Update ocr_engine.py

ocr_engine.py

@@ -31,20 +31,23 @@ def estimate_brightness(img):
     gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
     return np.mean(gray)
 
-def preprocess_image(img, scale=1.0):
+def preprocess_image(img, scale=1.0, method='clahe'):
     """Preprocess image for better OCR accuracy."""
     if scale != 1.0:
         img = cv2.resize(img, None, fx=scale, fy=scale, interpolation=cv2.INTER_CUBIC)
         save_debug_image(img, f"01_preprocess_scaled_{scale}")
     gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
-    # Apply bilateral filter to preserve edges
-    denoised = cv2.bilateralFilter(gray, 9, 15, 15)
+    # Gentle denoising
+    denoised = cv2.bilateralFilter(gray, 7, 10, 10)
     save_debug_image(denoised, "02_preprocess_bilateral")
-    # Enhance contrast using CLAHE
-    clahe = cv2.createCLAHE(clipLimit=3.0, tileGridSize=(8, 8))
-    enhanced = clahe.apply(denoised)
-    save_debug_image(enhanced, "03_preprocess_clahe")
-    # Sharpen the image
+    # Enhance contrast
+    if method == 'clahe':
+        clahe = cv2.createCLAHE(clipLimit=3.5, tileGridSize=(8, 8))
+        enhanced = clahe.apply(denoised)
+    else:  # Histogram equalization
+        enhanced = cv2.equalizeHist(denoised)
+    save_debug_image(enhanced, f"03_preprocess_{method}")
+    # Sharpen
     kernel_sharpening = np.array([[-1, -1, -1], [-1, 9, -1], [-1, -1, -1]])
     sharpened = cv2.filter2D(enhanced, -1, kernel_sharpening)
     save_debug_image(sharpened, "04_preprocess_sharpened")
@@ -54,11 +57,11 @@ def correct_rotation(img):
     """Correct image rotation using Hough Transform."""
     try:
         edges = cv2.Canny(cv2.cvtColor(img, cv2.COLOR_BGR2GRAY), 50, 150)
-        lines = cv2.HoughLinesP(edges, 1, np.pi / 180, threshold=80, minLineLength=50, maxLineGap=10)
+        lines = cv2.HoughLinesP(edges, 1, np.pi / 180, threshold=50, minLineLength=40, 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) > 3:
+            if abs(angle) > 2:
                 (h, w) = img.shape[:2]
                 center = (w // 2, h // 2)
                 M = cv2.getRotationMatrix2D(center, angle, 1.0)
@@ -76,64 +79,66 @@ def detect_roi(img):
         save_debug_image(img, "05_original")
         brightness_map = cv2.GaussianBlur(cv2.cvtColor(img, cv2.COLOR_BGR2GRAY), (15, 15), 0)
         
-        # Try multiple scales for preprocessing
-        scales = [1.0, 1.5, 0.75]
+        # Try multiple scales and methods
+        scales = [1.0, 1.5, 0.5]
+        methods = ['clahe', 'hist']
         for scale in scales:
-            preprocessed = preprocess_image(img, scale)
-            block_size = max(11, min(31, int(img.shape[0] / 20) * 2 + 1))
-            thresh = cv2.adaptiveThreshold(preprocessed, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
-                                           cv2.THRESH_BINARY_INV, block_size, 5)
-            _, otsu_thresh = cv2.threshold(preprocessed, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)
-            combined_thresh = cv2.bitwise_and(thresh, otsu_thresh)
-            save_debug_image(combined_thresh, f"06_roi_combined_threshold_scale_{scale}")
-            
-            # Morphological operations
-            kernel = np.ones((5, 5), np.uint8)
-            dilated = cv2.dilate(combined_thresh, kernel, iterations=2)
-            eroded = cv2.erode(dilated, kernel, iterations=1)
-            save_debug_image(eroded, f"07_roi_morphological_scale_{scale}")
-            
-            contours, _ = cv2.findContours(eroded, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
-            
-            if contours:
-                img_area = img.shape[0] * img.shape[1]
-                valid_contours = []
-                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] if scale == 1.0 else cv2.resize(brightness_map, (img.shape[1], img.shape[0])))
-                    aspect_ratio = w / h
-                    if (200 < area < (img_area * 0.95) and 
-                        0.5 <= aspect_ratio <= 15.0 and w > 50 and h > 20 and roi_brightness > 60):
-                        valid_contours.append((c, roi_brightness))
-                        logging.debug(f"Contour: Scale={scale}, Area={area}, Aspect={aspect_ratio:.2f}, Brightness={roi_brightness:.2f}")
+            for method in methods:
+                preprocessed = preprocess_image(img, scale, method)
+                block_size = max(9, min(31, int(img.shape[0] / 25) * 2 + 1))
+                thresh = cv2.adaptiveThreshold(preprocessed, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
+                                               cv2.THRESH_BINARY_INV, block_size, 3)
+                _, otsu_thresh = cv2.threshold(preprocessed, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)
+                combined_thresh = cv2.bitwise_and(thresh, otsu_thresh)
+                save_debug_image(combined_thresh, f"06_roi_combined_threshold_scale_{scale}_{method}")
                 
