SQUIRREL DETECTOR / ESP32 APPLICATION
Squirrel Watch
AI-powered squirrel detection system using ESP32-S3 CAM, Edge Impulse, and Arduino Mega.
About this project
About this project
Two microcontrollers split the job: an ESP32-S3 CAM takes the picture and runs an Edge Impulse model, then raises GPIO1 when the `squirrel` class passes 0.50. That 3.3 V signal crosses a SparkFun BOB-12009 level converter into digital pin 7 of an Arduino Mega 2560, which lights a red LED for a squirrel or a green one for none. The model was trained on 140 images, and Edge Impulse reported 65.2% validation and 81.5% test-set accuracy; lighting, distance, background and camera angle can change that. Both firmwares are single `.ino` sketches, one per board.
From the README
From the README
- Feature
- detection threshold of 0.50
- Accuracy
- | Validation accuracy | 65.2% |
- Input
- The Arduino Mega reads the signal on Digital Pin 7.
- Build
- SparkFun BOB-12009 logic-level converter
Hardware
Boards: ESP32-S3
ESP32-S3. Verify the exact board revision, wiring and power requirements in the upstream build guide.
Media credit & rights
Upstream project media · rights remain with its owner · Credit: dishadometi
Screenshots from the upstream README. Code and firmware were not executed. License notices are included.
View original source ↗ Copyright & removal · Reviewed
More projects on these boards
Cardboard Robot Claude Status Display DFR1154 AI Camera Desktop Buddy ESP32 PC Remote Family Hub · Voice Notes
See what an ESP32 project looks like, and whether it runs on your board.
FILTERS
Browse by form factor. Check the exact board, chip and peripherals before installing.
Try a project
Newest listings
Give your project a little spotlight.
Repository links do not imply author endorsement or verified hardware compatibility. Check the upstream safety guidance before powering or flashing a build.