A Virtual Restoration and Display System and Method for Grotto Statues Based on AI

Abstract

This invention discloses an AI-based virtual restoration and display system and method for grotto statues, belonging to the field of grotto statue restoration technology. It includes a data acquisition module, an intelligent restoration processing module, and a display output module connected in sequence. This invention acquires complete 3D and texture data of the statues through a non-contact data acquisition module, completing the entire virtual restoration process in digital space without any physical contact or intervention with the grotto statues. This effectively avoids the risk of secondary damage caused by traditional physical restoration, conforming to the principle of minimal intervention in cultural relic protection. Simultaneously, this invention can generate multiple differentiated restoration plans through a CSN artificial intelligence model, supporting experts to switch, adjust, and revoke restoration operations at any time, making the restoration process completely reversible. Furthermore, a secure encryption unit completely records the operation trajectory and model parameters of the entire restoration process, achieving full traceability and auditability of the restoration process, meeting the standardized management requirements of cultural relic restoration.

Description

Technical Field

[0001] This invention belongs to the field of grotto image restoration technology, specifically, it relates to an AI-based virtual restoration and display system and method for grotto images. Background Technology

[0002] Grotto sculptures are an important immovable cultural heritage of my country, carrying artistic, religious, and socio-cultural information from different historical periods and possessing extremely high historical, artistic, and scientific value. Affected by natural weathering, geological disasters, and human damage, a large number of grotto sculptures in my country have suffered varying degrees of damage and loss, urgently requiring scientific protection, restoration, and utilization. Currently, the restoration and protection of grotto sculptures mainly relies on traditional manual physical restoration. This approach depends on the personal experience of cultural relic restoration experts, using methods such as patching, reinforcement, and coloring to physically intervene in the damaged sculptures, and is currently the mainstream method for grotto sculpture restoration. With the development of digital technology, some research institutions have begun to use 3D scanning technology to digitally archive grotto statues. Artificial virtual restoration schemes based on general 3D modeling software and AI restoration technology for 2D cultural relic images have also emerged. In terms of displaying and utilizing grotto statues, existing technologies are mainly divided into three categories: one is the physical display of the statues at their original sites after restoration; another is the 2D display through pictures, videos, online digital exhibition halls, etc.; and the third is a simple AR / VR virtual display based on a preset general 3D model, which can only achieve basic overlay viewing of a fixed model.

[0003] However, the aforementioned existing technologies still have certain shortcomings in practical applications: Traditional manual restoration requires restorers to directly contact the artifact itself, and the polishing and repair operations during the restoration process can easily cause secondary damage to the already fragile grotto statues. Moreover, if problems such as restoration deviations or stylistic inconsistencies occur after restoration, they cannot be reversed, which may violate the principle of "minimal intervention" in cultural relic protection. Furthermore, both physical restoration and existing artificial virtual restoration rely heavily on the personal experience and artistic literacy of cultural relic restoration experts, resulting in low restoration efficiency, long cycles, and restoration effects that are greatly affected by human factors, without forming a standardized restoration process. Existing AI restoration technologies are mostly developed for two-dimensional cultural relic images and cannot adapt to the restoration needs of large-scale, complex three-dimensional cultural relics such as grotto statues. They also have not built a specialized knowledge base for the styles of grotto statues from different regions and eras, and may not be able to automatically generate restoration plans that conform to cultural relic restoration standards.

[0004] Existing physical displays can only present the final state after restoration, and cannot show the public the restoration process and the differences before and after restoration; two-dimensional displays lack spatial immersion and cannot allow the public to intuitively experience the three-dimensional structural features of the statues; existing AR / VR displays mostly use pre-made fixed models, which cannot be linked with the actual data collected on-site from the damaged statues, resulting in insufficient positioning accuracy, poor virtual-real fusion effect, and a lack of interactive functions that meet the needs of public visits, and may not be able to meet the dual needs of cultural relic research and public education. Summary of the Invention

[0005] To address the aforementioned problems and technical deficiencies, the present invention adopts the following technical solution: an AI-based virtual restoration and display system for grotto statues, comprising a data acquisition module, an intelligent restoration processing module, and a display output module that are sequentially connected in communication.

