IoT and WSN-Driven Solid Waste Image Classification in Smart Cities Using the Attention BiGRU Model

Waste management has become an important issue in smart cities due to the dramatic increase in solid waste output caused by fast urbanisation and population growth. Inefficient garbage collection and disposal are results of using antiquated techniques in conventional municipal solid WMS. Modern technology advancements, especially in the field of solid waste image categorisation, are required to address the growing trash volume. The AtBiGRUCN model is introduced in this paper as a new DL that improves classification accuracy by combining RNNs, CNNs, and AM. In order to overcome obstacles in garbage can image processing, the model optimises feature extraction through the use of various combinations of shape, texture, and colour. To solve Text-CNN's positional feature loss problem, AtBiGRUCN learns spatial and contextual information effectively. Applying AtBiGRUCN with all feature combinations outperforms existing models, with experimental evaluations revealing a classification accuracy of 97.45%. Smart city waste management through solid waste image classification could be a promising area for DL, according to their results. More intelligent and automated garbage disposal systems are on the horizon, according to the suggested model's scalable and efficient solution, which will enhance sustainability and urban cleanliness in today's cities.