Preparing an Image for OCR

import cv2
import numpy as np
import pytesseract

def preprocess_for_ocr(image_path):
    """
    Complete OCR pre-processing pipeline.
    Each step improves accuracy for different types of images.
    """
    image = cv2.imread(image_path)

    # Step 1: Convert to grayscale
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

    # Step 2: Remove noise
    denoised = cv2.fastNlMeansDenoising(gray, h=10)

    # Step 3: Increase contrast (CLAHE)
    clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
    enhanced = clahe.apply(denoised)

    # Step 4: Binarise (threshold)
    _, binary = cv2.threshold(enhanced, 0, 255,
                              cv2.THRESH_BINARY + cv2.THRESH_OTSU)

    # Step 5: Remove noise from binary image
    kernel = np.ones((1, 1), np.uint8)
    cleaned = cv2.morphologyEx(binary, cv2.MORPH_OPEN, kernel)
    cleaned = cv2.morphologyEx(cleaned, cv2.MORPH_CLOSE, kernel)

    # Step 6: Add padding (Tesseract works better with margins)
    padded = cv2.copyMakeBorder(cleaned, 10, 10, 10, 10,
                                cv2.BORDER_CONSTANT, value=255)

    return padded

# Compare OCR on raw vs. preprocessed image
image = cv2.imread("document.jpg")
preprocessed = preprocess_for_ocr("document.jpg")

raw_text = pytesseract.image_to_string(image)
processed_text = pytesseract.image_to_string(preprocessed)

print(f"Raw characters: {len(raw_text)}")
print(f"Processed characters: {len(processed_text)}")
print("\nProcessed result:")
print(processed_text[:300])