AI-BASED VIDEO SURVEILLANCE FOR THREAT DETECTION AND PASSENGER SAFETY AT SAUDI AIRPORTS
TL;DR
It is argued that AI should augment rather than replace trained security personnel and that performance must be assessed using airport-specific measures such as detection probability, false-alarm burden, alert-to-intervention time, crowd-risk prediction, system availability and passenger-impact indicators.
Abstract
Saudi Arabia is rapidly expanding and modernising its aviation system under Vision 2030, with national targets that include handling 330 million passengers annually, connecting the Kingdom to more than 250 destinations, and strengthening its position as a global tourism and logistics hub. The scale and complexity of this transformation require airport security systems that can detect threats quickly while preserving passenger flow, privacy, operational resilience and public trust. This review examines the role of artificial intelligence (AI)-based video surveillance in supporting threat detection and passenger safety at Saudi airports. It synthesises research on object detection, multi-object tracking, video anomaly detection, crowd analytics, abandoned-object detection, violence recognition, perimeter monitoring, facial and person re-identification, multimodal sensor fusion, edge computing and human–AI decision support. The paper also evaluates the limits of current systems, including false alarms, dataset bias, domain shift, occlusion, adversarial manipulation, cyber risk, privacy concerns, explainability and the operational consequences of automation error in safety-critical environments. A Saudi airport implementation framework is proposed that integrates camera and sensor infrastructure, edge analytics, central AI services, airport operational databases, access-control platforms and security operations centres. The framework places human oversight, risk-based assurance, privacy-by-design, cybersecurity, model monitoring and phased validation at its core. The review argues that AI should augment rather than replace trained security personnel and that performance must be assessed using airport-specific measures such as detection probability, false-alarm burden, alert-to-intervention time, crowd-risk prediction, system availability and passenger-impact indicators. The study concludes that AI-enabled video surveillance can make a meaningful contribution to Vision 2030 when deployed through controlled pilots, representative Saudi datasets, interoperable architecture, transparent governance and continuous operational assurance.
