Drone Watcher 🛸

Detect and log drones using SenseCAP Watcher + Wio Terminal

---

Overview

A compact system that detects drones in real time using computer vision (TinyML or Vision LLM) and logs events to a Wio Terminal.

Designed to be:

• Simple to replicate
• Low-cost
• Modular and hackable

(https://www.youtube.com/watch?v=HC7j8KhW98I)](https://www.youtube.com/watch?v=HC7j8KhW98I)

---

Features

• Real-time drone detection (Vision LLM or TinyML)
• Visual + audio alerts (screen, LED, speaker)
• UART communication between devices
• On-screen event log (Wio Terminal)
• Persistent logging to microSD (/drone_log.txt)
• Tripod-ready enclosure
• No custom ML training required (optional)

---

How it works

1. Detection SenseCAP Watcher monitors the scene through its camera.

2. AI Processing Uses either:

   • Custom TinyML model
   • Default Vision LLM

3. Trigger On detection:

   • Screen updates
   • LED flashes
   • Sound alert

4. Logging Event is sent via UART to Wio Terminal and stored on microSD.

---

Hardware

• SenseCAP Watcher (AI camera)
• Wio Terminal (SAMD51)
• Jumper wires (male-male)
• microSD card
• 3D printed enclosure (optional)
• Tripod (1/4”-20 mount)

---

Wiring (UART)

SenseCAP Watcher	Wio Terminal
GND	GND
5V	5V
Pin 19 (TX)	Pin 10 (RX)
Pin 20 (RX)	Pin 8 (TX)

---

Software

Arduino IDE setup

Add board manager URL:

https://files.seeedstudio.com/arduino/package_seeeduino_boards_index.json

Select:

Seeeduino Wio Terminal

Install libraries:

#include <Seeed_Arduino_FS.h>
#include <ArduinoJson.h>

---

SenseCAP Configuration

Using SenseCraft App:

1. Pair device
2. Select SenseCAP Watcher
3. Define detection in natural language

Example:

"If you see a drone, notify"

Then configure:

• Model: TinyML or Vision LLM
• Actions: LED / sound / screen / UART
• Capture frequency

---

Usage

• Device starts monitoring automatically on power
• Wio Terminal displays detection log
• Use right button to scroll
• Press button to clear log

Log file:

/drone_log.txt

---

## References

* https://wiki.seeedstudio.com/sensecraft-ai/tutorials/sensecraft-ai-pretrained-models-for-watcher/
* https://wiki.seeedstudio.com/Wio-Terminal-Getting-Started/

---

## Notes

* Vision LLM works well for standard objects like drones
* TinyML is useful for custom detections
* UART allows easy integration with other systems

---

## License

MIT