Update ocr_engine.py

ocr_engine.py

@@ -5,30 +5,36 @@ from PIL import Image
 
 def extract_weight_from_image(pil_img):
     try:
-        # Convert PIL to OpenCV
+        # Convert PIL image to OpenCV format
         img = pil_img.convert("RGB")
-        img = np.array(img)
-        img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
+        img_np = np.array(img)
+        img_cv = cv2.cvtColor(img_np, cv2.COLOR_RGB2BGR)
 
         # Convert to grayscale
-        gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+        gray = cv2.cvtColor(img_cv, cv2.COLOR_BGR2GRAY)
 
-        # Thresholding to highlight digits
-        _, binary = cv2.threshold(gray, 120, 255, cv2.THRESH_BINARY_INV)
+        # Adaptive Thresholding for 7-segment LCD
+        processed = cv2.adaptiveThreshold(
+            gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY_INV, 15, 10
+        )
 
-        # Resize for better OCR
-        resized = cv2.resize(binary, None, fx=2, fy=2, interpolation=cv2.INTER_CUBIC)
+        # Resize to enhance small text
+        resized = cv2.resize(processed, None, fx=2, fy=2, interpolation=cv2.INTER_LINEAR)
 
-        # Run OCR with digit whitelist
-        config = r'--oem 3 --psm 6 -c tessedit_char_whitelist=0123456789.'
-        raw_text = pytesseract.image_to_string(resized, config=config)
+        # OCR config tuned for digit blocks
+        config = r'--oem 3 --psm 7 -c tessedit_char_whitelist=0123456789.'
 
-        print("🔍 OCR Raw Output:", repr(raw_text))  # Show in Hugging Face logs
+        # Run OCR
+        text = pytesseract.image_to_string(resized, config=config)
+
+        print("🔍 RAW OCR OUTPUT:", repr(text))
+
+        # Clean the text
+        weight = ''.join(c for c in text if c in '0123456789.')
+        weight = weight.strip()
 
-        # Filter for digits only
-        weight = ''.join(filter(lambda c: c in '0123456789.', raw_text))
         confidence = 95 if weight else 0
-        return weight.strip(), confidence
+        return weight, confidence
 
     except Exception as e:
         print("❌ OCR Error:", str(e))