[0006] The data acquisition module is used to acquire three-dimensional point cloud data, surface texture color data, and on-site spatial coordinate data of the grotto statues; The intelligent restoration processing module is a locally deployed edge processing unit used to receive multi-source data transmitted by the data acquisition module, complete data calibration and matching, identify the damaged areas of the statue based on the pre-trained CSN artificial intelligence model, generate a virtual restoration plan that conforms to the cultural relic restoration specifications, and output a complete three-dimensional digital model after restoration. The display output module is used to receive the complete three-dimensional digital model, complete the precise registration of the virtual model with the original physical statues of the grotto through the spatial positioning unit, output the display content in a virtual-real fusion manner, and support user interactive operation.

[0007] Preferably, the data acquisition module includes a 3D laser scanner, a high-definition texture acquisition camera, an ambient light compensation sensor, and a positioning and orientation device; the 3D laser scanner is used to acquire 3D point cloud data of the grotto statues, the high-definition texture acquisition camera is used to acquire surface texture and color data of the statues, the positioning and orientation device is used to acquire on-site spatial coordinate data of the statues, and the ambient light compensation sensor is used to acquire on-site lighting data to assist in texture calibration.

[0008] Furthermore, the intelligent restoration processing module includes an edge computing server, a CSN model inference acceleration card, and a data storage array. The edge computing server is used to complete the format conversion, calibration, and matching of multi-source data, and establish a one-to-one correspondence between point cloud data, texture color data, and spatial coordinate data. The CSN model inference acceleration card is used to load a pre-trained grotto style knowledge base and run the CSN model to complete the extraction of statue features, identification of incomplete areas, generation of restoration schemes, and model reconstruction. The data storage array is used for encrypted archiving of multi-source data, restoration models, and process data.

[0009] Furthermore, the intelligent repair processing module also includes a security encryption unit, which is used to encrypt the transmission process from the data acquisition module to the processing module and the storage process of the repair model, while recording the operation trajectory of the repair process to achieve full traceability.

[0010] Preferably, the display output module includes an AR display unit, a UWB positioning base station, and an interaction unit; the UWB positioning base station is used to establish a three-dimensional spatial coordinate system on site to achieve millimeter-level real-time spatial positioning; the AR display unit includes at least one of AR glasses and a handheld terminal, used to spatially register and overlay the restored virtual model with the original physical image; the interaction unit is used to receive user interaction commands and complete operations such as switching between views before and after restoration, adjusting model transparency, and zooming in to view details.

[0011] Preferably, the display output module further includes a projection display unit, which includes a short-throw projector, a fusion processor and an infrared positioning system, used to project the restored virtual model to the corresponding position of the original site of the grotto statue through projection fusion, so as to realize the in-situ display of virtual and real fusion.

[0012] A method for virtual restoration and display of grotto statues based on AI, comprising the following steps: S1. Data Acquisition: Acquire the 3D point cloud data, surface texture and color data, and on-site spatial coordinate data of the target grotto statue; S2. Intelligent Virtual Restoration: After format conversion, calibration and correlation matching of the collected multi-source data, the system identifies the missing areas of the statue based on the pre-trained CSN artificial intelligence model, generates at least two differentiated restoration schemes that conform to the cultural relic restoration standards, and outputs the complete three-dimensional digital model after restoration after confirmation. S3. Virtual-Real Fusion Display: Establish a spatial positioning coordinate system on site, spatially register the restored 3D digital model with the original physical statue, and complete the in-situ display in a virtual-real fusion manner, while supporting user interaction.

[0013] Preferably, in step S2, the CSN artificial intelligence model is preloaded with a grotto style knowledge base, which contains the three-dimensional structural features, texture and color rules, and restoration case library of grotto statues from different regions, eras, and schools; the model completes the extraction of statue features through point cloud feature extraction network and texture feature extraction network, identifies the missing areas through three-dimensional semantic segmentation and defect detection algorithms, and generates a restoration plan following the principles of minimal intervention and restoration to the original state. Furthermore, in step S3, a three-dimensional spatial coordinate system with the same source as the collected data is established by deploying UWB positioning base stations in the display area, and the positioning error is controlled at the millimeter level; the feature points of the physical statue are captured by the AR terminal and matched with the corresponding feature points of the virtual model to complete spatial registration, so that the virtual model and the physical statue are completely aligned in terms of spatial position, scale and angle.

