ocr_engine.py

import easyocr
import numpy as np
import re
import cv2

reader = easyocr.Reader(['en'], gpu=False)

def extract_weight_from_image(pil_img):
    try:
        img = np.array(pil_img)

        # Convert to grayscale
        gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)

        # Apply bilateral filter to reduce noise while keeping edges
        filtered = cv2.bilateralFilter(gray, 11, 17, 17)

        # Apply binary threshold
        _, thresh = cv2.threshold(filtered, 150, 255, cv2.THRESH_BINARY_INV)

        # Find contours
        contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

        if not contours:
            return "No weight detected", 0.0

        # Get the largest contour assuming it's the display area
        largest_contour = max(contours, key=cv2.contourArea)
        x, y, w, h = cv2.boundingRect(largest_contour)

        # Add padding
        pad = 10
        x, y = max(x - pad, 0), max(y - pad, 0)
        cropped = gray[y:y+h+pad, x:x+w+pad]

        # OCR on cropped area
        result = reader.readtext(cropped, detail=0)
        combined = " ".join(result)
        print("Detected Text:", combined)

        # Match weight patterns like 52.30 or 003.25
        match = re.search(r"\b\d{2,4}\.?\d{0,2}\b", combined)
        if match:
            return match.group(), 95.0
        else:
            return "No weight detected", 0.0

    except Exception as e:
        return f"Error: {str(e)}", 0.0