Vision-Based Autonomous UAV Landing: A Comprehensive Review of Technologies, Techniques, and Applications

Autonomous landing capabilities are critical for unmanned aerial vehicles (UAVs) operating in challenging environments, yet remain among UAV operations’ most technically demanding aspects. This paper presents a comprehensive review of vision-based autonomous landing systems for UAVs, emphasising fiducial marker-based approaches. After systematically examining 143 papers published between 2018 and 2025, we critically analyse the evolution from traditional landing methods to advanced vision-based systems and evaluate their performance across diverse operational conditions. The review provides detailed analyses of marker design considerations, detection algorithms, performance metrics, and the impact of environmental factors like illumination, weather, and terrain on landing accuracy. We explore hardware implementation challenges, comparing FPGA, SoC, and GPU-based solutions for real-time vision processing on resource-constrained platforms. The integration of artificial intelligence accelerators and multi-sensor fusion approaches is examined, with quantitative comparisons of landing accuracy improvements across different methodologies. Contemporary trends, including swarm-based collaborative landing systems, self-reconfiguring markers, and machine learning-based adaptation to different environments, are discussed alongside regulatory frameworks and safety considerations. Real-world industrial implementations and technology transfer challenges are analysed to bridge the gap between academic research and commercial applications. This systematic review targets researchers, engineers, and stakeholders developing autonomous UAV landing systems for diverse applications, including emergency response, logistics, inspection, and environmental monitoring.