With the rapid development of digital technology, the digital preservation of murals has become an important means of cultural heritage protection. As immovable cultural heritage, frescoes face threats such as natural erosion and human damage. Traditional mural preservation methods are often difficult to fully capture the geometric features and texture details of murals, while modern 3D reconstruction and texture mapping techniques provide new possibilities for high-precision digitization of murals. In this study, the combination of texture mapping and incremental SFM algorithm is used to realize the complete digitization of geometric forms and texture details of murals by establishing the relationship between point cloud data and high-definition image data. The experimental results show that the processing time of the improved ISIFT feature matching algorithm on low-resolution mural images is 63.19 s, which is lower than the other four algorithms; in sparse point cloud construction, the incremental SFM algorithm has the highest accuracy with reconstruction error between 0 and 0.02, although it takes a relatively long time (1.7 s to 2.0 s); in dense reconstruction, the depth fusion method based on the dense In dense reconstruction, the number of point clouds generated by the depth fusion-based point cloud construction method is significantly higher than that of the SFS algorithm and the C/PMVS algorithm, and the results are more superior. In addition, the texture mapping and rendering technique proposed in this study nearly doubles the efficiency of the traditional mural splicing technique. The study confirms that the fresco digitization scheme proposed in this paper not only can preserve the geometric and textural information of frescoes more accurately, but also greatly improves the processing efficiency, which provides a precise and efficient technical path for the digital preservation and inheritance of frescoes.
