Update VisionMate: Added OCR utility, updated README, main.py, and requirements

main.py

@@ -13,10 +13,11 @@
 from langchain_core.prompts import PromptTemplate
 from langchain_core.messages import HumanMessage
 
-
 from tts_utils import speak
 from config import TTS_ENGINE, IP_WEBCAM_URL
 
+from ocr_utils import extract_text_from_image
+
 # Load Gemini API key from .env
 load_dotenv()
 api_key = os.getenv("API_KEY")
@@ -145,10 +146,18 @@ def listen_for_scan():
         scan_triggered = False
         status = "Analyzing surroundings..."
         speak("Analyzing surroundings")
+
+        # ✅ Gemini AI description
         desc = process_frame(frame)
         speak(desc)
 
+         # ✅ OCR text detection
+        text = extract_text_from_image(frame)
+        if text.strip():
+            speak(f"Detected text: {text}")
+
         # Alert on specific signs
+        lower_desc = desc.lower()
         if "stop sign" in desc.lower() or "stop" in desc.lower():
             speak("Stop! There's a stop sign.")
         elif "red light" in desc.lower():

ocr_utils.py

@@ -0,0 +1,21 @@
+import pytesseract
+import cv2
+import numpy as np
+
+def extract_text_from_image(image: 'np.ndarray') -> str:
+    """
+    Extract text from a given image using pytesseract OCR.
+
+    Args:
+        image: Input image as a numpy array (OpenCV format)
+
+    Returns:
+        Extracted text as a string
+    """
+    # Convert image to grayscale (OCR works better on gray)
+    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+
+    # Use pytesseract to extract text
+    text = pytesseract.image_to_string(gray)
+
+    return text

requirements.txt

@@ -1,3 +1,4 @@
+# Existing dependencies
 opencv-python
 pyttsx3
 gTTS
@@ -10,4 +11,7 @@ playsound
 python-dotenv
 langchain-google-genai
 langchain-core
-insightface
+
+# Add InsightFace
+insightface
+pytesseract