Adaptive Geographic Scene Expression Driven by User Needs: A Case Study of the Twin Highway Tunnel Scene

ABSTRACT Virtual geographic scenes, as a key component of virtual geographic environments (VGEs), play an essential role in conveying complex geographic information. However, the existing geographic visualization methods have problems such as separation from user needs and poor multi‐terminal adaptation, resulting in low user cognitive efficiency. To address these limitations, an adaptive geographic scene representation framework grounded in user requirements is proposed. A systematic mapping relationship between user requirements and scene representations was first established through rigorous analysis of multi‐platform user interaction patterns. Subsequently, data‐driven strategies for dynamic scene organization and scheduling were developed based on comprehensive terminal capability profiling. Furthermore, an integrated closed‐loop framework was developed, incorporating requirement analysis, task definition, scene organization, and interaction refinement to support adaptive scene description, analysis, and exploration. The efficacy of the proposed method was systematically assessed through eye‐tracking cognitive experiments and participant‐based questionnaires. Experimental results demonstrated that the proposed methodology facilitated smooth scene interaction rendering on contemporary terminal devices. In comparison with the original scenes, user cognitive efficiency in the designated areas of interest exhibited a 90.19% improvement, thereby establishing a novel approach for enhanced geographic information transmission.