adwaith5002/Smart-Fridge — reverse-engineered prompt

Build me a smart fridge fruit freshness detector for an ESP32 CAM.

I want it to use a small CNN model trained in Edge Impulse to look at fruit through the camera and decide if it’s fresh or rotten. It should run directly on the ESP32 CAM, not depend on a cloud server, and print the result clearly over serial so I can test it while connected to my computer.

Please include the Arduino code for the ESP32 CAM, the Edge Impulse exported model setup, and simple instructions for wiring, flashing, and testing with the Arduino IDE. Use the AI Thinker ESP32 CAM board, a USB to TTL adapter, and a basic temperature sensor if needed for fridge condition checking.

Keep it practical for a prototype. I mainly need the camera capture, image resize to the model input, inference, and readable fresh or rotten output with confidence. Look up current ESP32 CAM and Edge Impulse docs online if you need to.