[0014] Furthermore, it also includes step S4, data archiving and maintenance: encrypting and archiving the operation trajectory, model parameters, and metadata of the entire repair process, recording user interaction data to optimize the displayed content, and supporting remote experts to provide annotation guidance and adjust the solution through the system.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention adopts a fully digital technical approach, which acquires complete three-dimensional and texture data of the statue through a non-contact data acquisition module, and completes virtual restoration in the digital space throughout the process. No physical contact or intervention is required on the statue itself, which effectively avoids the risk of secondary damage caused by traditional physical restoration and conforms to the principle of "minimal intervention" in cultural relic protection. At the same time, this invention can generate multiple sets of differentiated restoration plans through the CSN artificial intelligence model, which supports experts to switch, adjust and cancel restoration operations at any time. The restoration process is completely reversible. At the same time, the operation trajectory and model parameters of the entire restoration process are completely recorded through a secure encryption unit, so as to realize the full traceability and auditability of the restoration process and meet the standardized management requirements of cultural relic restoration.

[0016] (2) The present invention adopts a CSN artificial intelligence model with a pre-trained grotto style knowledge base, which can automatically complete the extraction of statue features, identification of missing areas, and generation of restoration schemes that conform to the cultural relic restoration standards. Only experts are needed for final confirmation and fine-tuning, which greatly reduces the absolute dependence of the restoration work on the personal experience of experts, shortens the restoration cycle, and forms a standardized and replicable virtual restoration process for grotto statues, which can be adapted to the restoration needs of grotto statues in different regions and eras.

[0017] (3) This invention integrates the data acquisition module, intelligent repair processing module and display output module into a complete system that is connected in sequence, realizing a closed loop of the entire process from data acquisition and intelligent repair to virtual and real fusion display. It does not require multiple independent devices and multiple professional personnel to cooperate. The locally deployed edge processing unit can complete the entire process operation on the grotto site, thereby reducing the application threshold and cost, and is suitable for rapid deployment and promotion in grassroots grotto heritage sites.

[0018] (4) This invention achieves millimeter-level on-site spatial positioning of ±5mm through UWB positioning base stations, which can accurately spatially register the repaired virtual model with the original incomplete statue, achieving seamless integration of virtual and real; at the same time, it provides two modes: AR personal immersive display and projection multi-person synchronous display, supporting the front and back views of the repair. Figure 1 With multiple interactive operations such as key switching, stepless adjustment of model transparency, and zooming in on details, it can not only provide research assistance for cultural relic restoration experts, but also intuitively present the state of the statue before and after restoration and its original historical appearance to the public, realizing the dual value of cultural relic protection and public education. Attached Figure Description

[0019] In the attached diagram: Figure 1 This is a system diagram from an embodiment of the present invention; Figure 2 This is a flowchart of the steps in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. Generally, the components of the embodiments of the present invention described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0021] Example 1: like Figure 1 As shown, an AI-based virtual restoration and display system for grotto statues includes a data acquisition module, an intelligent restoration processing module, and a display output module, which are connected in sequence. The specific composition, equipment selection, and connection relationship of each module are as follows: Data acquisition module: The data acquisition module includes a 3D laser scanner, a high-definition texture acquisition camera, an ambient light compensation sensor, and a positioning and orientation device. Each device communicates with the local acquisition terminal via a gigabit wired network.

[0022] In this embodiment, the FAROFocus S350 3D laser scanner, with a ranging accuracy of ±2mm, is used to perform omnidirectional scanning of the grotto statues and acquire their 3D point cloud data. The high-definition texture acquisition camera, a Sony A7R4 full-frame SLR camera with a 50mm f / 2.8 fixed-focus lens and 61 million effective pixels, is used to acquire texture and color data of the statue surface. The positioning and orientation device, a Huace X12 GNSS receiver in conjunction with a total station, is used to acquire the on-site spatial coordinate data of the statues. The ambient light compensation sensor, a Konica Minolta CL-500A lux meter, is used to acquire on-site illumination data within the grotto to assist in subsequent texture and color calibration.

