Application of Cross-Platform Adaptive AR in Open Landscape Heritage

Abstract

Aiming at the problems of hardware performance bottleneck, data fragmentation, and single evaluation dimension in AR exhibition of landscape cultural heritage, this study puts forward a set of closed-loop framework of “data collection-semantic reconstruction-adaptive rendering-experience evaluation.” By combining laser scanning, UAV tilt photography, and multi-spectral texture mapping, a multi-source data model with both geometric accuracy and texture realism is constructed. An adaptive rendering algorithm based on the mechanism of “perception-decision-execution-feedback” is studied and designed. By using the strategy of frame prefetching and semantic coordination, the visual fidelity is guaranteed, and the interaction fluency is significantly improved. The experimental results show that the algorithm can effectively reduce the cognitive load and prolong the user's stay time on multi-terminals such as HoloLens 2. This achievement provides a quantifiable technical paradigm and evaluation standard for digital protection and cross-platform immersive communication of cultural heritage.