An evidence-constrained method for LOD reachability assessment in the virtual reconstruction of historic architectural complexes

Historic architectural complexes preserve important information on urban spatial structure, religious institutions, and socio-cultural evolution. However, war damage, urban redevelopment, and loss of historical records make their original forms difficult to verify, while virtual reconstruction often relies on modelling capability, display requirements, or expert judgement, weakening the correspondence between evidence conditions and geometric detail. This study proposes an Evidence-Constrained LOD Assessment (ECLA) method to introduce evidence conditions into the control of geometric representation levels. The method integrates multi-source evidence attribute modelling, Evidence Support Level (ESL) assessment, ESL-LOD interval mapping, and an intersection constraint model to determine the maximum achievable LOD, LOD max , for each object. A case study of Chongshan Temple in Taiyuan shows that ECLA produces differentiated LOD representations according to evidence support strength and prevents uniformly high-detail modelling of poorly documented objects, providing a traceable path for accuracy control and uncertainty representation.