-                if valid_contours:
-                    contour, _ = max(valid_contours, key=lambda x: x[1])
-                    x, y, w, h = cv2.boundingRect(contour)
-                    if scale != 1.0:
-                        x, y, w, h = [int(v / scale) for v in (x, y, w, h)]
-                    padding = 120
-                    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, f"08_detected_roi_scale_{scale}")
-                    logging.info(f"Detected ROI with dimensions: ({x}, {y}, {w}, {h}) at scale {scale}")
-                    return roi_img, (x, y, w, h)
+                # Morphological operations
+                kernel = np.ones((3, 3), np.uint8)
+                dilated = cv2.dilate(combined_thresh, kernel, iterations=2)
+                eroded = cv2.erode(dilated, kernel, iterations=1)
+                save_debug_image(eroded, f"07_roi_morphological_scale_{scale}_{method}")
+                
+                contours, _ = cv2.findContours(eroded, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
+                
+                if contours:
+                    img_area = img.shape[0] * img.shape[1]
+                    valid_contours = []
+                    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] if scale == 1.0 else cv2.resize(brightness_map, (img.shape[1], img.shape[0])))
+                        aspect_ratio = w / h
+                        if (100 < area < (img_area * 0.95) and 
+                            0.3 <= aspect_ratio <= 20.0 and w > 40 and h > 15 and roi_brightness > 50):
+                            valid_contours.append((c, roi_brightness))
+                            logging.debug(f"Contour: Scale={scale}, Method={method}, 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)
+                        if scale != 1.0:
+                            x, y, w, h = [int(v / scale) for v in (x, y, w, h)]
+                        padding = 150
+                        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, f"08_detected_roi_scale_{scale}_{method}")
+                        logging.info(f"Detected ROI with dimensions: ({x}, {y}, {w}, {h}) at scale {scale}, method {method}")
+                        return roi_img, (x, y, w, h)
         
         logging.info("No suitable ROI found, attempting fallback criteria.")
         # Fallback with relaxed criteria
-        preprocessed = preprocess_image(img)
+        preprocessed = preprocess_image(img, method='clahe')
         thresh = cv2.adaptiveThreshold(preprocessed, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
-                                       cv2.THRESH_BINARY_INV, block_size, 8)
+                                       cv2.THRESH_BINARY_INV, block_size, 5)
         save_debug_image(thresh, "06_roi_fallback_threshold")
         contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
-        valid_contours = [c for c in contours if 100 < cv2.contourArea(c) < (img.shape[0] * img.shape[1] * 0.95) and 
-                          0.3 <= cv2.boundingRect(c)[2]/cv2.boundingRect(c)[3] <= 20.0]
+        valid_contours = [c for c in contours if 50 < cv2.contourArea(c) < (img.shape[0] * img.shape[1] * 0.95) and 
+                          0.2 <= cv2.boundingRect(c)[2]/cv2.boundingRect(c)[3] <= 25.0]
         if valid_contours:
             contour = max(valid_contours, key=cv2.contourArea)
             x, y, w, h = cv2.boundingRect(contour)
-            padding = 120
+            padding = 150
             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]
@@ -152,13 +157,13 @@ def detect_roi(img):
 def detect_segments(digit_img, brightness):
     """Detect seven-segment patterns in a digit image."""
     h, w = digit_img.shape
-    if h < 10 or w < 8:
+    if h < 8 or w < 6:
         return None
 
