python-snippets/ocr/README.md

"""Get image timestamps from Swift camera"""

import os from glob import glob from datetime import datetime from tqdm import tqdm from PIL import Image import pandas as pd import pytesseract

input_dir = 'jpg'

jpg_names = glob(os.path.join(input_dir, '*.jpg')) dates = []

for jpg_name in tqdm(jpg_names): im = Image.open(jpg_name)

# Crop image
w, h = im.size
im = im.crop((w - 550, h - 40, w, h))

# Perform OCR with tesseract
text = pytesseract.image_to_string(im, lang='eng')

# Convert to datetime object
dates.append(datetime.strptime(text, '%d/%m/%Y %H:%M:%S'))

Save as csv

df = pd.DataFrame(data=dates, index=jpg_names, columns=['date']) df.index.name = 'file' df.to_csv('image-dates.csv')