Method and system for implementing a three-dimensional optical system

Abstract

The present application relates to the field of three-dimensional optical technology, and more particularly to a method and system for implementing a three-dimensional optical system. The method comprises the following steps: scanning a target object by an optical three-dimensional scanning device to obtain three-dimensional geometric data; collecting surface texture data of the target object to obtain the surface texture data; constructing a three-dimensional model using the three-dimensional geometric data and the surface texture data to obtain an initial three-dimensional digital model; identifying and marking a damaged area of the initial three-dimensional digital model to obtain three-dimensional model damaged area data; and generating three-dimensional model damaged prediction data by simulating restoration according to the three-dimensional model damaged area data. The present application can efficiently generate a high-quality three-dimensional digital model and provide excellent user interaction experience by collecting data, accurate modeling, automatic damage prediction and repair through a three-dimensional optical system, and is suitable for the field of cultural relic restoration.

Description

Technical Field

[0001] The present invention relates to the field of three-dimensional optical technology, and in particular to a method and system for realizing a three-dimensional optical system. Background Art

[0002] Current three-dimensional optical systems often exhibit significant limitations when dealing with irregular shapes or fine structures. For example, when scanning complex cultural heritage artifacts (such as broken pottery, finely carved artifacts, or irregularly shaped stone tools), the system is prone to scanning blind spots or geometric distortion, resulting in insufficient accuracy in the generated three-dimensional model. In addition, for target objects with transparent, reflective, or high-contrast surfaces, traditional scanning equipment has difficulty accurately capturing their details. The three-dimensional modeling systems on the market can usually capture the complete shape of the object, but for areas that are damaged or missing, existing systems cannot automatically and effectively complete the reconstruction. Existing methods mostly rely on manual inference and completion, which is not only time-consuming and labor-intensive, but also easily leads to models that are not realistic or accurate enough. In addition, for applications in the field of cultural heritage protection, how to perform reasonable and traceable repairs based on existing model features remains a technical challenge. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide a method and system for implementing a three-dimensional optical system to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for implementing a three-dimensional optical system includes the following steps:

[0005] Step S1: Scanning a target object with an optical 3D scanning device to obtain 3D geometric data; collecting surface texture data of the target object to obtain surface texture data; constructing a 3D model using the 3D geometric data and the surface texture data to obtain an initial 3D digital model;

[0006] Step S2: Identify and mark defective areas of the initial three-dimensional digital model to obtain defective area data of the three-dimensional model; perform simulation restoration based on the defective area data of the three-dimensional model to generate defect prediction data of the three-dimensional model;

[0007] Step S3: restoring and reconstructing the initial three-dimensional digital model according to the three-dimensional model defect prediction data to obtain a reconstructed three-dimensional digital model;

[0008] Step S4: Optimizing the surface smoothness of the reconstructed three-dimensional digital model and performing detail enhancement to obtain an optimized three-dimensional digital model;

[0009] Step S5: Visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

[0010] The present invention uses an optical 3D scanning device to collect geometric and surface texture data of a target object, enabling the precise construction of an initial 3D digital model. This approach not only improves modeling accuracy but also captures rich surface details. By identifying and marking defective areas in the 3D model and then simulating the restoration of these areas, the defective areas can be accurately predicted, generating high-quality defect prediction data for the 3D model. This provides a reliable basis for subsequent restoration and reconstruction. Using the defect prediction data to restore and reconstruct the initial 3D model, the missing portions of the target object can be restored, improving the model's integrity and authenticity. By optimizing surface smoothness and enhancing details, the visual quality of the 3D model can be effectively enhanced, making it more detailed and realistic, meeting the model quality requirements of various application scenarios. The optimized 3D digital model can be displayed and interacted with in real time, supporting user interaction with the 3D model, enhancing the model's usability and user experience, and making it suitable for a variety of display platforms, including virtual reality and augmented reality. Multi-platform adaptation ensures smooth display of the 3D model on a variety of devices, improving the system's flexibility and universality.

[0011] Preferably, the present invention further provides a three-dimensional optical system for executing the above-mentioned method for realizing the three-dimensional optical system, wherein the three-dimensional optical system comprises:

[0012] The 3D model initial construction module is used to scan the target object through an optical 3D scanning device to obtain 3D geometric data; collect surface texture data of the target object to obtain surface texture data; and construct a 3D model using the 3D geometric data and surface texture data to obtain an initial 3D digital model.

[0013] The defect area identification and prediction module is used to identify and mark the defect area of ​​the initial three-dimensional digital model to obtain the defect area data of the three-dimensional model; simulate and restore the defect area data of the three-dimensional model to generate the defect prediction data of the three-dimensional model;

[0014] A three-dimensional model restoration and reconstruction module is used to restore and reconstruct the initial three-dimensional digital model according to the three-dimensional model defect prediction data to obtain a reconstructed three-dimensional digital model;

[0015] The model optimization and detail enhancement module is used to optimize the surface smoothness of the reconstructed three-dimensional digital model and enhance the details to obtain an optimized three-dimensional digital model;

[0016] The three-dimensional model visualization display module is used to visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

[0017] The present invention uses an optical 3D scanning device to scan a target object. The system accurately acquires 3D geometric and surface texture data, ensuring high-precision construction of the initial 3D digital model. This lays a solid foundation for subsequent modeling, restoration, and optimization. By identifying and marking defective areas in the initial 3D model and performing predictive restoration based on the defective area data, the system can accurately identify the defective parts in the model and simulate the features of the defective areas. This function can significantly improve the accuracy and efficiency of model restoration. Utilizing the defect prediction data, the system accurately restores and reconstructs the initial 3D digital model, restoring the morphology and details of the defective areas, ensuring the final 3D digital model is more complete and realistic. By optimizing the surface smoothness and enhancing the details of the reconstructed 3D model, the system improves the model's visual quality and detail, further enhancing the accuracy and realism of the 3D model and making it more expressive during presentation. The system provides real-time display and interactive features, allowing users to intuitively and flexibly view and manipulate the 3D model. This feature allows users to not only deeply observe the model's details but also interact from different perspectives, enhancing the user experience and the flexibility of model presentation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0019] Figure 1 Schematic diagram of the steps of the method for implementing the three-dimensional optical system of the present invention;

[0020] Figure 2 for Figure 1 Detailed step flow diagram of step S1;

[0021] Figure 3 for Figure 1 Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0022] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0023] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0024] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0025] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for realizing a three-dimensional optical system, the method comprising the following steps:

