Extracting all text from a video

[jehoshua]

I have run the following code on a sample video/mp4 to extract all images. It does it frame by frame, so there are many thousands of image files created, ..lol

#!/usr/bin/env python

# Importing all necessary libraries 
import cv2 
import os 
  
# Read the video from specified path 
cam = cv2.VideoCapture("/home/********/Downloads/Python_scripts/test1.mp4") 
  
try: 
      
    # creating a folder named data 
    if not os.path.exists('data'): 
        os.makedirs('data') 
  
# if not created then raise error 
except OSError: 
    print ('Error: Creating directory of data') 
  
# frame 
currentframe = 0
  
while(True): 
      
    # reading from frame 
    ret,frame = cam.read() 
  
    if ret: 
        # if video is still left continue creating images 
        name = './data/frame' + str(currentframe) + '.jpg'
        print ('Creating...' + name) 
  
        # writing the extracted images 
        cv2.imwrite(name, frame) 
  
        # increasing counter so that it will 
        # show how many frames are created 
        currentframe += 1
    else: 
        break
  
# Release all space and windows once done 
cam.release() 
cv2.destroyAllWindows() 

but as I require the text, not the images, have tested the following on just one file ..

#!/usr/bin/env python

import subprocess
subprocess.run(["tesseract", "data/frame998.jpg",  "stdout"])

and called it as follows

python3 subprocess_test.py

and it works perfectly. Now I wanted to modify the first block of code to NOT write out the image files, but simply convert to text, and then if the text is the same as the last processed text, next frame, ..ELSE write out a text file and save the contents. This appears to work okay..

#!/usr/bin/env python

# Importing all necessary libraries 
import cv2 
import os 

#import subprocess
import pytesseract

# Read the video from specified path 
cam = cv2.VideoCapture("/home/********/Downloads/Python_scripts/test1.mp4") 

# Set the text file contents to null
text_file_old = ""
  
try: 
      
    # creating a folder named data 
    if not os.path.exists('data'): 
        os.makedirs('data') 
  
# if not created then raise error 
except OSError: 
    print ('Error: Creating directory of data') 
  
# frame 
currentframe = 0
  
while(True): 
      
    # reading from frame 
    ret,frame = cam.read() 
  
    if ret: 
        # if video is still left continue creating images 
        name = './data/frame' + str(currentframe) + '.jpg'

        text_file_new =  pytesseract.image_to_string(frame)
        
        if text_file_new != text_file_old:
            
            # write contents to a file
            filename = './data/text' + str(currentframe) + '.txt'
            file1 = open(filename, "w")
            file1.write(text_file_new) 
            file1.close() 

            text_file_old = text_file_new
        
        # increasing counter so that it will 
        # show how many frames are created 
        currentframe += 1
        
    else: 
        break
  
# Release all space and windows once done 
cam.release() 
cv2.destroyAllWindows() 

With the writing of the text file, can it be done differently or more efficiently ? The script has been running only about 5 minutes, yet running fdupes already show a few duplicate files (contents). No doubt if the video showed a frame123, then frame124 and then output frame123 again, there appears to be little I can do in the python code/execution. Just run fdupes afterwards ??

Also, some of the text is nearly the same. This is where there is two frames with the same text content, yet the frame image may be slightly different due to image distortions, etc. Can CV2 or pytesseract clean the image ??