ocr_engine.py

import pytesseract
import numpy as np
import cv2
import re
import logging
from datetime import datetime
import os
from PIL import Image

# Set up logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')

# Directory for debug images
DEBUG_DIR = "debug_images"
os.makedirs(DEBUG_DIR, exist_ok=True)

def save_debug_image(img, filename_suffix, prefix=""):
    """Save image to debug directory with timestamp."""
    timestamp = datetime.now().strftime("%Y%m%d_%H%M%S_%f")
    filename = os.path.join(DEBUG_DIR, f"{prefix}{timestamp}_{filename_suffix}.png")
    if isinstance(img, Image.Image):
        img.save(filename)
    elif len(img.shape) == 3:
        cv2.imwrite(filename, cv2.cvtColor(img, cv2.COLOR_BGR2RGB))
    else:
        cv2.imwrite(filename, img)
    logging.info(f"Saved debug image: {filename}")

def estimate_brightness(img):
    """Estimate image brightness."""
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    return np.mean(gray)

def preprocess_image(img):
    """Preprocess image for OCR with aggressive contrast and noise reduction."""
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    brightness = estimate_brightness(img)
    # Aggressive CLAHE
    clahe_clip = 6.0 if brightness < 80 else 4.0
    clahe = cv2.createCLAHE(clipLimit=clahe_clip, tileGridSize=(8, 8))
    enhanced = clahe.apply(gray)
    save_debug_image(enhanced, "01_preprocess_clahe")
    # Minimal blur to preserve edges
    blurred = cv2.GaussianBlur(enhanced, (3, 3), 0)
    save_debug_image(blurred, "02_preprocess_blur")
    # Multi-scale thresholding
    block_size = max(9, min(25, int(img.shape[0] / 20) * 2 + 1))
    thresh = cv2.adaptiveThreshold(blurred, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
                                   cv2.THRESH_BINARY_INV, block_size, 7)
    # Morphological operations
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (3, 3))
    thresh = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=2)
    thresh = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, kernel, iterations=3)
    save_debug_image(thresh, "03_preprocess_morph")
    return thresh, enhanced

def correct_rotation(img):
    """Correct image rotation using edge detection."""
    try:
        gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
        edges = cv2.Canny(gray, 30, 100, apertureSize=3)
        lines = cv2.HoughLinesP(edges, 1, np.pi / 180, threshold=40, minLineLength=20, maxLineGap=10)
        if lines is not None:
            angles = [np.arctan2(line[0][3] - line[0][1], line[0][2] - line[0][0]) * 180 / np.pi for line in lines]
            angle = np.median(angles)
            if abs(angle) > 0.5:
                h, w = img.shape[:2]
                center = (w // 2, h // 2)
                M = cv2.getRotationMatrix2D(center, angle, 1.0)
                img = cv2.warpAffine(img, M, (w, h))
                save_debug_image(img, "00_rotated_image")
                logging.info(f"Applied rotation: {angle:.2f} degrees")
        return img
    except Exception as e:
        logging.error(f"Rotation correction failed: {str(e)}")
        return img

def detect_roi(img):
    """Detect region of interest with aggressive contour filtering."""
    try:
        save_debug_image(img, "04_original")
        thresh, enhanced = preprocess_image(img)
        brightness_map = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
        block_sizes = [max(9, min(25, int(img.shape[0] / s) * 2 + 1)) for s in [10, 15, 20]]
        valid_contours = []
        img_area = img.shape[0] * img.shape[1]
        
        for block_size in block_sizes:
            temp_thresh = cv2.adaptiveThreshold(enhanced, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
                                                cv2.THRESH_BINARY_INV, block_size, 7)
            kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))
            temp_thresh = cv2.morphologyEx(temp_thresh, cv2.MORPH_CLOSE, kernel, iterations=3)
            save_debug_image(temp_thresh, f"05_roi_threshold_block{block_size}")
            contours, _ = cv2.findContours(temp_thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
            