[0023] Intelligent Restoration Processing Module: The intelligent restoration processing module is a locally deployed edge processing unit, located in the cultural relic protection workstation at the grotto site. It includes an edge computing server, a CSN model inference acceleration card, a data storage array, and a security encryption unit. Each unit is connected to the server motherboard via a PCIe bus, and internal data transmission uses a gigabit intranet.

[0024] In this embodiment, the edge computing server is a rack-mounted server equipped with an Intel Xeon 8375C CPU and 256GB of memory. It is used to complete the format conversion, calibration, and matching of multi-source data, and establish a one-to-one correspondence between point cloud data, texture color data, and spatial coordinate data. The CSN model inference acceleration card uses an NVIDIA RTX A6000 professional graphics card with 48GB of built-in GDDR6 video memory. It is used to load the pre-trained grotto style knowledge base and run the CSN model to complete the extraction of statue features, identification of missing areas, generation of restoration schemes, and model reconstruction. The data storage array uses an 8-bay RAID5 disk array with a total capacity of 64TB for encrypted archiving of multi-source data, restoration models, and process data. The security encryption unit uses an encryption chip equipped with the national cryptographic SM4 algorithm to encrypt the entire process of data transmission and model storage, while recording all operation trajectories of the restoration process, so as to achieve traceability and auditability of the entire restoration process.

[0025] Display output module: The display output module includes an AR display unit, a UWB positioning base station, an interactive unit, and a projection display unit. Each unit communicates with the intelligent repair and processing module through a wired or wireless local area network.

[0026] In this embodiment, four UWB positioning base stations, all from Tsinghua Research Institute, are deployed at the four corners of the statue display area. These stations are used to establish a three-dimensional spatial coordinate system with the same source as the collected data, achieving millimeter-level real-time spatial positioning with a resolution of ±5mm. The AR display unit includes Microsoft HoloLens 2 AR glasses and a handheld tablet terminal running Android, used to overlay the restored virtual model with the original physical statue after spatial registration. The interactive unit includes a multi-touch screen, an Orbbec 3D gesture recognition camera, and an iFlytek voice interaction module, used to receive user interaction commands and perform operations such as switching between views before and after restoration, adjusting model transparency, and zooming in on details. The projection display unit includes four Panasonic PT-FRZ690C short-throw laser projectors, a fusion processor, and an infrared positioning system, used to project the restored virtual model onto the corresponding position on the original statue site through projection fusion, achieving a virtual-real fusion in-situ display for multiple viewers simultaneously.

[0027] like Figure 2 As shown, the virtual restoration and display method for grotto statues based on the above system specifically includes the following steps: S1. Data Acquisition: This step uses a data acquisition module to obtain the 3D point cloud data, surface texture and color data, and on-site spatial coordinate data of the target grotto statue. The specific operation is as follows: A 3D laser scanner is used to scan the statue from all directions. The scanning distance is controlled at 2-5m and the scanning resolution is set to 3mm / 10m. Scanning data from 12 stations are collected. After stitching and noise reduction, the complete 3D point cloud data of the statue is obtained. The point cloud data includes spatial point coordinates and normal vector information and is exported in PLY format. A high-definition texture acquisition camera was used to take multi-angle surround shots of the image under uniform lighting conditions. The shooting aperture was set to f / 8, ISO to 100, and shutter speed to 1 / 125s. A total of 120 high-definition images were acquired to obtain the surface texture and color data of the image. The images were in JPG format. At the same time, the lighting data of each shooting position was recorded by an ambient light compensation sensor for subsequent texture distortion correction and color calibration. Using a positioning and orientation device, and taking the benchmark control points inside the grotto as a reference, the on-site spatial coordinate data of the statues were collected. The coordinate information was then linked to the WGS84 coordinate system and the local independent coordinate system to complete the calibration and establish the spatial coordinate benchmark of the statues.

[0028] After data collection is completed, the local acquisition terminal performs format standardization conversion on all data and transmits it to the intelligent repair processing module via a gigabit wired network. The transmission process is encrypted throughout by the security encryption unit using the national cryptographic SM4 algorithm to prevent data leakage.

