A Scan-to-AR Knowledge Management Framework for Cultural Landscapes

Abstract

This study constructs and validates a knowledge-centric “Scan-to-AR” closed-loop framework for cultural landscape documentation, spanning knowledge capture, representation, transfer, and internalization. Applied to three Western Sichuan ancient towns, the framework integrates multi-source scanning, adaptive mesh optimization, cross-platform AR rendering, and five-dimensional evaluation. The mesh algorithm reduced memory usage by 42% with <3 cm accuracy. Cross-platform testing maintained >30 FPS across iOS, Android, and HoloLens. A visitor study (n=60) showed AR tours improved experience scores by nearly 30% over traditional tours, with texture fidelity negatively correlated with information efficiency (r=-0.63). Theoretically, it operationalizes SECI and extends transfer theory by identifying representational fidelity as a knowledge stickiness antecedent. Practically, it offers a lightweight solution for heritage preservation and immersive dissemination.