Artificial intelligence-powered visual internet of things in smart cities: A comprehensive review

The field of smart cities has seen significant advancements in recent years to improve citizens' quality of life. Technologies such as the Internet of Things (IoT) and Edge Computing (EC), along with Artificial Intelligence (AI), are being utilized to achieve this goal. This study focuses on a specific branch of IoT known as Visual IoT, which uses digital cameras as sensors and relies on visual data. Advances in AI have enabled researchers to integrate AI models into camera-based edge devices, increasing the use of AI-powered Visual IoT systems in smart cities. However, since the energy consumption in battery-powered systems is naturally a concern, being deployed outdoors for visual data gathering with the integration of AI-based processing raises a significant challenge. This paper examines AI-powered Visual IoT systems in smart cities with a special emphasis on energy efficiency. Our goal is not only to evaluate how AI is used in Visual IoT systems in the context of smart cities but also to evaluate the level of consideration given to the energy efficiency aspect in the reviewed studies. Furthermore, we explore all of the methods used to address it. Through our work, readers will gain insights into the current landscape of Visual IoT in smart cities and an understanding of how much importance is placed on energy consumption in AI-integrated solutions.