[0029] S2, Intelligent Virtual Repair: This step uses an intelligent restoration processing module to perform format conversion, calibration, and correlation matching on the collected multi-source data. Based on a pre-trained CSN artificial intelligence model, it identifies the incomplete areas of the statue and generates at least two differentiated restoration schemes that conform to the cultural relic restoration standards. After confirmation, the restored complete 3D digital model is output. The specific operations are as follows: Data preprocessing and correlation matching: After receiving multi-source data, the edge computing server performs noise reduction, simplification, and stitching optimization on the point cloud data, performs distortion correction and color consistency calibration on the texture image, and correlates the processed point cloud data, texture color data, and spatial coordinate data to establish a one-to-one correspondence between 3D spatial points, texture pixels, and spatial coordinates, and stores them in the data storage array. CSN Model Loading and Feature Extraction: The CSN model inference accelerator card starts the inference environment and loads the pre-trained grotto style knowledge base. The knowledge base contains more than 2,000 sets of complete grotto statues from different regions, eras, and schools in China, including Yungang, Longmen, and Maijishan, with three-dimensional structural features, texture and color rules, and case studies of damaged restoration. At the same time, it loads the weights of sub-models for point cloud processing, image segmentation, and 3D reconstruction. The CSN model extracts the geometric contours, structures, and edge features of damaged areas of the statues through the point cloud feature extraction network, and extracts the surface texture, color distribution, and material features of the statues through the texture feature extraction network. Combined with spatial coordinate information, it completes the spatial positioning of the features, achieving a precise correspondence between the features and the original spatial location of the statues. Incomplete Area Identification and Repair Scheme Generation: Based on extracted multi-source features, the CSN model automatically identifies three incomplete areas in the point cloud model of the statue through 3D semantic segmentation and incomplete detection algorithms. Combining texture and color information, it judges the boundaries, degree of damage, and original texture direction of the incomplete areas, and completes the accurate calibration of the incomplete areas in 3D space, generating a feature report of the incomplete areas. Based on the knowledge base of grotto styles and the Song Dynasty style characteristics of the statues, and following the principle of "minimal intervention and restoration to its original state" in cultural relic restoration, the model generates three differentiated repair schemes for the calibrated incomplete areas. The differences are reflected in the 3D geometric reconstruction, texture and color restoration, and proportional adaptation of the incomplete areas. The model performs 3D model fusion simulation on each repair scheme, checks the geometric continuity and texture transition, and eliminates schemes that do not meet the repair principles. Finally, it outputs three feasible repair schemes, each of which includes a 3D model with the repaired area, a texture layer, and a repair scheme description. Scheme Confirmation and Fine-tuning: The system simultaneously loads the original statue model and three restoration scheme models on a professional visual interface, supporting multi-view viewing, model comparison, detail magnification, and overlay comparison operations for cultural relic restoration experts. Experts, combining their professional experience, cultural relic protection principles, and the historical background of the statue from the Song Dynasty, select the optimal restoration scheme based on dimensions such as stylistic consistency, geometric and textural transitions, and conformity to the characteristics of different periods and schools. They then use the interface's 3D modeling fine-tuning tools to fine-tune the restoration area of ​​the optimal scheme, including adjusting geometric contours and curvature to achieve seamless geometric structure connections, modifying texture layers, and adjusting color saturation to achieve natural texture and color transitions. After completing the fine-tuning, the experts confirm the final restoration scheme. The system automatically saves the restoration model parameters, texture layers, and coordinate calibration information, and records all the experts' fine-tuning operations through a secure encryption unit, ensuring traceability of the restoration process. Complete 3D digital model output: Based on the confirmed final restoration plan, the system deeply integrates the fine-tuned restoration area with the point cloud, texture, and coordinate data of the original statue and reconstructs the 3D model. This completes the point cloud model meshing, precise texture mapping, and final coordinate binding, generating a high-precision 3D mesh model. The system converts the reconstructed model into multi-format compatible files, including the lightweight glTF format adapted for AR or projection display devices and the high-precision OBJ format that retains complete information. At the same time, metadata containing acquisition time, restoration experts, restoration plan, and coordinate information is added to the model, encrypted, and stored in the data storage array. Simultaneously, the lightweight model is sent to the display output module.