[0026] Step S1: Scanning a target object with an optical 3D scanning device to obtain 3D geometric data; collecting surface texture data of the target object to obtain surface texture data; constructing a 3D model using the 3D geometric data and the surface texture data to obtain an initial 3D digital model;

[0027] Step S2: Identify and mark defective areas of the initial three-dimensional digital model to obtain defective area data of the three-dimensional model; perform simulation restoration based on the defective area data of the three-dimensional model to generate defect prediction data of the three-dimensional model;

[0028] Step S3: restoring and reconstructing the initial three-dimensional digital model according to the three-dimensional model defect prediction data to obtain a reconstructed three-dimensional digital model;

[0029] Step S4: Optimizing the surface smoothness of the reconstructed three-dimensional digital model and performing detail enhancement to obtain an optimized three-dimensional digital model;

[0030] Step S5: Visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

[0031] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a flow chart showing the steps of a method for implementing a three-dimensional optical system according to the present invention. In this example, the method for implementing a three-dimensional optical system includes the following steps:

[0032] Step S1: Scanning a target object with an optical 3D scanning device to obtain 3D geometric data; collecting surface texture data of the target object to obtain surface texture data; constructing a 3D model using the 3D geometric data and the surface texture data to obtain an initial 3D digital model;

[0033] This embodiment of the present invention utilizes a structured light scanner, which consists of a projector, a camera, and a computer control system. The projector projects a series of light strips onto the surface of a target object and, based on the deformation pattern captured by the camera, calculates the spatial coordinates of each point on the surface. A structured light scanner projects a known pattern of light strips onto the surface, uses the camera to detect the deformation of the reflected light strips, and uses triangulation to calculate the three-dimensional coordinates of each point. By scanning from multiple viewpoints, the system can accurately obtain three-dimensional geometric information about the object's surface. By scanning the target object with a structured light scanner, the system can obtain three-dimensional geometric data of the target object's surface. The appropriate scanning viewpoint and angle are selected based on the size and shape of the target object. During the scanning process, the object needs to be rotated or the scanner position adjusted multiple times to ensure coverage of all angles and avoid shadows or obstructions. The appropriate scanning accuracy setting is selected based on the complexity of the surface details. For fine textures or tiny features, the system can select a higher resolution for scanning to ensure accuracy that meets subsequent modeling requirements. After scanning, the resulting three-dimensional geometric data includes information such as the spatial coordinates and normal vectors of each point on the object's surface. While acquiring 3D geometric data, surface texture data of the target object must be collected for texture mapping during 3D modeling. A high-resolution camera is used to scan the texture of the target object, typically using a directional light source to avoid shadows and clearly capture surface details. During this process, the lighting angle and camera position must be optimized to obtain high-quality images of the object's surface texture. The captured images are processed, including color correction, denoising, and texture stitching. UV mapping is then used to accurately align the surface texture image with the 3D geometric data, generating complete 3D surface texture data. Surface texture data includes the object's surface color, gloss, and reflectance. The 3D geometric and surface texture data are then used to construct a 3D model. The captured 3D geometric data is processed, including noise removal and repairing any holes or overlaps in the scan. A triangulated meshing algorithm is used to convert the data into a continuous 3D mesh model. The surface texture data is mapped onto the processed 3D mesh model to ensure proper registration of the texture with the geometric model. Texture mapping algorithms are used to enhance texture detail and realism. Finally, the 3D model is constructed by fusing the 3D geometric and surface texture data, resulting in an initial 3D digital model.

[0034] Step S2: Identify and mark defective areas of the initial three-dimensional digital model to obtain defective area data of the three-dimensional model; perform simulation restoration based on the defective area data of the three-dimensional model to generate defect prediction data of the three-dimensional model;

[0035] Embodiments of the present invention identify defective areas in a model. Defective areas typically appear as incomplete or missing portions of a surface in the scanned data, such as holes or discontinuities caused by limited scanning angles, reflections from the object's surface, or damage to the object itself. A mesh analysis algorithm is used to process the 3D model. This involves calculating the distribution of the model's mesh faces and vertices. For each mesh face, the system calculates the connectivity of its adjacent faces, identifying areas of voids or damage. Common algorithms, such as point cloud-based data integration and connectivity analysis, can be used to identify discontinuous areas on the object's surface. By performing in-depth analysis of the mesh model, machine learning algorithms are used to automatically identify areas that differ from the surrounding area. Once defective areas are identified, the system marks them and assigns them different colors or shapes to facilitate subsequent processing. The marked data is saved as a defective area dataset, which serves as the basis for subsequent repair and restoration. The resulting defective area data for the 3D model includes the spatial location, shape, size, and missing geometric information of the defective area. Interpolation or fitting algorithms are used to predict the defective area. In addition to geometric shape prediction, texture reconstruction is equally important. Texture loss is usually caused by the failure to completely cover the surface of the target object during the scanning process. To this end, the texture of the missing area can be inferred based on the texture data of the surrounding area through the interpolation technology of texture mapping. In the defective area of ​​the model, an interpolation algorithm is used to extract color information from the known area to generate a smoothly transitioned texture. Based on the texture features of the surrounding area, texture synthesis technology is used to restore the missing texture information. Through the above-mentioned geometric and texture reconstruction, simulated restoration data of the defective area of ​​the three-dimensional model is obtained. This data includes the reconstructed geometric shape and texture details of the defective area, as well as the restoration prediction of the object's integrity. The three-dimensional model defect prediction data is obtained, including the restored geometric information, predicted texture and surface details of the defective area.

[0036] Step S3: restoring and reconstructing the initial three-dimensional digital model according to the three-dimensional model defect prediction data to obtain a reconstructed three-dimensional digital model;

