PigeonPal

Our project tackles the issue of food waste and uncontrolled pigeon feeding in urban areas. We’ve built an AI-powered smart pigeon feeder that efficiently dispenses food, ensuring sustainable feeding without waste. It’s eco-friendly and helps urban wildlife coexist better with humans. Our system integrates multiple components, containing an ESP32-CAM with a frame detection on Edge Impulse to detect pigeons using image recognition, and we got our pigeon images with a web crawler. When a pigeon is detected, the cam triggers the ESP32, which triggers the servo motor to open the door. Our code integrates image processing with hardware control, and we ensured that the whole system operates reliably in different conditions. We faced challenges like managing power consumption for the servo motor and making a connection between our ESP32 and our ESP32 cam. We tried to connect our servo directly to the cam, which should work theoretically, but it didn’t, so we resorted back to using an ESP32. We learned a lot about hardware integration, AI optimization, and sustainable design. We faced challenges integrating the AI model on the ESP32 due to limited resources, but we worked together to optimize it. Our team handled both hardware and software components equally, ensuring smooth collaboration. We had 3 ECE students and 1 Mech, so the tasks were split evenly; each member contributed based on their strengths. To conclude, this smart pigeon feeder promotes sustainable feeding, prevents food waste, and supports urban wildlife. Our system is scalable and could be deployed in parks and urban areas to encourage controlled pigeon feeding. In the future, we could improve it by integrating more sensors and refining our AI model further. It could also be incorporated with IoT to track pigeon species (potentially other bird species as well). As well, we can also add a solar panel to further add to our sustainability contributions.