Update ocr_processing.py

ocr_processing.py

@@ -32,16 +32,36 @@ def correct_lighting(image):
 
 # OCR Extraction using EasyOCR
 def extract_text(image, langs=['en']):
-    reader = easyocr.Reader(langs, gpu=False)  # GPU=True لو عندك دعم CUDA
-    results = reader.readtext(image, detail=0)
+    reader = easyocr.Reader(langs, gpu=False)
+    results = reader.readtext(image, detail=0, paragraph=True)
     text = "\n".join(results)
     return text
 
+    
+def enhance_for_ocr(image):
+    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+    # Adaptive Threshold
+    th = cv2.adaptiveThreshold(
+        gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
+        cv2.THRESH_BINARY, 15, 10
+    )
+    return th
+
+
 # Full pipeline
 def process_image(file, langs=['en']):
     img = Image.open(file).convert('RGB')
     img_cv = np.array(img)
+
+    # 1. تصحيح الانحراف
     img_cv = deskew(img_cv)
+
+    # 2. تصحيح الإضاءة
     img_cv = correct_lighting(img_cv)
+
+    # 3. تحسين للنصوص
+    img_cv = enhance_for_ocr(img_cv)
+
+    # 4. استخراج النصوص
     text = extract_text(img_cv, langs)
     return text