[0037] In this embodiment of the present invention, 3D model defect prediction data is loaded into the system. The system then integrates this data with the defective areas in the initial 3D digital model, preparing for geometric reconstruction. The defective area is reconstructed based on the geometric features of the surrounding area using algorithms such as nearest neighbor interpolation, surface fitting, or least squares. For example, if the defective area is a large surface, the system can perform surface fitting using the surrounding intact surface to create a smooth, seamless surface. For complex defective areas, geometric repair can be performed using mesh-based repair algorithms. These algorithms extend the mesh information of the surrounding area to the defective area, reconstructing the geometric surface of the missing portion and ensuring a smooth transition with the surrounding mesh. For smaller defective areas, the system can fill in the geometric data of the defective area using methods such as boundary completion, convex hull algorithms, and Laplace smoothing. The reconstructed geometry is compared with the original 3D model to verify the effectiveness of the repair of the defective area. This process can be verified by checking accuracy and comparing the differences between the models before and after reconstruction. If necessary, the system can perform local optimization and detail adjustments. Through geometric reconstruction, a reconstructed 3D digital model with complete geometric structure is obtained, and all defective areas have been effectively repaired. Based on the three-dimensional model defect prediction data, the system maps the texture image to the reconstructed geometric surface. For large-scale defective areas, the system fills in the missing texture details through texture interpolation or texture image generation technology. In order to make the texture blend seamlessly with the surrounding area, the system can use gradient interpolation or texture patching algorithms to ensure that the color difference, reflectivity and glossiness of the texture transition smoothly between the defective area and the surrounding area to avoid visible splicing marks. For smaller defective areas, the system can also perform texture restoration based on the local features of the surrounding texture area. By extracting the color, details and texture direction of similar textures, the system can accurately fill in the missing areas and restore the details of the object. For particularly complex texture reconstruction problems, deep learning technology can be used to generate more natural texture repair effects. Through texture reconstruction, a reconstructed three-dimensional digital model with complete texture details is obtained.

[0038] Step S4: Optimizing the surface smoothness of the reconstructed three-dimensional digital model and performing detail enhancement to obtain an optimized three-dimensional digital model;

[0039] This embodiment of the present invention optimizes the mesh of a 3D model using a Laplace smoothing algorithm to remove noise and irregular geometric surfaces. Based on the neighborhood relationships of mesh vertices, this algorithm adjusts the position of each vertex to ensure a smoother position and consistency with surrounding vertices, reducing sharp edges and protruding points. Layered smoothing is performed based on the model's local regions and global morphology. While preserving more detail in localized areas, the overall surface is smoothed more strongly to reduce the unnaturalness of transitional regions. Incorporating an adaptive smoothing algorithm, the system intelligently selects the smoothing intensity based on the specific conditions of the defective and reconstructed regions. For areas with surface irregularities, the smoothing algorithm moderately reduces their impact, while enhancing smoothing for flat areas or seams to ensure seamless transitions. This smoothing optimization results in a 3D digital model with smoother surfaces and more natural transitions. After smoothing the surface of the 3D digital model, the next step is to enhance the model's details, particularly the realism of surface features and textures. A multi-scale detail enhancement algorithm is employed to restore and enhance minute surface details. This process primarily leverages surface concavity and convexity information to enhance small-scale variations in the model's surface, improving the model's realism. For example, for archaeological artifacts, it is necessary to enhance tiny cracks or wear marks on the surface. During the texture enhancement process, the system refines and adjusts the texture based on a combination of texture data and geometric information. Through multi-level refinement technology, the texture details of the object surface are carefully restored, such as improving the surface gloss, color gradient, texture density, etc., making the model more visually attractive. Combined with image enhancement technology, especially the enhancement method based on deep learning, the texture performance of the model can be further refined. Through convolutional neural networks (CNN) or generative adversarial networks (GAN), the system can generate richer and more natural surface textures, especially on some more complex object surfaces, which can effectively fill in the missing details and improve the naturalness of the texture. After detail enhancement, an optimized three-dimensional digital model is obtained.

[0040] Step S5: Visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

[0041] The embodiments of the present invention use real-time rendering technology to render 3D models, presenting realistic lighting effects and texture details. By setting different light source directions, intensities, and material properties, the model's surface, color, reflection, and other characteristics are rendered more realistically. The system supports multiple perspective rotations, allowing users to view the 3D model from different angles. By dragging the mouse, touching, or using keyboard shortcuts, users can freely adjust the perspective to view various sides and details of the model. To improve performance, level of detail technology is used to reduce the model's detail level when viewed from a distance and increase it when viewed up close. This ensures a high-quality display while optimizing loading speed. Through the visualization process, high-quality 3D digital model display data is generated. This data supports display on various display devices, adapting to different resolutions and performance requirements. User interaction is supported through click, drag, zoom, and other operations. For example, users can click on the model surface to view detailed information or drag the model to rotate and adjust the viewing angle. Through real-time display and interaction, users enjoy a more intuitive and immersive experience, allowing them to freely view and interact with the 3D digital model.

[0042] The present invention uses an optical 3D scanning device to collect geometric and surface texture data of a target object, enabling the precise construction of an initial 3D digital model. This approach not only improves modeling accuracy but also captures rich surface details. By identifying and marking defective areas in the 3D model and then simulating the restoration of these areas, the defective areas can be accurately predicted, generating high-quality defect prediction data for the 3D model. This provides a reliable basis for subsequent restoration and reconstruction. Using the defect prediction data to restore and reconstruct the initial 3D model, the missing portions of the target object can be restored, improving the model's integrity and authenticity. By optimizing surface smoothness and enhancing details, the visual quality of the 3D model can be effectively enhanced, making it more detailed and realistic, meeting the model quality requirements of various application scenarios. The optimized 3D digital model can be displayed and interacted with in real time, supporting user interaction with the 3D model, enhancing the model's usability and user experience, and making it suitable for a variety of display platforms, including virtual reality and augmented reality. Multi-platform adaptation ensures smooth display of the 3D model on a variety of devices, improving the system's flexibility and universality.

[0043] Preferably, step S1 includes the following steps:

[0044] Step S11: Scanning the target object at multiple angles using an optical 3D scanning device, and collecting surface geometric information to obtain 3D geometric data;

[0045] Step S12: performing noise reduction processing on the three-dimensional geometric data, and aligning and fusing the noise reduction processing results to obtain high-precision three-dimensional data;

[0046] Step S13: collecting surface texture information of the target object from multiple angles using a camera to obtain surface texture data;

[0047] Step S14: Mapping the surface texture data to high-precision three-dimensional data, and constructing a three-dimensional model to obtain an initial three-dimensional digital model.

[0048] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in the embodiment of the present invention, step S1 includes the following steps:

[0049] Step S11: Scanning the target object at multiple angles using an optical 3D scanning device, and collecting surface geometric information to obtain 3D geometric data;

[0050] The embodiment of the present invention uses a high-resolution structured light scanner to obtain object surface data from multiple different angles through multiple scans. Laser scanning can quickly capture point cloud data on the surface of an object, while structured light scanning is suitable for high-precision scanning of complex curved surfaces and small objects. The object is placed in the scanning area, and the optical device performs an all-round scan around the target object to ensure that comprehensive geometric data is collected from different angles. During the scanning process, the device uses structured light projection technology to obtain high-density point cloud data. The data collected by each scan includes the coordinates, shape and surface features of the object surface, generating a three-dimensional point cloud. By integrating the point cloud data scanned from multiple angles, complete three-dimensional geometric data is ultimately obtained.

