Revolutionizing Urban Waste Management: Cutting-Edge Image Processing for Precise Solid Waste Classification

Due to rapid urbanization and population growth, urban waste management has become one of the major challenges faced by cities around the world. Solid waste classification and sorting are necessary for intensifying recycling, safeguarding environment and manage additional landfilling of materials that are reusable or recyclable, and are getting in landfill of solid waste in certain extent. In this study, an original ensemble model using the VGG19 and EfficientNet architectures is proposed to classify solid waste in urban area. This proposed model was applied on a dataset consists of 2,527 number of images in to six classes to increase the accuracy of classification as well as to maximize the efficiency of waste. It preprocesses the incoming data using various techniques such as image resizing, normalization and augmentation to make the dataset more diverse and robust. The ensemble mechanism performed better combining the detailed feature extraction skills of VGG19 and versatile and scalable architecture of EfficientNet. On test set, it outperformed individual models and the accuracy for the proposed model was 98.75% which closely followed, previously obtain the best result.