[0030] S3, Virtual-Real Fusion Display: This step uses the output module to establish a spatial positioning coordinate system on-site, spatially registering the restored 3D digital model with the original physical statue to achieve in-situ display through virtual-real fusion, and supports user interaction. This embodiment also provides AR display mode and projection display mode, as detailed below: AR Display Mode Deployment: On-site Spatial Positioning System Setup: Four UWB positioning base stations were deployed at the four corners of the statue display area. The spacing of the base stations was calibrated, the signal was debugged and networked. Based on the spatial coordinate information of the statue obtained in the acquisition phase, an on-site three-dimensional spatial coordinate system with the same source as the original data was established. The positioning error was controlled within ±5mm, laying a spatial benchmark for the accurate superposition of the virtual model. AR terminal device configuration and model adaptation: Import the output lightweight 3D digital model in glTF format into AR glasses and handheld tablet terminals to complete the hardware and software adaptation of the model and the device, optimize the model rendering parameters, and ensure that the device can load and display smoothly; at the same time, synchronize the on-site 3D spatial coordinate system parameters to all AR terminals to achieve the unification of device positioning and on-site spatial reference. Spatial registration between virtual model and physical statue: By combining the visual recognition module of the AR terminal with UWB positioning data, 12 preset feature points of the physical statue are captured. The captured physical feature points are accurately matched with the corresponding feature points of the virtual 3D model to complete the spatial registration. This ensures that the restored virtual model is completely aligned with the original damaged statue in terms of spatial position, scale, and angle, achieving seamless integration of virtual and real spaces. This ensures that when users view the virtual restored area from different perspectives, there is no visual offset between the virtual restored area and the physical statue. Interactive Function Debugging: Users can debug interactive functions in the AR terminal using gestures or voice commands to fix forward and backward vision issues. Figure 1 The system tested key switching, infinitely adjustable virtual model transparency, and zoom-in viewing of detailed images. It also tested the synchronous display effect across multiple terminals to ensure that the superimposed position and display status of the virtual model are consistent across different users' AR devices, thus completing the AR display deployment.

[0031] Projection display mode deployment: The repaired virtual model is precisely projected onto the incomplete position of the image after multi-channel fusion correction by a short-throw projector and multi-channel fusion processor. The infrared positioning system identifies the user's position and adjusts the projection angle and brightness to achieve in-situ display of virtual and real fusion. Users can interact through the touch screen to switch between different repair schemes and view the comparison effect before and after repair. It is suitable for public education scenarios where multiple people can watch at the same time.

[0032] S4. Data archiving and maintenance: The system encrypts and archives the operation trajectory, model parameters, and metadata of the entire restoration process, storing them in a data storage array for subsequent retrieval, modification, and reuse. The system automatically records user interaction data, compiles statistics on the statue details that users pay attention to, and frequently used interactive functions, and uses this information to optimize the display content and interaction logic. At the same time, it supports remote cultural relic restoration experts to annotate, guide, and adjust restoration plans through the system's remote access interface, enabling cross-regional collaborative research and training.

[0033] The above embodiments only illustrate preferred embodiments of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications, improvements, and substitutions without departing from the concept of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. An AI-based virtual restoration and display system for grotto statues, characterized in that, It includes a data acquisition module, an intelligent repair processing module, and a display and output module that are connected in sequence. The data acquisition module is used to acquire three-dimensional point cloud data, surface texture color data, and on-site spatial coordinate data of the grotto statues; The intelligent restoration processing module is a locally deployed edge processing unit used to receive multi-source data transmitted by the data acquisition module, complete data calibration and matching, identify the damaged areas of the statue based on the pre-trained CSN artificial intelligence model, generate a virtual restoration plan that conforms to the cultural relic restoration specifications, and output a complete three-dimensional digital model after restoration. The display output module is used to receive the complete three-dimensional digital model, complete the precise registration of the virtual model with the original physical statues of the grotto through the spatial positioning unit, output the display content in a virtual-real fusion manner, and support user interactive operation.

2. A virtual restoration and display system for grotto statues based on AI according to claim 1, characterized in that, The data acquisition module includes a 3D laser scanner, a high-definition texture acquisition camera, an ambient light compensation sensor, and a positioning and orientation device. The 3D laser scanner is used to acquire 3D point cloud data of the grotto statues, the high-definition texture acquisition camera is used to acquire surface texture and color data of the statues, the positioning and orientation device is used to acquire on-site spatial coordinate data of the statues, and the ambient light compensation sensor is used to acquire on-site lighting data to assist in texture calibration.

