Murals carry cultural significance and historical information, and are an important channel for understanding ancient social norms, artistic styles, and religious beliefs. At present, the digitization of murals is an important technical means for the protection of cultural heritage. Orthogonal images of murals play a vital role in high-precision recording, preservation, academic research, educational expansion, mural protection, digital exhibition and dissemination. At present, orthogonal images of murals are mostly realized by plane projection, but this method is not suitable for making orthogonal images of arched and dome-shaped murals. To address this problem, this paper proposes a method for generating orthogonal expansion images of arched and dome-shaped murals. This method combines a three-dimensional virtual space simulation model with an RTT virtual camera and adopts a spatial reference orthogonal ray scanning model. First, the detailed three-dimensional color model is fitted to the geometric reference of cylindrical and spherical objects to determine its parameters. Next, for the cylindrical murals on the arch, the orientation of the model is initialized using quaternions, and the viewport matrix is adjusted to obtain the required resolution. Then, the RTT camera is used to perform line orthogonal projection in the viewport, and the fringe projection image is generated by rotating around the cylinder axis according to the inversely calculated rotation angle. For the murals on the dome ceiling, this method is used to segment them according to a certain longitude, and the circumscribed cylinder of the fitted sphere is rotated to perform cylindrical orthogonal line scanning in the segmented area. These individual orthogonal line scan images are carefully spliced together to form a complete orthogonal unfolded image. Finally, a fringe projection image is generated with the central meridian of the unfolded part as the center line, and the fringe projection images are spliced together to obtain the final orthogonal unfolded image. Experiments show that compared with existing methods, this method can generate two-dimensional orthogonal unfolded images with high texture fidelity, minimal texture deformation, and uniform deformation distribution. This study provides a novel perspective on the orthogonal unfolding of quasi-cylindrical and quasi-spherical painted objects, and provides an accurate and diverse data basis for the digitization of murals.
Read full abstract