Got it, let's break this down simply—no need to mess with C++ code or config files directly. tesserocr lets you set those same Tesseract parameters right in your Python script using the SetVariable method.

1. Restricting to a Custom Set of Characters #

Just like the tessedit_char_whitelist config in C++, you can define exactly which characters Tesseract should look for. Here's how to do it in Python:

from tesserocr import PyTessBaseAPI

# Use a context manager to handle the API instance cleanly
with PyTessBaseAPI() as api:
    # Define your allowed characters here (adjust as needed)
    allowed_chars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789!@#$%"
    api.SetVariable("tessedit_char_whitelist", allowed_chars)
    
    # Load your image and run recognition
    api.SetImageFile("your_target_image.png")
    recognized_text = api.GetUTF8Text()
    print(recognized_text.strip())

The SetVariable method directly maps to Tesseract's internal configuration—you're essentially passing the same parameter you'd set in the C++ config file, but straight from your Python code.

2. Restricting to Only Digits #

For your specific use case of limiting recognition to numbers, just set the whitelist to the digits 0-9:

from tesserocr import PyTessBaseAPI

with PyTessBaseAPI() as api:
    # Restrict recognition to digits only
    api.SetVariable("tessedit_char_whitelist", "0123456789")
    
    api.SetImageFile("image_with_numbers.png")
    print(api.GetUTF8Text().strip())

Quick Bonus: Blacklisting Characters #

If you ever need to exclude specific characters instead of whitelisting, use tessedit_char_blacklist instead. For example, to block all symbols:

api.SetVariable("tessedit_char_blacklist", "!@#$%^&*()")

Hope this works for you—no C++ docs required! 😊

内容的提问来源于stack exchange，提问作者WesR