Framework for Organizing Viewsheds From Multiple Observation Points: Enabling Bidirectional Landscape Visibility Information Queries

Viewshed analysis holds significant value in landscape assessment and management. However, current research faces dual limitations: inefficient organization of viewsheds from multiple observation points and inability to analyze bidirectional visibility relationships between multiple observation points and ground cells. This study develops a 3D tensor (observation point–row–column) that integrates detailed visibility relationships between landscape units and ground cells and proposes a framework for organizing viewsheds from multiple observation points that supports bidirectional queries. The framework implements a three-stage optimization strategy: 1) decimal encoding of observation point dimension data, 2) a mapping mechanism from ground cells to observation point sets, and 3) run-length encoding-based organization of mapping data. Experimental results demonstrate that when using ASCII storage format, the framework achieves a 98% storage space reduction compared to conventional viewshed matrix methods while maintaining comparable query efficiency, with storage requirements similar to those of NetCDF format. Moreover, with 7-Zip compression, the framework achieves further space savings: a 40% reduction relative to viewshed matrix methods and a 70% reduction compared to NetCDF format. This research not only enriches the theory of multiple-observation-point viewshed analysis but also provides technical reference for geographic information product development, offering important application value for landscape conservation and planning.