[0051] Step S12: performing noise reduction processing on the three-dimensional geometric data, and aligning and fusing the noise reduction processing results to obtain high-precision three-dimensional data;

[0052] The embodiment of the present invention uses a point cloud-based noise reduction algorithm to remove noise points generated by environmental interference or equipment errors during the scanning process. This step removes abnormal data points that do not conform to geometric laws by analyzing the local density and distance of the point cloud data. The iterative nearest point algorithm is used to accurately align the point cloud data obtained from multiple scanning angles. Through feature point matching or iterative optimization based on minimum error, the data points under different perspectives are accurately connected. The aligned data is fused and the scanning information from different angles is combined to generate a unified high-precision three-dimensional point cloud model. During the fusion process, the weighted average or error minimization strategy is used to ensure the accuracy of each point cloud data. High-precision three-dimensional geometric data is obtained through noise reduction processing, alignment and fusion.

[0053] Step S13: collecting surface texture information of the target object from multiple angles using a camera to obtain surface texture data;

[0054] The embodiments of the present invention use a high-resolution digital camera or a dedicated texture acquisition device to ensure that the subtle texture and color details of the target object's surface can be captured. The device supports shooting from different angles and multiple directions to cover all surfaces of the object. The object is fixed in the scanning area, and the camera device is rotated or moved to ensure that all surfaces of the target object are covered. The camera device captures a series of images at different angles to obtain high-quality texture data. In order to avoid the influence of lighting changes on texture data, the system uses a unified light source or soft light technology to ensure that the collected texture data is clear and free of shadow interference. Through this process, a series of high-precision surface texture data is obtained.

[0055] Step S14: Mapping the surface texture data to high-precision three-dimensional data, and constructing a three-dimensional model to obtain an initial three-dimensional digital model.

[0056] Embodiments of the present invention use texture mapping technology to accurately match surface texture data obtained from multi-angle camera equipment with high-precision 3D geometric data. Specifically, the pixels of the texture image are mapped to the corresponding surface of the 3D model using the spatial relationship between the camera calibration data and the 3D point cloud data. A UV coordinate system is generated on the 3D geometric data, and the UV coordinates serve as the basis for texture mapping. An unfolding algorithm divides the object surface into several regions, and the texture image is accurately mapped onto these regions, ensuring accurate texture alignment. By fusing the 3D geometric data and texture information, an initial 3D digital model is generated.

[0057] The present invention uses an optical three-dimensional scanning device to scan the target object from multiple angles, and combined with the collection of surface geometric information, it can obtain high-precision three-dimensional geometric data of the object. This method can comprehensively and accurately capture the complex geometric shape of the object, ensuring the accuracy and details of the model. The three-dimensional geometric data is subjected to noise reduction processing, and through alignment and fusion operations, the noise in the data can be effectively removed to improve the data quality. This step ensures that the final generated three-dimensional data is more realistic and clear, reducing the complexity of post-processing. By collecting surface texture information from multiple angles through a camera device, the surface details of the object can be obtained with high quality. This process enhances the texture effect of the model and provides rich surface data for subsequent model optimization and repair. Mapping the surface texture data to high-precision three-dimensional data and constructing a three-dimensional model helps to generate a complete digital model that includes the object's geometric shape and surface texture. This integrated method improves the efficiency and accuracy of three-dimensional modeling. Through the above steps, the initial three-dimensional digital model can be generated quickly and accurately, providing high-quality basic data for subsequent processing, optimization and application.

[0058] Preferably, step S2 includes the following steps:

[0059] Step S21: meshing the initial three-dimensional digital model to obtain meshed data of the initial three-dimensional model;

[0060] Step S22: identifying and marking surface defect areas of the initial three-dimensional digital model based on the initial three-dimensional model gridding data to obtain defect area data of the three-dimensional model;

[0061] Step S23: adjusting the defect boundary precision according to the defect area data of the 3D model to obtain accurate defect area data of the 3D model;

[0062] Step S24: extracting geometric and texture features from the defective area data of the precise three-dimensional model to obtain feature data of the defective area of ​​the three-dimensional model;

[0063] Step S25: Using the feature data of the defective area of ​​the three-dimensional model, a defective area restoration simulation is performed on the initial three-dimensional digital model to generate three-dimensional model defect prediction data.

[0064] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S2 in the embodiment of the present invention, step S2 includes the following steps:

[0065] Step S21: meshing the initial three-dimensional digital model to obtain meshed data of the initial three-dimensional model;

[0066] The embodiment of the present invention adopts three-dimensional meshing technology to convert the point cloud data of the initial three-dimensional digital model into a mesh structure composed of multiple polygonal units. The meshing process includes connecting adjacent points to generate facets according to the geometric structure of the model to form a structured three-dimensional mesh model. The density of the mesh is adjusted according to the complexity of the object surface. For complex areas, a higher density mesh is used to ensure accuracy; for flat areas, a lower density mesh is used to improve computing efficiency. The density of the mesh should match the surface details of the object to ensure the accuracy of the model and the feasibility of subsequent processing. After meshing, the three-dimensional model mesh data obtained includes vertex coordinates, boundary information and a mesh structure composed of facets. These data provide the basis for subsequent defect area identification, boundary adjustment and feature extraction. After meshing, the meshed data of the initial three-dimensional model is obtained.

[0067] Step S22: identifying and marking surface defect areas of the initial three-dimensional digital model based on the initial three-dimensional model gridding data to obtain defect area data of the three-dimensional model;

[0068] The embodiment of the present invention analyzes the gridded data of the initial three-dimensional digital model and applies algorithms (such as point cloud-based density analysis, geometric difference detection, or texture consistency analysis) to identify surface defect areas. Defective areas usually appear as discontinuous, missing, or abnormal areas on the surface, which are caused by occlusion during the scanning process, object damage, or incomplete texture information. The identified defective areas are marked using a marking algorithm. Each defective area is assigned a unique label and clearly identified in the three-dimensional model. This process involves refining the boundaries and shapes of the defective areas to ensure the accuracy of the marking. Through defective area identification and marking, three-dimensional model defective area data containing detailed defect location and shape information is generated.

[0069] Step S23: adjusting the defect boundary precision according to the defect area data of the 3D model to obtain accurate defect area data of the 3D model;

[0070] The embodiment of the present invention identifies the boundary position of the defective area by analyzing the data of the defective area of ​​the three-dimensional model. Then, a boundary smoothing algorithm (such as curve fitting, least squares method or weighted average method) is used to fine-tune the boundary of the defective area to remove errors or noise. At this time, the adjusted boundary can more accurately reflect the actual missing part of the object surface. Based on the boundary adjustment results, combined with the geometric features of the adjacent areas, the shape of the defective area is further optimized. Local optimization is performed on each point of the boundary to ensure that the boundary of the defective area transitions naturally with other parts of the object to avoid obvious faults or unnatural curves. The adjusted defect boundary data forms accurate three-dimensional model defective area data.

