultiProcess' #  The image storage location after processing 
       
'''
       
2. Enter the secondary folder and read the file 
       
'''
       
dir_names = os.listdir(path)    #  Read the name of the secondary folder 
       
# print(dir_names)
       
for dir in tqdm(dir_names):
       
    path_1 = os.path.join(path, dir) #  Get the path of the secondary folder  || secondary 1 start 
       
    # print(path_1)
       
    pic_names = os.listdir(path_1)    # Read the name of the picture 
       
    print(pic_names)
      
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.

Output results ：

100%|██████████| 3/3 [00:00<00:00, 5863.43it/s]
['1_1_1.jpg', '1_1_2.jpg']
['1_2_1.jpg', '1_2_2.jpg']
['1_3_1.jpg', '1_3_2.jpg']

At this time, the name and address of the picture have been obtained —— You can start working

The next work is equivalent to the operation of a single picture , Because we cycle to the bottom （ It may not be very accurate , Probably means I found the picture , What do you want? Now ）
The code is for you , The notes are very clear ！！！

       
       import os
       
import cv2
       
from tqdm import tqdm
       
'''
       
1. Set the necessary read path and storage path 
       
'''
       
# path = os.getcwd() #  Get the current path 
       
#  The processed image is not under the current file , You can add absolute paths directly 
       
path = '/media/pzw/0E50196C0E50196C/weixin/MultiProcess/0'   #  First level folder  0  start 
       
## print(path) #  It's customary to make sure every step is correct , You can directly ignore 
       
dst_path0 = '/media/pzw/0E50196C0E50196C/weixin/MultiProcess/0_pro' #  The image storage location after processing 
       
'''
       
2. Enter the secondary folder and read the file 
       
'''
       
dir_names = os.listdir(path)    #  Read the name of the secondary folder 
       
# print(dir_names)
       
# tqdm Is to display the current read to n It's a folder , Of course, you can change to another position , If you don't know me, please read the previous article 
       
for dir in tqdm(dir_names):
       
    path_1 = os.path.join(path, dir) #  Get the path of the secondary folder  || secondary 1 start 
       
    # print(path_1)
       
    pic_names = os.listdir(path_1)    # Read the name of the picture 
       
    # print(pic_names)
       
    dst_path1 = os.path.join(dst_path0,dir) #  Corresponding secondary storage path 
       
    '''
       
    3. Working with pictures 
       
     At this time, it can be considered as an operation in a single file , After the operation in this folder is completed ,
       
     The next folder will be executed , In this way, it is easy to understand how to do 
       
     This passage may affect your understanding here , But it's to tell you how to think . You can delete it when reading the code 
       
    '''
       
    for pic_name in pic_names:
       
        new_path = os.path.join(path_1, pic_name)
       
        # print(new_path) #  The reading path of the picture 
       
        dst_pic_name = pic_name.split('.')[0]+'>_<.jpg' #  Modify the name of the processed picture , I will cry after being repaired 
       
        # print(dst_pic_name)
       
        dst_path = os.path.join(dst_path1, dst_pic_name)
       
        # print(dst_path)
       
        #  Read the picture 
       
        img_0 = cv2.imread(new_path, cv2.IMREAD_GRAYSCALE) #  Become a grayscale image 
       
        #  Here you can add the operations you want , tailoring , Change and so on 
       
        #  Determine whether the storage path exists , non-existent , Create a new path 
       
        isExists = os.path.exists(dst_path1)
       
        if not isExists:
       
            os.makedirs(dst_path1)
       
        cv2.imwrite(dst_path, img_0)
      
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.