An IoT-based smart door locking system that uses the ESP32-CAM to stream live video, capture images, and simulate a facial recognition-based unlocking mechanism.
⚠️ This is a half-functional prototype. Due to a hardware mismatch (wrong ESP32 module), face recognition could not be implemented on-device. However, the streaming, capture, SD card logging, and lock mechanism were successfully tested.
- Captures live video from the ESP32-CAM module.
- Takes still images on command.
- Planned: Saves images to an SD card.
- Planned: Controls a relay and servo motor to simulate door unlocking.
- Planned: On-device face recognition with access control.
- ESP32-CAM (Ai Thinker module, not sure)
- Servo Motor (SG90)
- Relay Module (5V)
- SD Card Module (SPI mode)
- Breadboard + Jumper Wires
- 5V Power Supply
| Component | Pin Connection |
|---|---|
| Servo | GPIO13 |
| Relay | GPIO12 |
| SD Card CS | GPIO5 |
| SD Card MISO | GPIO2 |
| SD Card MOSI | GPIO15 |
| SD Card SCK | GPIO14 |
- ✅ Live video stream from ESP32-CAM via browser.
- ✅ Capturing a still image from stream.
- ❌ Face recognition (due to incorrect ESP32 module without proper PSRAM support).
- ❌ On-device facial match and access logic.
- ❌ Servo and relay control.
- ❌ Capturing and storing images to SD card.
code/– Arduino sketches for streaming, locking, and SD card logging.report/– Final project report PDF.notes/– Known issues, debug logs, and what we learned.images/– Project images.
- Download the repository
- Get the correct hardware
- Make connections
- Run and flash the code with Arduino IDE
Even with partial success, this project taught us:
- How to work with ESP32-CAM hardware.
- Basics of IoT hardware interaction (camera, servo, relay).
- Debugging hardware-level issues.
- Importance of checking specs before building 😅.
- Aditya Chavan – Project Lead, Programming, Documentation @Aditya-Chavan
- Vrundarank Parite – Hardware Assembly, Testing, Co-Development @Vrundarank
See the full report in /report/Final_Report.pdf.
- Use correct ESP32-CAM board with PSRAM.
- Integrate OpenCV-based face matching via external processor (e.g., Raspberry Pi).
- Cloud-based image logging.
- Add Telegram/Email alerts on failed recognition attempts.
Thanks for checking out our project!