Enhancing Sustainability in Smart Cities: IoT-Enabled with Fog and Haze Mitigation Using EDN-GTM Technology

Abstract

the growing reliance on Information and Communication Technologies (ICT) has prompted a shift towards smart city concepts, especially as over 70 per cent of the global population is expected to reside in urban areas by 2050. The smart city paradigm addresses the challenges of managing critical infrastructure by integrating intelligent ICT solutions across various sectors such as the economy, mobility, environment, people, living spaces, and governance. Central to this vision is the Internet of Things (IoT) role, acting as a pivotal network connecting devices and enabling seamless interactions in urban settings. Efficient energy usage is a key aspect of IoT-enabled smart cities, given the increasing demand for power with population growth. Lighting, a fundamental need, consumes a significant portion of energy, and an SLS becomes crucial for effective energy management in a smart city leveraging IoT technologies. The proposed system aims to enhance image quality and address fog-related challenges in Smart City applications by integrating IoT-based technology. In urban environments, foggy conditions can significantly impact visibility, affecting surveillance and monitoring systems. The system employs a comprehensive approach, utilising sensor data, image processing techniques, and deep learning models for efficient fog detection and removal. Key objectives include fog removal and image quality enhancement, focusing on adaptability to varying environmental conditions. The system dynamically adjusts fog removal processes based on real-time fog density information collected from cloud-based sensors. The dataset for training and evaluation is obtained from Kaggle, comprising fully foggy images. Pre-processing involves grey conversion and wavelet transform filters, simplifying image representation and extracting relevant features for subsequent stages. Segmentation with a region of interest (ROI) approach optimises processing efficiency by focusing on information-rich areas. Enhanced Dehazing Network (EDN) and Guided Transmission Map (GTM) models are separately implemented for haze removal and transmission map refinement. A hybrid approach combines the strengths of both models, aiming for superior fog removal and image quality improvement. Performance is assessed using metrics like PSNR, SSIM, TMQI, and FSIMc, providing a comprehensive evaluation. The proposed system, leveraging deep learning, adaptive processing, and IoT integration, offers an effective solution for mitigating the impact of foggy conditions in Smart City scenarios.