[0071] Step S24: extracting geometric and texture features from the defective area data of the precise three-dimensional model to obtain feature data of the defective area of ​​the three-dimensional model;

[0072] The embodiment of the present invention extracts the shape features of the defective area by analyzing the data of the defective area of ​​the precise three-dimensional model. The geometric contour of the defective area can be modeled using an edge detection algorithm or a surface smoothness calculation method to capture the spatial structural features of the missing part. The texture analysis algorithm is used to extract texture features from the adjacent surfaces of the defective area. These features may include surface color difference, pattern arrangement, surface details, etc., which are used to supplement the texture in subsequent simulations. The extracted geometric and texture features are combined to form a comprehensive set of defective area feature data. These data include the spatial morphology, surface texture of the defective part, and its transition features with the surrounding intact areas. By extracting geometric and texture features, detailed three-dimensional model defective area feature data is generated.

[0073] Step S25: Using the feature data of the defective area of ​​the three-dimensional model, a defective area restoration simulation is performed on the initial three-dimensional digital model to generate three-dimensional model defect prediction data.

[0074] The embodiments of the present invention use shape matching, texture filling, and interpolation algorithms to restore and simulate the defective area based on the geometric and texture features in the feature data of the defective area of ​​the 3D model. By analyzing the spatial structure and surface texture of the defective area, an appropriate algorithm is selected to fill the defective area. Based on the known complete area data, similarity matching methods or inverse modeling techniques are used to generate a reasonable geometric form in the defective area. The defective part is filled in based on the texture features of the surrounding area to form complete prediction data. After the simulation and restoration process, the 3D model defect prediction data is generated.

[0075] The present invention can convert complex three-dimensional data into more operational grid data by gridding the initial three-dimensional digital model, which is convenient for subsequent analysis and processing. Gridding provides structured basic data for further identification and repair of defective areas. Defective area identification and marking based on gridded data can effectively locate the defective parts in the three-dimensional model. This process provides a clear target area for subsequent defect repair and optimization through precise marking. The boundary accuracy of the defective area is adjusted to ensure the accuracy of the defective area data. This step improves the accuracy of the repair, avoids errors during repair, and ensures that the final repair result is more accurate. By extracting the geometric and texture features of the defective area, more detailed defective area information can be obtained. This provides the necessary detailed data for subsequent defect repair and model reconstruction, making the repair process more efficient and realistic. Using the defective area feature data for restoration simulation to generate three-dimensional model defect prediction data, the missing parts can be accurately predicted and supplemented, ensuring the integrity and realism of the three-dimensional model after repair.

[0076] Preferably, step S25 includes the following steps:

[0077] Step S251: constructing a preliminary interpolation model of the defective region based on the feature data of the defective region of the three-dimensional model to obtain a preliminary interpolation model of the defective region;

[0078] The embodiment of the present invention uses the characteristic data of the defective area of ​​the three-dimensional model as the input for constructing the interpolation model. Using these characteristic data, combined with the neighborhood information, the spatial structure and geometric shape of the defective area are determined. An appropriate interpolation algorithm, such as polynomial interpolation, radial basis function interpolation or spline interpolation, is selected to perform preliminary reconstruction based on the geometric characteristics of the defective area. Through the interpolation algorithm, a preliminary geometric shape model of the defective area is generated. By modeling the spatial relationship between the known part and the defective part, a preliminary defective area interpolation model is constructed, which can reasonably estimate the shape and structure of the defective part and ensure geometric consistency with the surrounding area. The preliminary defective area interpolation model obtained by this step.

[0079] Step S252: optimizing the defect area geometry of the preliminary defect area interpolation model using the defect area data of the accurate three-dimensional model to obtain optimized defect area geometry data;

[0080] The embodiment of the present invention takes the preliminary defect area interpolation model and the precise three-dimensional model defect area data as input. The precise three-dimensional model data contains more detailed information, including more precise boundaries, curvatures and texture changes. A shape optimization algorithm or a constrained optimization method, such as a deformable mesh algorithm, shape matching technology or a rigid / non-rigid registration method, is used to optimize the geometric morphology. The goal is to make the preliminary interpolation model more consistent with the geometric morphology of the surrounding area and to correct errors or irregularities in the interpolation model. Through constrained optimization, the geometric morphology of the defect area is gradually adjusted to conform to the geometric features of the precise model. Special attention is paid to the smooth transition of the boundary between the defect area and the adjacent area to avoid morphological breaks or abrupt changes. Ultimately, the optimized geometric morphology data of the defect area is obtained.

[0081] Step S253: reconstructing the texture of the defective area according to the feature data of the defective area of ​​the three-dimensional model and the optimized geometric data of the defective area to obtain texture restoration data of the defective area;

[0082] The embodiment of the present invention inputs optimized geometric data of the defective area and characteristic data of the defective area of ​​the three-dimensional model. Texture mapping technology is used to accurately apply the texture data to the geometric surface of the defective area. Specific methods may include texture reconstruction based on UV mapping, or using a texture synthesis algorithm to extract sample textures from the surrounding area to fill the defective area. During the texture reconstruction process, taking into account the transition between the texture of the defective area and the surrounding area, a gradient smoothing process is used to avoid visual abruptness. In addition, a multi-scale texture generation method can also be used to enhance texture details and ensure a natural transition of textures. Finally, texture restoration data of the defective area is obtained.

[0083] Step S254: Comprehensively verifying the optimized defect area geometry data and the defect area texture restoration data to obtain defect area restoration data;

[0084] The embodiment of the present invention utilizes a geometric accuracy assessment method for a three-dimensional model to verify the optimized geometric morphological data of the defective area and check whether its boundary smoothly connects with the surrounding structure of the defective area. The continuity and accuracy of the morphology are verified by a curvature analysis tool. The texture restoration data of the defective area is compared and analyzed with the original texture of the initial three-dimensional model, with a focus on checking the color, texture details and lighting consistency. Image processing technology is used to align and evaluate the textures to ensure that there are no obvious boundary differences. Based on the geometric and texture verification results, the existing deviations are automatically or manually adjusted. The geometry and texture mapping are readjusted based on the characteristics of the surrounding area to ultimately generate accurate defective area restoration data.

