A refined 3D reconstruction method for irregular ancient buildings

Abstract. The refined three-dimensional (3D) reconstruction of ancient buildings using unmanned aerial vehicles (UAVs) is a crucial task in cultural heritage preservation. However, the complex structure, diverse shapes and non-regularity of ancient buildings make it difficult for traditional 3D reconstruction methods to fully cover all the details, resulting in poor quality reconstruction models. Therefore, this study proposes a novel 3D reconstruction method tailored for irregular ancient architecture, which adopts a "structural surface" guided strategy, and plans routes through the idea of "structural surfaces segmentation and then designing the routes", so as to obtain a high-resolution image set, and to ensure that the reconstruction model's fineness and completeness. Additionally, to tackle the slow speed and time-consuming nature of traditional modeling algorithms in establishing image adjacency relationships, this paper presents a method that utilizes flight path information to build these relationships. To validate the effectiveness of the proposed method, the Yellow Crane Tower in Wuhan, Hubei Province, China, was selected as a case study. Results indicate that the proposed method successfully enables the refined reconstruction of the Yellow Crane Tower, providing significant insights for the 3D reconstruction of irregular ancient buildings.