     segments = {
         'top': (int(w*0.1), int(w*0.9), 0, int(h*0.25)),
         'middle': (int(w*0.1), int(w*0.9), int(h*0.45), int(h*0.55)),
-        'bottom': (int(w*0.1), int看到的: int(w*0.9), int(h*0.75), h),
+        'bottom': (int(w*0.1), int(w*0.9), int(h*0.75), h),
         'left_top': (0, int(w*0.3), int(h*0.1), int(h*0.5)),
         'left_bottom': (0, int(w*0.3), int(h*0.5), int(h*0.9)),
         'right_top': (int(w*0.7), w, int(h*0.1), int(h*0.5)),
@@ -175,7 +180,7 @@ def detect_segments(digit_img, brightness):
             continue
         pixel_count = np.sum(region == 255)
         total_pixels = region.size
-        segment_presence[name] = pixel_count / total_pixels > (0.2 if brightness < 80 else 0.4)
+        segment_presence[name] = pixel_count / total_pixels > (0.15 if brightness < 80 else 0.35)
 
     digit_patterns = {
         '0': ('top', 'bottom', 'left_top', 'left_bottom', 'right_top', 'right_bottom'),
@@ -195,8 +200,8 @@ def detect_segments(digit_img, brightness):
     for digit, pattern in digit_patterns.items():
         matches = sum(1 for segment in pattern if segment_presence.get(segment, False))
         non_matches_penalty = sum(1 for segment in segment_presence if segment not in pattern and segment_presence[segment])
-        score = matches - 0.2 * non_matches_penalty
-        if matches >= len(pattern) * 0.7:
+        score = matches - 0.15 * non_matches_penalty
+        if matches >= len(pattern) * 0.65:
             score += 1.0
         if score > max_score:
             max_score = score
@@ -208,9 +213,9 @@ def detect_segments(digit_img, brightness):
 def custom_seven_segment_ocr(img, roi_bbox):
     """Perform custom OCR for seven-segment displays."""
     try:
-        preprocessed = preprocess_image(img)
+        preprocessed = preprocess_image(img, method='clahe')
         brightness = estimate_brightness(img)
-        thresh_value = 80 if brightness < 80 else 0
+        thresh_value = 60 if brightness < 80 else 0
         _, thresh = cv2.threshold(preprocessed, thresh_value, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)
         save_debug_image(thresh, "09_roi_thresh_for_digits")
 
@@ -221,20 +226,20 @@ def custom_seven_segment_ocr(img, roi_bbox):
 
         batch_size = max(4, min(16, int(img.shape[0] * img.shape[1] / 100000)))
         results = easyocr_reader.readtext(thresh, detail=1, paragraph=False, 
-                                          contrast_ths=0.2, adjust_contrast=1.2, 
-                                          text_threshold=0.5, mag_ratio=4.0, 
-                                          allowlist='0123456789.', batch_size=batch_size, y_ths=0.3)
+                                          contrast_ths=0.1, adjust_contrast=1.3, 
+                                          text_threshold=0.3, mag_ratio=6.0, 
+                                          allowlist='0123456789.', batch_size=batch_size, y_ths=0.4)
         
-        logging.info(f"EasyOCR results: {results}")
+        logging.info(f"EasyOCR results (seven-segment): {results}")
         if not results:
-            logging.info("EasyOCR found no digits.")
+            logging.info("EasyOCR found no digits in seven-segment OCR.")
             return None
 
