AI-Powered Smart Door Guardian with Face Recognition & IoT A

A smart security system that uses computer vision + IoT to detect, recognize, and respond to visitors at your door. It combines a Raspberry

IntermediateShowcase (no instructions)6 hours156

AI-Powered Smart Door Guardian with Face Recognition & IoT A

Things used in this project

Hardware components

Raspberry Pi 4 Model B

Raspberry Pi Camera Module V2

Story

Overview

A smart security system that uses computer vision + IoT to detect, recognize, and respond to visitors at your door. It combines a Raspberry Pi 4 Model B, camera module, and cloud alerts to create a real-time intelligent door guardian.

It can:

Recognize known faces
Detect unknown visitors
Send alerts to your phone
Trigger a buzzer / lock system

Key Features

🎥 Real-time face detection & recognition
📲 Instant Telegram alerts with image
🔐 Optional door lock control
📡 Remote monitoring via web dashboard
🧠 Local AI (no heavy cloud dependency)

Software Stack

Python 3
OpenCV
face_recognition library
Telegram Bot API
Flask (for dashboard)

System Architecture

Camera → Face Detection → Face Recognition
                ↓
         Known / Unknown
          ↓           ↓
     Unlock Door    Send Alert
                        ↓
                Telegram Notification

Circuit Connections

Installation Steps

1. Setup Raspberry Pi

sudo apt update && sudo apt upgrade
sudo apt install python3-pip

2. Install dependencies

pip3 install opencv-python face_recognition flask requests
pip3 install opencv-python face_recognition flask requests

📸 Dataset Preparation

Create a folder:

dataset/
 ├── arpit/
 │    ├── img1.jpg
 │    ├── img2.jpg

Testing Procedure

Run script
Show face to camera

Verify:

Known → Relay triggers
Unknown → Alert sent

Results

Face recognition accuracy: ~90–95% (depends on lighting)
Alert latency: ~2–3 seconds
Works fully offline except notifications

Future Improvements

Add voice assistant (Amazon Alexa integration)
Use edge AI accelerator (Google Coral USB Accelerator)
Add cloud logging dashboard
Mobile app integration

🧠 Main Python Code

import cv2
import face_recognition
import os
import requests
import RPi.GPIO as GPIO

# GPIO Setup
GPIO.setmode(GPIO.BCM)
BUZZER = 18
RELAY = 23
GPIO.setup(BUZZER, GPIO.OUT)
GPIO.setup(RELAY, GPIO.OUT)

# Telegram config
TOKEN = "YOUR_BOT_TOKEN"
CHAT_ID = "YOUR_CHAT_ID"

def send_alert(image_path):
    url = f"https://api.telegram.org/bot{TOKEN}/sendPhoto"
    with open(image_path, 'rb') as img:
        requests.post(url, data={"chat_id": CHAT_ID}, files={"photo": img})

# Load known faces
known_encodings = []
known_names = []

dataset_path = "dataset"

for person in os.listdir(dataset_path):
    for img_name in os.listdir(f"{dataset_path}/{person}"):
        img = face_recognition.load_image_file(f"{dataset_path}/{person}/{img_name}")
        enc = face_recognition.face_encodings(img)[0]
        known_encodings.append(enc)
        known_names.append(person)

# Start camera
video = cv2.VideoCapture(0)

while True:
    ret, frame = video.read()
    rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)

    faces = face_recognition.face_locations(rgb)
    encodings = face_recognition.face_encodings(rgb, faces)

    for (top, right, bottom, left), face_encoding in zip(faces, encodings):
        matches = face_recognition.compare_faces(known_encodings, face_encoding)
        name = "Unknown"

        if True in matches:
            name = known_names[matches.index(True)]
            GPIO.output(RELAY, GPIO.HIGH)
        else:
            GPIO.output(BUZZER, GPIO.HIGH)
            cv2.imwrite("alert.jpg", frame)
            send_alert("alert.jpg")

        cv2.rectangle(frame, (left, top), (right, bottom), (0,255,0), 2)
        cv2.putText(frame, name, (left, top-10),
                    cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0,255,0), 2)

    cv2.imshow("Smart Guardian", frame)

    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

video.release()
cv2.destroyAllWindows()
GPIO.cleanup()