<p>Crack and defect detection in civil engineering structures, including pavements, buildings, bridges, and roads, is vital for ensuring structural integrity, public safety, and long-term asset sustainability. While traditional detection methods have been widely used, they are often inefficient, inaccurate, and lack scalability, especially in real-time monitoring scenarios. The upward traffic volumes and growing complexity of structural systems aggravate these challenges. The key objective of this study is to discuss and review the currently utilized technologies as on crack detection and highlight the gap and future direction. Therefore, this review systematically analyzes modern developments in automated crack and defect detection. Besides, a review study was exploited to provide and discuss state-of-the-art Deep Learning (DL) approaches, as well as exploring the integration with advanced technologies as Internet of Things (IoT) and Digital Twins to facilitate real-time monitoring and predictive cack detection for smart road management in the context of smart cities. The study explores the limitations of traditional approaches and estimates how emerging knowledge can address problems for instance detection accuracy, workflow automation, and complex data management using Artificial Intelligence (AI) approaches as DL. The outcomes display the transformative potential of DL-based techniques when combined with IoT and Digital Twins, presenting smart crack detection, real-time data processing, and predictive potentials for smart road monitoring in the future. Also, this work explores key challenges, and future research scope, including enlightening interoperability, and enhancing real-time decision-making of crack detection for intelligent road management. This work presents a critical review for researchers and practitioners, suggesting a roadmap for leveraging advanced DL approaches for detecting pavement cracks methods to promise smart and resilient road infrastructure.</p>