         digits_info = []
         for (bbox, text, conf) in results:
             (x1, y1), (x2, y2), (x3, y3), (x4, y4) = bbox
             h_bbox = max(y1, y2, y3, y4) - min(y1, y2, y3, y4)
-            if (text.isdigit() or text == '.') and h_bbox > 6:
+            if (text.isdigit() or text == '.') and h_bbox > 5:
                 x_min, x_max = int(min(x1, x4)), int(max(x2, x3))
                 y_min, y_max = int(min(y1, y2)), int(max(y3, y4))
                 digits_info.append((x_min, x_max, y_min, y_max, text, conf))
@@ -248,7 +253,7 @@ def custom_seven_segment_ocr(img, roi_bbox):
                 continue
             digit_img_crop = thresh[y_min:y_max, x_min:x_max]
             save_debug_image(digit_img_crop, f"11_digit_crop_{idx}_{easyocr_char}")
-            if easyocr_conf > 0.9 or easyocr_char == '.':
+            if easyocr_conf > 0.85 or easyocr_char == '.':
                 recognized_text += easyocr_char
             else:
                 digit_from_segments = detect_segments(digit_img_crop, brightness)
@@ -280,7 +285,7 @@ def extract_weight_from_image(pil_img):
         img = correct_rotation(img)
 
         brightness = estimate_brightness(img)
-        conf_threshold = 0.7 if brightness > 150 else (0.5 if brightness > 80 else 0.3)
+        conf_threshold = 0.65 if brightness > 150 else (0.45 if brightness > 80 else 0.25)
 
         roi_img, roi_bbox = detect_roi(img)
         if roi_bbox:
@@ -288,10 +293,10 @@ def extract_weight_from_image(pil_img):
             conf_threshold *= 1.1 if roi_area > (img.shape[0] * img.shape[1] * 0.5) else 1.0
 
         custom_result = custom_seven_segment_ocr(roi_img, roi_bbox)
-        if custom_result:
+        if custom_result and custom_result != '0':
             try:
                 weight = float(custom_result)
-                if 0.001 <= weight <= 2000:
+                if 0.0001 <= weight <= 5000:
                     logging.info(f"Custom OCR result: {custom_result}, Confidence: 95.0%")
                     return custom_result, 95.0
                 else:
@@ -300,78 +305,122 @@ def extract_weight_from_image(pil_img):
                 logging.warning(f"Custom OCR result '{custom_result}' is not a valid number.")
 
         logging.info("Custom OCR failed or invalid, falling back to enhanced EasyOCR.")
-        preprocessed_roi = preprocess_image(roi_img)
-        block_size = max(11, min(31, int(roi_img.shape[0] / 20) * 2 + 1))
+        preprocessed_roi = preprocess_image(roi_img, method='hist')
+        block_size = max(9, min(31, int(roi_img.shape[0] / 25) * 2 + 1))
         final_roi = cv2.adaptiveThreshold(preprocessed_roi, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
-                                          cv2.THRESH_BINARY_INV, block_size, 8)
+                                          cv2.THRESH_BINARY_INV, block_size, 5)
         save_debug_image(final_roi, "12_fallback_adaptive_thresh")
 