            for c in contours:
                area = cv2.contourArea(c)
                x, y, w, h = cv2.boundingRect(c)
                roi_brightness = np.mean(brightness_map[y:y+h, x:x+w])
                aspect_ratio = w / h
                if (500 < area < (img_area * 0.5) and 
                    0.5 <= aspect_ratio <= 6.0 and w > 80 and h > 40 and roi_brightness > 60):
                    valid_contours.append((c, area * roi_brightness))
                    logging.debug(f"Contour (block {block_size}): Area={area}, Aspect={aspect_ratio:.2f}, Brightness={roi_brightness:.2f}")
        
        if valid_contours:
            contour, _ = max(valid_contours, key=lambda x: x[1])
            x, y, w, h = cv2.boundingRect(contour)
            padding = max(25, min(70, int(min(w, h) * 0.5)))
            x, y = max(0, x - padding), max(0, y - padding)
            w, h = min(w + 2 * padding, img.shape[1] - x), min(h + 2 * padding, img.shape[0] - y)
            roi_img = img[y:y+h, x:x+w]
            save_debug_image(roi_img, "06_detected_roi")
            logging.info(f"Detected ROI: ({x}, {y}, {w}, {h})")
            return roi_img, (x, y, w, h)
        
        logging.info("No ROI found, using full image.")
        save_debug_image(img, "06_no_roi_fallback")
        return img, None
    except Exception as e:
        logging.error(f"ROI detection failed: {str(e)}")
        save_debug_image(img, "06_roi_error_fallback")
        return img, None

def detect_digit_contour(digit_img, brightness):
    """Simplified contour-based digit recognition."""
    try:
        h, w = digit_img.shape
        if h < 20 or w < 10:
            logging.debug("Digit image too small for contour detection.")
            return None

        # Normalize image
        pixel_count = np.sum(digit_img == 255)
        total_pixels = digit_img.size
        density = pixel_count / total_pixels
        if density < 0.1 or density > 0.8:
            return None

        # Contour analysis
        contours, _ = cv2.findContours(digit_img, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
        if not contours:
            return None

        contour = max(contours, key=cv2.contourArea)
        x, y, cw, ch = cv2.boundingRect(contour)
        if cw < 5 or ch < 10:
            return None

        aspect = cw / ch
        area_ratio = cv2.contourArea(contour) / (cw * ch)

        # Simplified digit patterns
        if aspect > 0.2 and aspect < 0.4 and area_ratio > 0.5:
            return '1'
        elif aspect > 0.5 and area_ratio > 0.6:
            if density > 0.5:
                return '8'
            elif density > 0.3:
                return '0'
        elif aspect > 0.4 and area_ratio > 0.5:
            if density > 0.4:
                return '3'
            elif density > 0.3:
                return '2'
        elif aspect > 0.3 and area_ratio > 0.4:
            return '5' if density > 0.3 else '7'
        elif aspect > 0.2 and area_ratio > 0.3:
            return '4' if density > 0.2 else '9'
        return None
    except Exception as e:
        logging.error(f"Contour digit detection failed: {str(e)}")
        return None

def perform_ocr(img, roi_bbox):
    """Perform OCR with Tesseract and contour-based fallback."""
    try:
        thresh, enhanced = preprocess_image(img)
        brightness = estimate_brightness(img)
        pil_img = Image.fromarray(enhanced)
        save_debug_image(pil_img, "07_ocr_input")
        
        # Tesseract with aggressive numeric config
        custom_config = r'--oem 3 --psm 7 -c tessedit_char_whitelist=0123456789.'
        text = pytesseract.image_to_string(pil_img, config=custom_config)
        logging.info(f"Tesseract raw output: {text}")
        
