Update ocr_engine.py

ocr_engine.py

@@ -5,53 +5,54 @@ from PIL import Image
 import easyocr
 import os
 
-# Initialize OCR Reader
+# ✅ Initialize OCR reader only once
 reader = easyocr.Reader(['en'], gpu=False)
 
 def preprocess_image(image):
-    """Preprocess the image to improve OCR accuracy"""
-    # Convert to grayscale
+    """
+    Preprocess the image to improve OCR detection.
+    Converts to grayscale and applies adaptive threshold.
+    """
     gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
-
-    # Apply threshold to isolate digits
-    _, thresh = cv2.threshold(gray, 150, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)
-
+    # Apply adaptive threshold to isolate digits better
+    thresh = cv2.adaptiveThreshold(
+        gray, 255,
+        cv2.ADAPTIVE_THRESH_MEAN_C,
+        cv2.THRESH_BINARY_INV,
+        11, 10
+    )
     return thresh
 
 def extract_weight_from_image(pil_image):
     try:
-        # Convert PIL to OpenCV format
+        # ✅ Step 1: Convert to OpenCV format
         image = np.array(pil_image.convert("RGB"))
 
-        # Print image shape for debugging
-        print("Image shape:", image.shape)
-
-        # Preprocess for better OCR accuracy
+        # ✅ Step 2: Preprocess image
         processed = preprocess_image(image)
 
-        # Save debug image
-        debug_img = Image.fromarray(processed)
+        # ✅ Step 3: Optional - Save debug image
         debug_path = "debug_processed_image.png"
-        debug_img.save(debug_path)
-        print(f"✅ Processed image saved to: {debug_path}")
+        Image.fromarray(processed).save(debug_path)
+        print(f"[DEBUG] Saved preprocessed image to {debug_path}")
 
-        # Run OCR on processed image
+        # ✅ Step 4: Run EasyOCR
         result = reader.readtext(processed)
 
-        print("✅ OCR Results:")
+        print("🔍 OCR Results:")
         for r in result:
-            print(f"Text: '{r[1]}' | Confidence: {r[2] * 100:.2f}%")
+            print(f"  • Text: '{r[1]}' | Confidence: {r[2]*100:.2f}%")
 
-        # Try to find numeric weight
+        # ✅ Step 5: Look for a decimal number like 53.25
         weight = None
         confidence = 0.0
 
         for detection in result:
-            text = detection[1]
+            text = detection[1].replace(",", ".")  # Handle comma decimal (if any)
             conf = detection[2]
 
-            # Match numbers like 53.25 or 45
-            match = re.search(r"\b\d+(\.\d+)?\b", text)
+            # Look for numbers like 53.25 or 100
+            match = re.search(r"\b\d{1,3}(\.\d{1,2})?\b", text)
             if match:
                 weight = match.group()
                 confidence = conf
@@ -63,5 +64,5 @@ def extract_weight_from_image(pil_image):
             return "No weight detected", 0.0
 
     except Exception as e:
-        print("❌ Exception during OCR:", str(e))
+        print(f"❌ OCR Error: {e}")
         return f"Error: {str(e)}", 0.0