Algorithms for Recognition in Marker-Based and Spatial Augmented Reality

Augmented Reality has emerged as a transformative technology with applications spanning education, healthcare, manufacturing, and cultural heritage. Despite rapid advancements, one of the most critical challenges in Augmented Reality remains the accurate recognition and seamless integration of virtual objects into dynamic real-world environments. This study investigates the algorithmic foundations of marker-based and spatial Augmented Reality recognition techniques. Marker-based methods provide stability through fiducial markers, while spatial approaches utilize simultaneous localization and mapping and feature detection to achieve markerless interaction. Mathematical formulations are presented to model homography estimation and 3D point reconstruction. A comparative computational experiment demonstrates that marker-based systems achieve higher recognition accuracy in controlled environments, whereas spatial systems provide greater flexibility and adaptability in complex, unstructured settings. The results highlight the necessity of hybrid approaches that combine the strengths of both paradigms. The findings contribute to the development of robust Augmented Reality systems that balance accuracy, computational efficiency, and user experience.

Перевод пока недоступен