[0085] Step S255: Fusing the restored data of the defect area with the initial three-dimensional digital model to obtain a defect prediction three-dimensional model;

[0086] The embodiment of the present invention determines the boundary of the defective area corresponding to the initial three-dimensional digital model, and ensures that it is completely aligned with the geometric boundary of the defective area restoration data. Using a three-dimensional model processing algorithm, the geometric form of the defective area restoration data and the geometric data of the initial model are spliced ​​and smoothed to ensure a natural transition at the joint. The curvature adjustment algorithm is used to eliminate the abrupt boundary phenomenon. The texture restoration data of the defective area is mapped to the model surface and aligned with the surrounding texture data. Texture blending technology is used to optimize color and texture details to ensure uniform texture transitions. The fused three-dimensional model is checked for overall consistency, including geometric and texture consistency. Global adjustments are made when necessary to ensure model integrity, and ultimately a defect prediction three-dimensional model is generated.

[0087] Step S256: Utilizing the defect prediction three-dimensional model to perform defect area restoration simulation to obtain three-dimensional model defect prediction data.

[0088] The embodiments of the present invention determine the restoration simulation parameters, including physical properties, geometric accuracy requirements, and texture feature details, based on the actual application scenario of the target object. Using geometric and texture fitting algorithms, the feature data of the defective area in the defect prediction 3D model is integrated with the complete model to simulate and restore the final shape of the defective area. A stepwise iterative method is used to optimize the consistency of the model. Error analysis is performed on the restored 3D model defect prediction data to ensure a high degree of match with the geometric form and texture details of the original data. The 3D model defect prediction data containing the complete geometric form and surface texture is output.

[0089] The present invention can effectively fill the defective parts in the three-dimensional model while maintaining consistency with the surrounding areas by constructing a preliminary interpolation model of the defective area and optimizing the geometric morphology. This ensures that the repaired three-dimensional model is more natural in structure and avoids errors and inconsistencies in the manual repair process. Texture reconstruction using the defective area feature data and the optimized geometric morphology data helps to restore the visual effect of the defective area, ensuring that the three-dimensional model is not only repaired and perfected in geometry, but also achieves a high sense of reality in surface texture, making the final model more perfect in visual effect. By comprehensively verifying the geometry and texture of the defective area, the accuracy and consistency of all repair data are ensured, avoiding the quality of the final model affected by data conflicts or errors. By fusing the defective area restoration data with the initial three-dimensional digital model, it is possible to achieve accurate restoration of the defective area and ultimately generate a high-quality defect prediction three-dimensional model. Restoration simulation based on the defect prediction three-dimensional model can accurately predict and supplement the missing parts, providing a scientific basis for subsequent repair work. This helps to improve repair efficiency and reduce the loss of original data during the repair process.

[0090] Preferably, step S3 includes the following steps:

[0091] Step S31: geometrically aligning the three-dimensional model defect prediction data with the initial three-dimensional digital model to obtain a three-dimensional digital defect prediction combined model;

[0092] This embodiment of the present invention extracts significant geometric feature points from the 3D model defect prediction data and the initial 3D digital model, ensuring high-precision alignment of the base points. An iterative closest point alignment algorithm is used to adjust the pose of the two sets of data for precise spatial alignment. A global error analysis is performed on the aligned model, residual distribution is calculated, and areas with deviations are adjusted to ensure consistency in geometry and spatial position. The two aligned data sets are then fused to generate a complete 3D digital defect prediction combined model.

[0093] Step S32: generating a geometric grid of the defect area according to the three-dimensional digital defect prediction combined model, and filling the defect area to generate a three-dimensional digital model with a continuous grid structure;

[0094] The embodiment of the present invention automatically extracts the boundary data of the defect area from the three-dimensional digital defect prediction combined model to clarify the scope of mesh generation. A geometric mesh generation method based on Delaunay triangulation or surface reconstruction algorithm is applied to generate a triangular mesh that conforms to the model geometry in the defect area. The generated mesh is optimized, including eliminating overlapping meshes, smoothing boundaries, adjusting mesh density, etc., to ensure the continuity and stability of the mesh. According to the predicted geometric shape and texture features, the defect area is filled, and the mesh structure is seamlessly connected to the surrounding complete area to generate a three-dimensional digital model with a continuous mesh structure.

[0095] Step S33: optimizing the texture effect of the defective area of ​​the continuous grid structure three-dimensional digital model using the three-dimensional model defect prediction data to obtain a repaired three-dimensional digital model;

[0096] The embodiments of the present invention extract texture features of the defective area, including color, gloss, and texture pattern, from the 3D model defect prediction data. Using the extracted texture features, a texture mapping algorithm is used to accurately project the texture information of the defective area onto the corresponding position of the continuous mesh structure 3D model. Texture boundaries are smoothed to avoid discontinuities or abruptness at the texture junctions, ensuring a natural transition of the overall texture effect. Texture detail is enhanced in the defective area to simulate the texture characteristics of a real object, such as increasing texture resolution and adjusting light and shadow effects. After optimizing the texture effect, a repaired 3D digital model is generated.

[0097] Step S34: performing geometry and texture consistency verification on the repaired three-dimensional digital model, and optimizing the repaired three-dimensional digital model based on the consistency verification result to obtain a reconstructed three-dimensional digital model.

[0098] The embodiment of the present invention compares the geometry of the defective area of ​​the repaired three-dimensional digital model with the geometry of the original model, and checks the degree of matching of geometric parameters, including edge alignment, surface flatness and mesh density consistency. The texture of the repaired area is analyzed for consistency with the texture of the complete area to verify whether the color, gloss and pattern connection of the texture are natural and coordinated. The areas with geometric discontinuities or overly dense meshes are readjusted to improve the smoothness and consistency of the model surface. Fine-tuning is performed in areas where the texture does not match to ensure a seamless transition between the texture details and the complete parts. Combined with the results of the geometry and texture consistency verification, the model is corrected and improved as a whole to ensure that the defective area is highly consistent with the complete area. After verification and optimization processing, a reconstructed three-dimensional digital model is generated.