         batch_size = max(4, min(16, int(roi_img.shape[0] * roi_img.shape[1] / 100000)))
-        results = easyocr_reader.readtext(final_roi, detail=1, paragraph=False, 
-                                          contrast_ths=0.3, adjust_contrast=1.2, 
-                                          text_threshold=0.4, mag_ratio=5.0, 
-                                          allowlist='0123456789. kglb', batch_size=batch_size, y_ths=0.3)
-
-        # Secondary EasyOCR pass with different parameters
-        if not results:
-            logging.info("First EasyOCR pass failed, trying with relaxed parameters.")
+        ocr_passes = [
+            {'contrast_ths': 0.2, 'text_threshold': 0.3, 'mag_ratio': 6.0, 'y_ths': 0.4, 'label': 'first'},
+            {'contrast_ths': 0.1, 'text_threshold': 0.2, 'mag_ratio': 7.0, 'y_ths': 0.5, 'label': 'second'},
+            {'contrast_ths': 0.05, 'text_threshold': 0.1, 'mag_ratio': 8.0, 'y_ths': 0.6, 'label': 'third'}
+        ]
+        candidates = []
+
+        for ocr_pass in ocr_passes:
             results = easyocr_reader.readtext(final_roi, detail=1, paragraph=False, 
-                                              contrast_ths=0.2, adjust_contrast=1.5, 
-                                              text_threshold=0.3, mag_ratio=6.0, 
-                                              allowlist='0123456789. kglb', batch_size=batch_size, y_ths=0.4)
-            save_debug_image(final_roi, "12_fallback_adaptive_thresh_relaxed")
-
-        logging.info(f"EasyOCR results: {results}")
-        best_weight = None
-        best_conf = 0.0
-        best_score = 0.0
-        unit = None
-        for (bbox, text, conf) in results:
-            if 'kg' in text.lower():
-                unit = 'kg'
-                continue
-            elif 'g' in text.lower():
-                unit = 'g'
-                continue
-            elif 'lb' in text.lower():
-                unit = 'lb'
-                continue
-            text = re.sub(r"[^\d\.]", "", text)
-            if text.count('.') > 1:
-                text = text.replace('.', '', text.count('.') - 1)
-            text = text.strip('.')
-            if re.fullmatch(r"^\d*\.?\d*$", text):
-                try:
-                    weight = float(text)
-                    if unit == 'g':
-                        weight /= 1000  # Convert grams to kilograms
-                    elif unit == 'lb':
-                        weight *= 0.453592  # Convert pounds to kilograms
-                    range_score = 1.5 if 0.001 <= weight <= 2000 else 0.7
-                    digit_count = len(text.replace('.', ''))
-                    digit_score = 1.3 if 1 <= digit_count <= 8 else 0.8
-                    score = conf * range_score * digit_score
-                    if roi_bbox:
-                        (x_roi, y_roi, w_roi, h_roi) = roi_bbox
-                        roi_area = w_roi * h_roi
-                        x_min, y_min = int(min(b[0] for b in bbox)), int(min(b[1] for b in bbox))
-                        x_max, y_max = int(max(b[0] for b in bbox)), int(max(b[1] for b in bbox))
-                        bbox_area = (x_max - x_min) * (y_max - y_min)
-                        if roi_area > 0 and bbox_area / roi_area < 0.03:
-                            score *= 0.5
-                    if score > best_score and conf > conf_threshold:
-                        best_weight = text
-                        best_conf = conf
-                        best_score = score
+                                              contrast_ths=ocr_pass['contrast_ths'], 
+                                              adjust_contrast=1.4, 
+                                              text_threshold=ocr_pass['text_threshold'], 
+                                              mag_ratio=ocr_pass['mag_ratio'], 
+                                              allowlist='0123456789. kglb', 
+                                              batch_size=batch_size, 
+                                              y_ths=ocr_pass['y_ths'])
+            logging.info(f"EasyOCR results ({ocr_pass['label']} pass): {results}")
+            save_debug_image(final_roi, f"12_fallback_adaptive_thresh_{ocr_pass['label']}_pass")
+            
+            unit = None
+            for (bbox, text, conf) in results:
+                if 'kg' in text.lower():
+                    unit = 'kg'
+                    continue
+                elif 'g' in text.lower():
+                    unit = 'g'
+                    continue
+                elif 'lb' in text.lower():
+                    unit = 'lb'
+                    continue
+                text = re.sub(r"[^\d\.]", "", text)
+                if text.count('.') > 1:
+                    text = text.replace('.', '', text.count('.') - 1)
+                text = text.strip('.')
+                if re.fullmatch(r"^\d*\.?\d*$", text):
+                    try:
+                        weight = float(text)
+                        if unit == 'g':
+                            weight /= 1000
+                        elif unit == 'lb':
+                            weight *= 0.453592
+                        range_score = 1.5 if 0.0001 <= weight <= 5000 else 0.6
