Seeing What Matters: A Systematic Review of Aerial Person Detection for Drone-Based Search and Rescue

Abstract

Drones have become critical tools in search and rescue (SAR) missions by providing real-time situation awareness and access to difficult terrain. However, operators must process large amounts of visual data under time pressure, creating significant cognitive demands. Object detection algorithms may support operators by automatically identifying potential human targets within drone imagery, yet most research has focused on technical performance rather than human-system integration. This systematic review synthesized 67 studies published between 2010 and 2025 examining object detection algorithms for drone-based human detection in SAR contexts. Findings showed that research overwhelmingly emphasized algorithm accuracy and computational efficiency, with operator workload and team performance rarely evaluated. Most studies relied on static datasets, limiting ecological validity. Human limitations were seldom measured directly, constraining human-centered evaluations in SAR operations.