[0099] The present invention can accurately identify and locate defective areas by geometrically aligning the three-dimensional model defect prediction data with the initial three-dimensional digital model. The subsequent geometric grid generation and filling process ensures that the defective area is accurately restored, and the generated continuous grid structure three-dimensional digital model is more complete and natural, avoiding improper or unnatural connections that occur in manual restoration. Using the defect prediction data to optimize the texture effect of the repaired three-dimensional model can effectively restore the visual effect of the defective area and ensure that the texture style of the repaired part is consistent with that of the original model, thereby enhancing the overall visual realism and integrity. By performing geometric and texture consistency checks on the repaired three-dimensional digital model, inconsistencies or errors generated during the repair process can be detected and corrected. Further optimization is performed based on the consistency check results to ensure that the final reconstructed three-dimensional digital model has high consistency and high quality in terms of geometric shape and surface texture. Through optimization processing, the reconstructed three-dimensional digital model obtained not only has a better geometric structure, but also has a more perfect visual effect. The final result can be widely used in fields requiring high-precision three-dimensional reconstruction, such as cultural heritage protection, medical image processing, and industrial detection.

[0100] Preferably, step S4 includes the following steps:

[0101] Step S41: performing geometric mesh optimization on the reconstructed three-dimensional digital model to obtain a mesh-optimized three-dimensional digital model;

[0102] The embodiment of the present invention conducts a comprehensive inspection of the geometric mesh of the reconstructed three-dimensional digital model and analyzes the mesh distribution, including mesh size, density, shape, and whether there is overlap or staggering. For areas with uneven mesh distribution, a mesh reconstruction algorithm is used to re-divide the mesh so that its density is uniform and its shape is regular. Redundant meshes are streamlined, unnecessary vertices or edges are deleted, and the complexity of the model is reduced while maintaining geometric accuracy. By adjusting the mesh topology, sharp edges, extremely small or extremely large mesh units, and gaps or overlaps between meshes are eliminated, thereby enhancing the overall topological stability of the model. Mesh parameters such as resolution and smoothness are adjusted according to the geometric complexity and detail requirements of the model. Finally, a mesh-optimized three-dimensional digital model is generated.

[0103] Step S42: performing surface optimization on the mesh-optimized three-dimensional digital model and removing noise to obtain a surface-smoothed optimized three-dimensional digital model;

[0104] The embodiment of the present invention uses a surface analysis algorithm to detect the surface of the mesh-optimized three-dimensional digital model and identify surface noise areas. A Laplace smoothing algorithm or a bidirectional curvature constraint algorithm is used to adjust the surface vertex positions, eliminate small-scale fluctuations, and achieve smooth surface transitions. Gaussian filtering or non-local mean filtering methods are applied to remove high-frequency noise while retaining the main surface features. During the smoothing process, constraints are set for the key boundaries of the model to avoid distortion of the boundaries due to smoothing. The entire model surface is optimized for consistency to ensure that the smoothness of each part is coordinated and to avoid excessive or insufficient local smoothing. Ultimately, a surface-smoothed and optimized three-dimensional digital model is obtained.

[0105] Step S43: performing detail restoration on the surface smoothed and optimized three-dimensional digital model to obtain a detail enhanced three-dimensional digital model;

[0106] The embodiment of the present invention uses a multi-scale analysis method to extract detail features from a smoothed and optimized three-dimensional model. This can be achieved by using a method based on a Gaussian pyramid or a Laplace pyramid to extract the fine structure and texture details of the surface. The extracted detail information is locally enhanced, and a local magnification algorithm is used to enhance the surface details of the model. Using detail mapping technology, the enhanced detail information is superimposed on the original model surface to restore the tiny surface features lost due to smoothing. Morphological processing technology is used to repair the details of the geometric shape to ensure the consistency and coherence of the surface details. Accurate local reconstruction is performed on the complex areas of the model surface to ensure that the details are highly consistent with the original data. Based on the detail recovery data and combined with the texture information of the original three-dimensional model, the surface texture of the detail part is restored by texture mapping to enhance the surface texture. Texture synthesis technology is used to optimize local texture details to ensure that the details are consistent with the geometric surface. Finally, a detail-enhanced three-dimensional digital model is obtained.

[0107] Step S44: performing a global quality check on the detail-enhanced three-dimensional digital model to obtain three-dimensional digital model restoration parameters;

[0108] The embodiment of the present invention uses an automated tool to perform a global quality assessment of a three-dimensional digital model, mainly checking the model's geometry, texture consistency, and mesh quality. Evaluate the mesh quality of the model, including the shape of the triangular facets, the distribution of side lengths, the coherence of the mesh, etc., to ensure that there are no self-intersections, overlaps, or irregular facets. Check the seamlessness of the texture map to avoid stretching, distortion, or repeated texture problems. Compare the optimized three-dimensional model with the original model data to confirm whether there are any missing or abnormal protrusions on the surface to ensure that there are no significant errors in the geometric form. Based on the quality inspection results, the problematic areas are automatically identified, such as geometric defects such as holes, sharp angles, and skew on the model surface. The part that needs to be repaired is determined by an algorithm, and the defective area is marked. Based on the inspection results, repair parameters are generated, covering the repair method of the defective area, the optimization direction, the required geometric adjustment amount, and the texture mapping repair method, and finally the three-dimensional digital model repair parameters are obtained.

[0109] Step S45: adjusting and optimizing the detail-enhanced three-dimensional digital model based on the three-dimensional digital model restoration parameters to obtain an optimized three-dimensional digital model.

[0110] The embodiment of the present invention uses three-dimensional digital model repair parameters to make precise adjustments to problem areas in the detail-enhanced three-dimensional digital model. A geometric repair algorithm is applied to make adjustments to geometric inconsistencies or defects found during the inspection. Sharp corners, depressions, or discontinuous mesh parts are repaired to smooth the surface. Overly complex or irregular meshes are simplified or regenerated to improve the smoothness and stability of the geometric structure. Textures are adjusted using texture mapping repair parameters to eliminate stretching, breakage, and repetition to ensure seamless texture transitions. Texture fit is optimized through texture remapping, patching technology, and other methods to make the model surface more natural. After the adjustment is completed, global quality inspection is performed again through automated tools to verify the adjustment and optimization effects to ensure that the geometric shape, surface smoothness, texture consistency, and mesh structure all meet the optimization standards. The final optimized three-dimensional digital model is generated.

[0111] Through geometric mesh optimization, the present invention optimizes the mesh structure of the three-dimensional digital model, reduces redundant parts and irregular shapes in the model, and enhances the fineness and accuracy of the mesh, thereby ensuring higher geometric accuracy and laying a solid foundation for subsequent processing. The surface optimization and noise removal steps effectively remove unnecessary interference in the model, making the model surface smoother and finer. The denoising process improves the clarity and quality of the model, avoiding uneven texture or uneven surface caused by noise. Through detail recovery, the details of the model are enhanced, making the three-dimensional digital model more realistic and accurate, especially when representing complex surface features or tiny details, it can better preserve the true appearance of the object, enhance the visual effect and perceived quality. Global quality inspection can identify potential quality problems in the model and optimize it according to the repair parameters to ensure the global consistency and accuracy of the model's geometric structure and surface texture. This step helps to improve the stability and adaptability of the model, ensuring that it can meet the needs of different application scenarios. Through adjustment and optimization based on repair parameters, the final three-dimensional digital model not only has improved geometry and surface quality, but also has higher accuracy and visual consistency, making it suitable for demanding application scenarios such as virtual reality, medical imaging, industrial inspection, etc.

