Update ocr_engine.py

ocr_engine.py

@@ -9,29 +9,31 @@ def extract_weight_from_image(pil_img):
     try:
         img = np.array(pil_img)
 
-        # Preprocessing for OCR
+        # Resize if image is too large
+        if img.shape[1] > 1000:
+            img = cv2.resize(img, (1000, int(img.shape[0] * 1000 / img.shape[1])))
+
+        # Preprocessing
         gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
-        gray = cv2.resize(gray, None, fx=2, fy=2, interpolation=cv2.INTER_LINEAR)
+        gray = cv2.resize(gray, None, fx=2, fy=2, interpolation=cv2.INTER_CUBIC)
         gray = cv2.equalizeHist(gray)
-        gray = cv2.GaussianBlur(gray, (3, 3), 0)
-        inverted = cv2.bitwise_not(gray)
+        blurred = cv2.GaussianBlur(gray, (3, 3), 0)
+        thresh = cv2.adaptiveThreshold(
+            blurred, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, 11, 2
+        )
 
-        # OCR read
-        result = reader.readtext(inverted, detail=0)
+        # OCR
+        result = reader.readtext(thresh, detail=0)
         combined_text = " ".join(result)
-        print("OCR Raw Text:", combined_text)
+        print("OCR Text:", combined_text)
 
-        # First try matching with "kg"
-        match = re.search(r"(\d{1,4}(?:\.\d{1,2})?)\s*(kg|KG|Kg)", combined_text)
+        # Regex to match numbers with optional 'kg'
+        match = re.search(r"(\d{1,4}(?:\.\d{1,2})?)\s*(kg|KG|Kg)?", combined_text)
         if match:
-            return f"{match.group(1)} kg", 95.0
-
-        # Then try fallback match: only numbers
-        fallback = re.search(r"\d{1,4}(?:\.\d{1,2})?", combined_text)
-        if fallback:
-            return f"{fallback.group(0)} kg", 75.0
-
-        return "No weight detected kg", 0.0
+            weight = match.group(1)
+            return f"{weight} kg", 95.0
+        else:
+            return "No weight detected kg", 0.0
 
     except Exception as e:
         return f"Error: {str(e)}", 0.0