Addressing the critical need for autonomous navigation of low-altitude unmanned aerial vehicles (UAVs) in Global Navigation Satellite System (GNSS)-denied environments, this paper presents a comprehensive survey on absolute visual localization techniques based on a "retrieval-matching-pose estimation" framework. It begins by analyzing the challenges inherent to low-altitude UAV observation, such as significant imaging disparities, rapid scene scale variations, and object occlusions, there by elucidating the advantages of the hier-archical framework for large-scale, long-endurance localization tasks. Subsequently, the review systematically examines the technological evolution and current state-of-the-art across three core components: cross-view image retrieval, pixel-level feature matching, and UAV pose estimation, tracing the progression from traditional handcrafted features to deep learning paradigms. Finally, considering the requirements for deployment on airborne edge-computing platforms, the paper discusses the limitations of existing technologies and outlines promising future research directions. This survey is intended to serve as a valuable reference for both research and practical applications in absolute visual localization for low-altitude UAVs.