        # Clean and validate
        text = re.sub(r"[^\d\.]", "", text)
        if text.count('.') > 1:
            text = text.replace('.', '', text.count('.') - 1)
        text = text.strip('.')
        if text and re.fullmatch(r"^\d*\.?\d*$", text):
            text = text.lstrip('0') or '0'
            confidence = 98.0 if len(text.replace('.', '')) >= 3 else 95.0
            logging.info(f"Validated Tesseract text: {text}, Confidence: {confidence:.2f}%")
            return text, confidence

        # Fallback to contour-based detection
        logging.info("Tesseract failed, using contour-based detection.")
        contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
        digits_info = []
        for c in contours:
            x, y, w, h = cv2.boundingRect(c)
            if w > 15 and h > 20 and 0.2 <= w/h <= 1.2:
                digits_info.append((x, x+w, y, y+h))
        
        if digits_info:
            digits_info.sort(key=lambda x: x[0])
            recognized_text = ""
            prev_x_max = -float('inf')
            for idx, (x_min, x_max, y_min, y_max) in enumerate(digits_info):
                x_min, y_min = max(0, x_min), max(0, y_min)
                x_max, y_max = min(thresh.shape[1], x_max), min(thresh.shape[0], y_max)
                if x_max <= x_min or y_max <= y_min:
                    continue
                digit_crop = thresh[y_min:y_max, x_min:x_max]
                save_debug_image(digit_crop, f"08_digit_crop_{idx}")
                digit = detect_digit_contour(digit_crop, brightness)
                if digit:
                    recognized_text += digit
                elif x_min - prev_x_max < 15 and prev_x_max != -float('inf'):
                    recognized_text += '.'
                prev_x_max = x_max
            
            text = re.sub(r"[^\d\.]", "", recognized_text)
            if text.count('.') > 1:
                text = text.replace('.', '', text.count('.') - 1)
            text = text.strip('.')
            if text and re.fullmatch(r"^\d*\.?\d*$", text):
                text = text.lstrip('0') or '0'
                confidence = 92.0 if len(text.replace('.', '')) >= 3 else 90.0
                logging.info(f"Validated contour text: {text}, Confidence: {confidence:.2f}%")
                return text, confidence
        
        logging.info("No valid digits detected.")
        return None, 0.0
    except Exception as e:
        logging.error(f"OCR failed: {str(e)}")
        return None, 0.0

def extract_weight_from_image(pil_img):
    """Extract weight from a digital scale image."""
    try:
        img = np.array(pil_img)
        img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
        save_debug_image(img, "00_input_image")
        img = correct_rotation(img)
        brightness = estimate_brightness(img)
        conf_threshold = 0.9 if brightness > 100 else 0.7

        roi_img, roi_bbox = detect_roi(img)
        if roi_bbox:
            conf_threshold *= 1.15 if (roi_bbox[2] * roi_bbox[3]) > (img.shape[0] * img.shape[1] * 0.3) else 1.0

        result, confidence = perform_ocr(roi_img, roi_bbox)
        if result and confidence >= conf_threshold * 100:
            try:
                weight = float(result)
                if 0.01 <= weight <= 1000:
                    logging.info(f"Detected weight: {result} kg, Confidence: {confidence:.2f}%")
                    return result, confidence
                logging.warning(f"Weight {result} out of range.")
            except ValueError:
                logging.warning(f"Invalid weight format: {result}")

        logging.info("Primary OCR failed, using full image fallback.")
        result, confidence = perform_ocr(img, None)
        if result and confidence >= conf_threshold * 0.95 * 100:
            try:
                weight = float(result)
                if 0.01 <= weight <= 1000:
                    logging.info(f"Full image weight: {result} kg, Confidence: {confidence:.2f}%")
                    return result, confidence
                logging.warning(f"Full image weight {result} out of range.")
            except ValueError:
                logging.warning(f"Invalid full image weight format: {result}")

        logging.info("No valid weight detected.")
        return "Not detected", 0.0
    except Exception as e:
        logging.error(f"Weight extraction failed: {str(e)}")
        return "Not detected", 0.0