Plastic and Non-plastic Waste Classification Using Machine Learning Techniques

Because of its adaptability and scalability in packaging, plastic is being used more and consumed more, both of which are dramatically rising. Waste categorization and separation into different components is very import for the waste recycling and disposal. Typically, waste segregation is manual activity and to carry out this operation requires lot of efforts. This makes scaling only feasible for long-term monitoring across several sites. To address this issue, in this paper, waste recognition and classification technique is described. This work proposes a way to categorize wastes using their photos into three separate trash classes (plastic bottles, plastic bags, and non-plastics), based on the concept of computer vision and machine learning. In this paper, convolutional neural network model and VGG 16 model were used for waste classification. A total of 6512 images were used for the training and testing purpose. This paper also provides comparison of convolutional neural network model and VGG 16 model based on waste image classification.