Empirical Validation of Integrated Digital-to-Physical Workflow for Heritage Restoration

Overview of the Framework and Case Study

This section presents empirical results and validation of the integrated digital-to-physical workflow for historic heritage building (HHB) restoration, highlighting a case study of a masonry screen wall in Lanxi City, China.

Restoration Digitisation

The successful implementation of the Heritage Lifecycle Restoration System (HLRS) hinges on accurately capturing, analyzing, and reconstructing deteriorated structures in a digital realm.

Phase 1: 3D Reconstruction Accuracy

• Photogrammetry Method: Employed for high geometric accuracy using high-resolution imaging.
• Kinect Scanning: Used for real-time scanning, allowing rapid data collection.

Phase 2: Brick Detection Following Segmentation

Utilizing the YOLACT algorithm for precise identification and segmentation of bricks and mortar joints.

Phase 3: Modelling Method for 3D Model

Reconstructing the masonry structure digitally, ensuring true representation of size and arrangement.

Phase 4: Volumetric Design and Bounding Box Voxel Size Definition

Defining voxel size based on brick types to maintain structural integrity in reconstruction.

Phase 5: Input Matrix Setting, Rotation, and Connection Mechanisms

Implementing WFC algorithm for structural arrangement, ensuring adherence to original design logic.

Phase 6: Structure Aggregation for Digital Retrieval

Producing a cohesive digital model that guides the physical restoration process.

Phase 7: FEA for Structural Assessment

Performing finite element analysis (FEA) to validate structural integrity prior to physical restoration.

Robot System Innovation and Communications

Presenting the UR-MetaBridge framework for seamless integration of augmented reality (AR) with robotic manipulation.

Augmented and Automated Assembly

Validation of the workflow through physical repairs, highlighting the synergy of robotic precision and human oversight.

Evaluation Model

Establishing core indicators for 3D scanning accuracy and robotic grasping performance to optimize the restoration process.

Pose Estimation Accuracy Metrics

Implementing dual geometric metrics for assessing pose estimation in robotic systems, enhancing operational reliability and precision.

Integrating Digital and Physical Workflows for Heritage Restoration: A Case Study of HHB in Lanxi City

Introduction

The preservation of historical structures is a critical aspect of cultural heritage conservation. In this context, the empirical results from a recent project demonstrate the efficacy of an integrated digital-to-physical workflow for the restoration of deteriorated historic masonry. This post delves into the validation and findings of this approach, focusing on a historic masonry screen wall in Lanxi City, China.

The Challenge of Authentic Heritage Structures

Choosing a genuine, deteriorated historic structure over a controlled laboratory specimen presents unique challenges. Factors like material decay, non-standard geometries, and structural asymmetries complicate conservation efforts. However, testing the integrated digital workflow in this real-world scenario validates its robustness and practical applicability, even when confronted with the unpredictable conditions of an authentic heritage site.

Restoration Digitisation: A Systematic Approach

Successful restoration hinges on accurately capturing, analyzing, and reconstructing the structure in the digital realm. The workflow comprises several phases:

Phase 1: 3D Reconstruction Accuracy

Photogrammetry and Kinect Scanning:
Using a high-resolution SONY DSLR camera, researchers conducted photogrammetry with over 70% image overlap, achieving impressive geometric accuracy (mean deviations within ±10 mm). In parallel, real-time data collection was executed with a Microsoft Kinect sensor, providing rapid point cloud generation, albeit with slightly lower accuracy.

Phase 2: Brick Detection

The YOLACT algorithm facilitated precise identification of individual bricks, distinguishing them from surrounding mortar joints. Domain-specific constraints ensured that overlapping predictions were resolved, yielding a detailed map of the brickwork necessary for subsequent reconstruction.

Phase 3: 3D Model Reconstruction

Segmentation data informed the digital reconstruction of the masonry structure, ensuring the 3D model accurately mimicked the original form. This allowed for detailed planning regarding the restoration process.

Phase 4: Volumetric Design

The bounding box voxel size for the digital model was derived from the bricks’ type and scale, establishing a crucial input matrix to maintain structural integrity during restoration.

Phase 5: Input Matrix Setting and WFC Algorithm

Utilizing the Wave Function Collapse (WFC) algorithm, researchers established constraints based on local geometric relationships, ensuring each block’s arrangement adhered to the original design principles.

Phase 6: Digital Retrieval

The processed input matrix was aggregated into a cohesive digital model, serving as a blueprint for the physical restoration.

Phase 7: Finite Element Analysis (FEA)

Before actual construction, FEA assessed the structural integrity of the digital model under various loading conditions. The findings indicated that the reconstruction met acceptable stress and displacement thresholds.

Innovative Robot System and Communications

Seamless integration between immersive technology and industrial robots opens new avenues for restoration. The UR-MetaBridge framework connects virtual reality functionality with robotic manipulators to enable real-time control. This innovative system enhances the efficiency of operations while ensuring safety through gesture recognition and intelligent path planning.

Augmented and Automated Assembly

The efficacy of the digital workflow was validated through physical repairs on moderately damaged wall specimens. The approach marries robotic precision with human oversight, achieving a balance between efficiency and quality control.

Evaluation Model: Metrics for Success

An evaluation model was established to quantify the results of the restoration process. Three core indicators—geometric accuracy, surface flatness, and feature fidelity—were analyzed to ensure the reliability of outcomes. Additionally, the success rate of robotic grasping and pose estimation accuracy were thoroughly measured, providing a comprehensive understanding of the system’s performance.

Conclusion

The integrated digital-to-physical workflow for heritage restoration exemplifies a significant advancement in the field. By effectively bridging the gap between historical preservation and modern technology, this approach not only safeguards cultural legacy but also enhances the accuracy and efficiency of restoration processes. As we embrace innovative solutions, we continue to redefine the future of heritage conservation.