The development by a SUAI graduate will help keep birds fed and courtyards clean.
The problem of small litter in city courtyards and parks has long become a common issue. In a large city, standard cleaning measures are often insufficient: trash bins fill up faster than they can be emptied, manual labor requires the constant presence of staff, cleaning equipment is limited in terms of mobility, and environmental initiatives are often one‑off events.
An unusual technical solution was proposed by Ilya Zubkov, a graduate of the Institute of Information Technologies and Programming: the smart bird feeder he developed not only rewards birds for bringing in trash but also sorts the items itself. The device accepts an object, analyzes it using a neural network, and dispenses food only if the bird has brought in trash.
The system is based on the ESP32-CAM microcontroller, which combines a computing module, a camera, and Wi‑Fi support. The bird places the object in the receiving compartment, where an infrared sensor, a camera, and LED lighting are located. The sensor detects the presence of the object, the camera takes a photo, and then the image is transmitted to the computer vision model. After recognition, the servo drive takes over. It rotates the flap on which the object is lying, and the object falls into the waste container. If the object was recognized as waste, after the dump, the feed dispensing mechanism is activated, and food is released from the hopper into the tray. The portion size can be adjusted programmatically.
The developer plans to expand the dataset for training the model so that it can better recognize different types of waste, conduct tests under real‑world conditions, and refine the design taking weather into account. It is also considering the possibility of adding solar panels for autonomous power supply and QR codes to simplify device linking.
