Application of AI-Enabled Computer Vision Technology for Segregation of Industrial Plastic Wastes

Abstract

Plastic waste generation is growing alarmingly at the global level. Since a major chunk of it is nonbiodegradable, it is detrimental to the health of the ecology. Mixed plastic waste hampers the recycling initiatives and imposes a grave and threatening challenge to sustainability efforts. Commercial plastics differ in chemical composition, and disposing of them makes recycling cumbersome and results in low-quality plastic with poor market potential and economic viability. Further, recycling mixed plastic waste is an energyintensive process that enhances operational costs and exacerbates carbon footprint. Mostly, the mixed plastic waste is either landfilled or incinerated, emanates greenhouse gasses, pollutes the environment, and inflicts healthhazardous effects to both humans and ecology. Unsound management of mixed plastic waste leads to microplastic generation and the leaching of toxic chemicals into the environment. Therefore, it is imperative to deal with mixed plastic waste efficiently by proper handling, segregating, and processing for a sustainable future. Different methods, such as manual sorting, sensor-based sorting, magnetic density separation, electrostatic separators, near-infrared (NIR) spectroscopy-based methods, robotics, artificial intelligence (AI), etc., have come to the fore as emerging technologies for effective segregation of plastics. Researchers in the field of AI are aiming to use AI 234successfully and effectively for sorting mixed plastic waste. Advancements in AI, convolutional neural networks (CNNs), and deep neural networks are emerging as powerful tools for the segregation of mixed plastic waste. Thus, computer vision employing image-based classification is of great help in segregating the plastics into poly(ethylene terephthalate) (PET), highdensity poly(ethylene) (HDPE), poly(vinyl chloride) (PVC), low-density poly(ethylene) (LDPE), poly(propylene) (PP), poly(styrene) (PS), and others based on their resin identification code. This book chapter demonstrates the application of computer vision technology-driven, image-based classification for segregating mixed plastic waste, enabling its amenability to recycling processes and, in turn, facilitating plastic waste management.