Prototype of a Waste Classification System Based on Deep Learning

This article presents the implementation of hardware tools such as Raspberry Pi, cameras, sensors, motors, and controllers, along with software components like convolutional neural networks and a mobile application for waste classification. In the future, the proposed waste collector and classifier implementation will contribute to environmental care and environmental education. The project’s innovation lies in the automation of waste classification using neural networks, automatic notifications generated by the prototype when a container is full and transmitted to the mobile application via a web server, and the flexibility of the prototype for various environments, including educational, office, and industrial settings. The advancements in the project include the creation of a mobile application to monitor container levels, the construction of the prototype, training results of selected neural networks, and the evaluation of the final network with test images.