Virtual restoration guidance practice

A crucial part of "evidence-based restoration" is "virtual restoration," which simulates the restoration of a building's past using evaluated data to determine whether and to what extent the objectives of the first phase can be accomplished. If the results support restoration goals, then the restoration moves on to the second phase of data-based research, where the evidence is adjusted and analyzed until the virtual restoration goals are met. If the computerized response to the research questions is accurate and the goals are met, then the virtual restoration plan is valid, and the restoration moves to physical restoration.

The two types of virtual restoration are ornamental restoration and structural repair. The outsides of historically significant buildings are frequently restored using semantic automated restoration based on deep learning, while external virtual restoration is frequently handled using image processing technology. Wang et al. [54] performed the virtual restoration of cracked murals using top-hat to extract crack information, queer segmentation, morphological operations, and sample-based restoration algorithms, while Zou et al. [55] used U-Net to convert the unrecognizable color restoration problem into a semantic segmentation problem for some old and dilapidated buildings, where the naked eye could no longer distinguish colors.

It is crucial to restore the structural integrity of historic buildings. If structural heritage is not properly restored in terms of appearance, then it will lose its original flavor and value and look inappropriate, while improper structural restoration will cause such heritage to vanish. Because wood structural damage mechanisms are more complex, it is reasonable to integrate multidisciplinary, clear repair ground research problems, the development of a correct and reasonable repair strategy, technical solutions and engineering systems, and virtue engineering. Cracks, bending, rolling, tenon pulling, and other forms of structural damage are common in wooden structures; traditional structural repair and reinforcement rely on the craftsperson's experience and judgment [56–58]. The procedure and method for the virtual structural restoration of historic buildings can be summarized as the integration of analytical data through data-based research, combined with infrared thermography, three-dimensional stress waves, and other technologies, further combined with related experimental investigations, surveying the structure or component damage characteristics information, and the construction of a three-dimensional refinement of the building information model [59, 60]. Finite element static analysis and virtual repair are used to evaluate the safety performance before and after repair. Additionally, various cumulative damage time-varying models have been built based on member damage information to assess the reliability of member repair and provide reliable theoretical research for physical repair design [61, 62]. This research offers a trustworthy theoretical foundation for physical repair design.