+                        digit_count = len(text.replace('.', ''))
+                        digit_score = 1.4 if 1 <= digit_count <= 8 else 0.7
+                        score = conf * range_score * digit_score
+                        if roi_bbox:
+                            (x_roi, y_roi, w_roi, h_roi) = roi_bbox
+                            roi_area = w_roi * h_roi
+                            x_min, y_min = int(min(b[0] for b in bbox)), int(min(b[1] for b in bbox))
+                            x_max, y_max = int(max(b[0] for b in bbox)), int(max(b[1] for b in bbox))
+                            bbox_area = (x_max - x_min) * (y_max - y_min)
+                            if roi_area > 0 and bbox_area / roi_area < 0.02:
+                                score *= 0.4
+                        candidates.append((text, conf, score, unit))
                         logging.info(f"Candidate EasyOCR weight: '{text}', Unit: {unit or 'none'}, Conf: {conf}, Score: {score}")
-                except ValueError:
-                    logging.warning(f"Could not convert '{text}' to float during EasyOCR fallback.")
+                    except ValueError:
+                        logging.warning(f"Could not convert '{text}' to float during EasyOCR fallback.")
+
+        # Fallback to full image if no candidates
+        if not candidates:
+            logging.info("No candidates from ROI, trying full image.")
+            preprocessed_full = preprocess_image(img, method='hist')
+            final_full = cv2.adaptiveThreshold(preprocessed_full, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
+                                               cv2.THRESH_BINARY_INV, block_size, 5)
+            save_debug_image(final_full, "12_fallback_full_image")
+            results = easyocr_reader.readtext(final_full, detail=1, paragraph=False, 
+                                              contrast_ths=0.1, adjust_contrast=1.5, 
+                                              text_threshold=0.2, mag_ratio=7.0, 
+                                              allowlist='0123456789. kglb', batch_size=batch_size, y_ths=0.5)
+            logging.info(f"EasyOCR results (full image): {results}")
+            
+            unit = None
+            for (bbox, text, conf) in results:
+                if 'kg' in text.lower():
+                    unit = 'kg'
                     continue
-
-        if not best_weight:
+                elif 'g' in text.lower():
+                    unit = 'g'
+                    continue
+                elif 'lb' in text.lower():
+                    unit = 'lb'
+                    continue
+                text = re.sub(r"[^\d\.]", "", text)
+                if text.count('.') > 1:
+                    text = text.replace('.', '', text.count('.') - 1)
+                text = text.strip('.')
+                if re.fullmatch(r"^\d*\.?\d*$", text):
+                    try:
+                        weight = float(text)
+                        if unit == 'g':
+                            weight /= 1000
+                        elif unit == 'lb':
+                            weight *= 0.453592
+                        range_score = 1.2 if 0.0001 <= weight <= 5000 else 0.5
+                        digit_count = len(text.replace('.', ''))
+                        digit_score = 1.2 if 1 <= digit_count <= 8 else 0.6
+                        score = conf * range_score * digit_score * 0.8  # Penalty for full image
+                        candidates.append((text, conf, score, unit))
+                        logging.info(f"Candidate EasyOCR weight (full image): '{text}', Unit: {unit or 'none'}, Conf: {conf}, Score: {score}")
+                    except ValueError:
+                        logging.warning(f"Could not convert '{text}' to float during full image fallback.")
+
+        if not candidates:
             logging.info("No valid weight detected after all attempts.")
             return "Not detected", 0.0
 
+        # Select best candidate
+        best_weight, best_conf, best_score, best_unit = max(candidates, key=lambda x: x[2])
+        
         # Format the weight
         if "." in best_weight:
             int_part, dec_part = best_weight.split(".")
@@ -383,14 +432,14 @@ def extract_weight_from_image(pil_img):
 
         try:
             final_weight = float(best_weight)
-            if final_weight < 0.001 or final_weight > 2000:
-                best_conf *= 0.6
-            elif final_weight == 0 and best_conf < 0.9:
-                best_conf *= 0.7  # Penalize zero weights with low confidence
+            if final_weight < 0.0001 or final_weight > 5000:
+                best_conf *= 0.5
+            elif final_weight == 0 and best_conf < 0.95:
+                best_conf *= 0.6  # Penalize zero weights
         except ValueError:
             pass
 
-        logging.info(f"Final detected weight: {best_weight} kg, Confidence: {round(best_conf * 100, 2)}%")
+        logging.info(f"Final detected weight: {best_weight} kg, Confidence: {round(best_conf * 100, 2)}%, Unit: {best_unit or 'none'}")
         return best_weight, round(best_conf * 100, 2)
 
     except Exception as e: