Intelligent Waste Sorting System: Leveraging Arduino for Automated Trash Identification and Categorization

Abstract

As the all encompassing community evolves and urbanization accelerates, the refuse namely produce persists to rise. This dirty trash has extreme consequences to the environmental atmosphere, moving the equilibrium of the all encompassing tangible balance. Trash discovery electronics can swiftly and correctly recognize, categorize, and settle various kinds of litter to accomplish the mechanical transfer and reliable reusing of waste, and it can likewise advance the growth of a circular frugality. However, the prior trash refuse discovery science has few problems, to a degree reduced accuracy and a weak detection effect in difficult surroundings. Even YOLOv5 has realized good results in refuse discovery, the detection results cannot meet the necessities in complex sketches, so this paper uses, YOLOv5x, an enhanced YOLOv5 model. We, in this research work would like to infuse custom dataset, a comprehensive collection of images of waste categories divided into seven classes with the object detection models like YOLO towards automating trash detection and precise classification that helps in achieving better waste sorting accuracy. The research work core ambition is to exploit distinctive attributes of YOLO models combined with Arduino to give rise to a more effective trash sorting system. Traditional garbage management processes is very labour intensive work, making them time- consuming, costly, and prone to human error. To overcome these short comings, we propose the development of a deep learning-based model that can swiftly detect and classify various types of waste items in real-time. YOLO, a state-of- the-art object detection algorithm, will serve as the backbone of our model, ensuring fast and accurate identification of trash items. The integration of an Arduino microcontroller into our system adds a practical and interactive dimension. Arduino will facilitate seamless communication between the deep learning model and the physical world. This enables our system to trigger actions such as sorting, recycling, or alerting authorities, depending on the detected trash item and its classification.