Smart Doorbell with Face Recognition (ESP32-CAM): Complete FYP Build Guide

Why a Smart Doorbell Makes a Strong FYP

A face-recognition doorbell hits the sweet spot examiners look for: it's visual (the demo is genuinely impressive), it touches computer vision, embedded systems, and networking all in one build, and it solves a problem people actually recognise — knowing who's at the door before you open it.

Unlike a lot of sensor-and-buzzer projects, this one produces a photo and a name on your phone within seconds. That's the kind of live demo that keeps a viva panel's attention.

Hardware Components

Component	Purpose	Cost (RM)
ESP32-CAM (AI-Thinker)	Camera + microcontroller	RM 25–35
FTDI Programmer	Flash the ESP32-CAM	RM 8–12
Push Button	Doorbell trigger	RM 1
Buzzer	Audio confirmation	RM 2
PIR Motion Sensor (optional)	Wake camera on approach	RM 4–6
5V Power Supply	Continuous power	RM 10–15
Enclosure	Weatherproof housing	RM 10–20

Total hardware cost: RM 60–90

How It Works

1. Someone presses the button (or the PIR sensor detects motion)
2. ESP32-CAM captures a photo
3. The photo is compared against a small set of registered faces stored on the device or sent to a lightweight recognition API
4. If the face matches a known person, the doorbell logs it as "recognised"; if not, it's logged as "unknown visitor"
5. Either way, a Telegram bot sends the photo and result straight to your phone
6. The buzzer gives the visitor audio confirmation that the doorbell registered them

For FYP scope, most students use one of two approaches:

Option A — On-device with esp32-cam + Edge Impulse: faster response, works without internet, but recognition accuracy is lower with only a handful of training images.

Option B — Cloud-assisted with a Python/Flask backend: ESP32-CAM sends the photo to a small Flask server running OpenCV + face_recognition, which does the actual matching and pushes the Telegram alert. Slightly more setup, noticeably better accuracy — this is what we recommend for degree-level FYPs.

Wiring the ESP32-CAM

• 5V → VCC
• GND → GND
• Push button → GPIO 12 (with pull-down resistor)
• Buzzer → GPIO 13
• PIR sensor (optional) → GPIO 14

The ESP32-CAM doesn't have a built-in USB port, so you'll need the FTDI programmer for flashing: connect FTDI TX → ESP32 RX, FTDI RX → ESP32 TX, and bridge GPIO 0 to GND only while uploading code.

Core Logic (Pseudocode)

void loop() {
  if (buttonPressed() || motionDetected()) {
    camera_fb_t *photo = capturePhoto();
    sendToServer(photo);          // Flask backend does face matching
    String result = waitForResult();

    if (result == "recognised") {
      beep(1);
      sendTelegramAlert(photo, "Recognised visitor");
    } else {
      beep(3);
      sendTelegramAlert(photo, "Unknown visitor");
    }
  }
}

On the server side, a simple Flask endpoint using the face_recognition library can compare the incoming image against a folder of known faces and return a match name or "unknown" in under a second on a laptop-grade CPU.

Presentation Tips for the Viva

• Bring 2-3 registered faces and one "stranger" (a coursemate) for a live demo — panels want to see both the positive and negative case
• Show the Telegram log with timestamps as evidence of data persistence
• Be ready to explain the tradeoff between on-device and cloud-based recognition — this is a common follow-up question
• If you have time, add a simple web dashboard showing visitor history; it turns a hardware demo into a full-stack one

Scope Recommendations

For a Diploma FYP: Camera + button + Telegram alert with photo is enough For a Degree FYP: Add face matching (recognised vs unknown) + visitor log For Merit/Distinction: Add a web dashboard, multiple registered users, and a report of visit frequency

Need This Project Done?

Rectronx Circuits builds complete face-recognition doorbell systems — camera setup, recognition backend, Telegram integration, and full documentation. WhatsApp us your university's requirements and we'll send a free quote within 2 hours.

Looking for more inspiration? Browse 500+ FYP project titles by category and get a free quote.