Which Board for Autonomous Facial Recognition Drone?

Hello…I’m working on a pretty ambitious project—a drone that can autonomously identify a target using facial recognition and then shoot a dart using a slingshot mechanism. The drone is already equipped with a flight controller that handles all the flight dynamics, so the board I’m looking for would only need to process the facial recognition and send commands to the flight controller.

The key requirement is that the drone needs to be fully autonomous, with all the processing done on-board, without relying on a ground station. I was considering using a Raspberry Pi 4 for this, but it’s a bit on the expensive side, and I’m wondering if there are more cost-effective alternatives.

One concern I have is whether the drone moving too fast might cause blurring in the video feed, potentially affecting the accuracy of facial recognition. I assume this could be a result of low FPS, so I’m looking for a board that can handle high FPS and high-quality video processing.

I’ve heard that the ESP32 has a facial recognition library, but I’ve read that it suffers from a lot of delay. I also thought about using an STM32H7 since it has a higher clock rate than the Raspberry Pi 4, but I’m not sure if a higher clock rate directly translates to better FPS handling for facial recognition.

If anyone has experience with this or could suggest a more affordable yet powerful solution, I’d really appreciate your advice. Thanks in advance :wink:

For your ambitious drone project that autonomously identifies targets using facial recognition, consider using the NVIDIA Jetson Nano or Google Coral Dev Board as cost-effective alternatives to the Raspberry Pi 4. These boards offer sufficient processing power and GPU support for real-time video processing, essential for handling high FPS and reducing image blurring during fast movements. To further enhance facial recognition accuracy, select a high-quality camera capable of 60 FPS or more and implement image stabilization techniques. Additionally, optimize your software with efficient libraries like OpenCV to ensure low latency and effective decision-making. Testing in various flight conditions will help refine the performance of your drone.

1 Like

Hi,
For an autonomous facial recognition drone, consider using the NVIDIA Jetson Nano or Jetson Xavier board.