3. A virtual restoration and display system for grotto statues based on AI according to claim 1, characterized in that, The intelligent restoration processing module includes an edge computing server, a CSN model inference acceleration card, and a data storage array. The edge computing server is used to complete the format conversion, calibration, and matching of multi-source data, and establish a one-to-one correspondence between point cloud data, texture color data, and spatial coordinate data. The CSN model inference acceleration card is used to load a pre-trained grotto style knowledge base and run the CSN model to complete the extraction of statue features, identification of incomplete areas, generation of restoration schemes, and model reconstruction. The data storage array is used for encrypted archiving of multi-source data, restoration models, and process data.

4. A virtual restoration and display system for grotto statues based on AI according to claim 3, characterized in that, The intelligent repair processing module also includes a security encryption unit, which is used to encrypt the transmission process from the data acquisition module to the processing module and the storage process of the repair model, while recording the operation trajectory of the repair process to achieve full traceability.

5. A virtual restoration and display system for grotto statues based on AI according to claim 1, characterized in that, The display output module includes an AR display unit, a UWB positioning base station, and an interaction unit. The UWB positioning base station is used to establish a three-dimensional spatial coordinate system on site to achieve millimeter-level real-time spatial positioning. The AR display unit includes at least one of AR glasses and a handheld terminal, used to overlay and display the restored virtual model and the original physical statue after spatial registration. The interaction unit is used to receive user interaction commands and complete operations such as switching between views before and after restoration, adjusting model transparency, and zooming in to view details.

6. A virtual restoration and display system for grotto statues based on AI according to claim 1, characterized in that, The display output module also includes a projection display unit, which includes a short-throw projector, a fusion processor and an infrared positioning system. It is used to project the restored virtual model onto the corresponding position of the original site of the grotto statue through projection fusion, so as to realize the in-situ display of virtual and real fusion.

7. A method for virtual restoration and display of grotto statues based on AI, characterized in that, Includes the following steps: Step S1, Data Acquisition: Obtain the 3D point cloud data, surface texture and color data, and on-site spatial coordinate data of the target grotto statue; Step S2, Intelligent Virtual Restoration: After format conversion, calibration and correlation matching of the collected multi-source data, the incomplete areas of the statue are identified based on the pre-trained CSN artificial intelligence model, and at least two differentiated restoration schemes that conform to the cultural relic restoration specifications are generated. After confirmation, the complete three-dimensional digital model of the restoration is output. Step S3, Virtual-Real Fusion Display: Establish a spatial positioning coordinate system on site, spatially register the restored 3D digital model with the original physical statue, and complete the in-situ display in a virtual-real fusion manner, supporting user interaction.

8. The AI-based virtual restoration and display method for grotto statues according to claim 7, characterized in that, In step S2, the CSN artificial intelligence model is preloaded with a grotto style knowledge base, which contains the three-dimensional structural features, texture and color rules and restoration case library of grotto statues from different regions, eras and schools; the model completes the extraction of statue features through point cloud feature extraction network and texture feature extraction network, identifies the missing areas through three-dimensional semantic segmentation and defect detection algorithm, and generates a restoration plan following the principle of minimal intervention and restoration to its original state.

9. The AI-based virtual restoration and display method for grotto statues according to claim 7, characterized in that, In step S3, a three-dimensional spatial coordinate system with the same source as the collected data is established by deploying UWB positioning base stations in the display area, and the positioning error is controlled at the millimeter level; the feature points of the physical statue are captured by the AR terminal and matched with the corresponding feature points of the virtual model to complete spatial registration, so that the virtual model and the physical statue are completely aligned in terms of spatial position, scale and angle.

10. A method for virtual restoration and display of grotto statues based on AI according to claim 7, characterized in that, It also includes step S4, data archiving and maintenance: encrypting and archiving the operation trajectory, model parameters, and metadata of the entire repair process, recording user interaction data to optimize the displayed content, and supporting remote experts to provide annotation guidance and adjust the solution through the system.

Images