Waste Classification using Edge based Multilayer White Shark Dense Net-Spiking Neural Network from Urban Areas

Effective waste management is crucial for human health and a clean environment. Segregating waste based on types and classification is essential for recycling. Traditional methods, like manual sorting, are labor-intensive and error-prone, leading to environmental pollution and resource depletion and to classify the waste several deep learning algorithms used, but the accuracy is decreased, computational time, cost and complexity is increased. To overcome these issues, this work is proposed. This study proposes a Waste classification using Edge based Multilayer White Shark Dense Net-Spiking Neural Network (EBMW2SD-Net) from urban Areas. In this, for classifying the waste images urban areas are collected using Stanford TrashNet dataset. These images are collected from the various IoT sensors, so that, these images are full of noises, so, that these data are pre-processed and its features are extracted using the Orthonormal S-Transforms with Multiple Discrete (OSTMD). Then these data are classified using the Edge based Multilayer skill Dense Net-Spiking Neural Network method with seven categories (cardboard, glass, metal, organic, paper, plastic, and trash). Then, cardboard, organic, and paper class images are considered biodegradable waste, and other classes are considered non-biodegradable waste. The White Shark Optimization Algorithm (WSOA) further improves prediction accuracy and reduces error rates, with the entire model implemented on the Python platform. From the analysis the EMHS-ParCNN model attains 99.9% accuracy, 99.8% recall and attains better results compared with the existing methods.