[0112] Preferably, step S42 includes the following steps:

[0113] Step S421: monitoring and marking noise points on the surface of the mesh-optimized three-dimensional digital model to obtain three-dimensional model noise point marking data;

[0114] The embodiment of the present invention clarifies the definition of noise points. Usually, noise points appear as abnormal vertices or triangles in the grid that deviate from the surrounding point sets. These noise points are usually caused by scanning errors, inaccuracies in the reconstruction process, or data loss. Noise points are detected by analyzing the surface properties of the three-dimensional digital model using computer vision or graphics processing technology. These points appear as areas that are inconsistent with the neighborhood structure, have high local geometric fluctuations, or they do not match the topological structure of the surrounding area. Once a potential noise point is detected, it is marked by an automated algorithm. A geometric or topological standard is set, and points that exceed the standard are considered noise points. By analyzing the topological structure of adjacent grids, points that are seriously inconsistent with the surrounding grids are found and marked as noise points. All detected noise point information will be recorded, and eventually form the three-dimensional model noise point marking data.

[0115] Step S422: removing noise points from the mesh-optimized 3D digital model according to the 3D model noise point marking data to obtain a denoised 3D digital model;

[0116] The embodiment of the present invention utilizes the noise point marker data of the three-dimensional model to identify the noise points in the grid that do not conform to the normal geometric and topological structure. The noise points exist in the local area of ​​the model, and are usually manifested as abnormal connections with the surrounding grid points or geometric mutations. For each noise point, the geometric features of other points in its neighborhood are calculated and used as a reference to correct the position of the noise point. Weighted averaging or least squares method is usually used for repair. For noise points, the information of adjacent grid points is used for reconstruction. A local surface fitting algorithm (such as B-spline surface or Laplacian smoothing) is used to eliminate noise points and smooth the surrounding areas. Based on the noise point marker data, the identified noise points are deleted from the three-dimensional digital model. If the noise point cannot be removed directly, interpolation or smoothing technology is used to smooth the surrounding area to make its transition more natural. After the noise point removal process, the three-dimensional model grid structure is updated to obtain a denoised three-dimensional digital model.

[0117] Step S423: Optimizing the surface smoothness of the denoised three-dimensional digital model to obtain a preliminary smooth three-dimensional digital model;

[0118] Based on the denoised three-dimensional digital model, the embodiment of the present invention further optimizes the model surface to make it smoother, eliminates local discontinuities or rough parts caused by noise removal, and thus improves the smoothness of the overall geometric shape. For each grid point, the difference between it and the adjacent points is calculated, and the position of each point is adjusted by weighted averaging to reduce the high-frequency noise on the surface. Usually, multiple iterations are performed to make the grid surface smoother. The grid is smoothed by the B-spline surface fitting method to reduce excessive surface features. Gaussian filtering is applied to the grid surface to smooth the position of each grid point while maintaining the overall geometric structure. The degree of smoothing is controlled by setting a suitable Gaussian kernel function. The above-mentioned smoothing technology is applied to the denoised three-dimensional digital model to gradually optimize the surface geometry of the model. During the smoothing process, attention is paid to retaining edges and detail features to avoid loss of geometric features due to excessive smoothing. After completion, a preliminary smoothed three-dimensional digital model is obtained.

[0119] Step S424: performing feature enhancement on the preliminary smooth three-dimensional digital model to obtain a surface smoothed optimized three-dimensional digital model.

[0120] Based on the initially smoothed 3D digital model, the present invention further restores and enhances its important geometric features, such as edges, corners, and uneven details, to make it more consistent with the morphology and structure of the real object. High-pass filtering or sharpening algorithms are used to restore details on the smooth surface, highlighting edges and texture information to make the model details clearer. Local geometric features of the model surface are weighted to enhance details in certain areas. For example, during the smoothing process, key edges or prominent structures are specifically enhanced to prevent loss of detail due to oversmoothing. Surface fitting techniques are used to supplement smooth areas of the model with details, reconstructing complex geometric forms and enhancing the surface details of the object. Texture details are enhanced to enhance surface features, making them more natural and realistic, especially for objects with texture or uneven surfaces. Local features of the initially smoothed 3D model are enhanced to preserve details and enhance edges and important structures. Appropriate enhancement algorithms are used to ensure that the model's geometric features are restored while not affecting the overall smoothing effect. Ultimately, a smoothed and optimized 3D digital model is obtained.

[0121] By monitoring and marking noise points on the model's surface, the present invention enables the system to effectively identify and remove unnecessary noise, thereby producing a de-noised 3D digital model. This noise removal process significantly improves the model's accuracy and prevents noise from affecting subsequent processing or visualization. The surface smoothness of the de-noised 3D digital model is optimized to ensure a smoother surface. Optimizing surface smoothness reduces surface irregularities, making the surface more uniform and improving visual quality and model accuracy, particularly advantageous in applications requiring detailed visualization. Feature enhancement is performed on the initially smoothed 3D digital model to enhance key surface details and texture features. Feature enhancement can highlight important structures and details, improving the model's precision and realism. This is particularly suitable for applications requiring high-detail representation, such as industrial inspection and virtual reality. The entire process, through the sequential steps of noise removal, surface optimization, and feature enhancement, gradually improves the quality of the 3D model, ensuring more realistic details, more accurate structures, and greater adaptability to meet demanding application scenarios.

[0122] Preferably, step S43 includes the following steps:

[0123] Step S431: identifying key detail areas of the surface smoothed and optimized three-dimensional digital model to obtain detail area marking data of the three-dimensional model;

[0124] The embodiment of the present invention identifies and marks detail areas in a surface-smoothed and optimized three-dimensional digital model that are critical to the final visual effect or function. By calculating the geometric features of the three-dimensional model surface, such as curvature, normal direction, and edge sharpness, areas where the surface changes are more significant are identified. For example, a higher curvature value usually indicates that the area is a detail area, such as a sharp edge, hole, or protrusion. Based on different surface features, an adaptive algorithm is used to dynamically adjust the recognition criteria. For the geometric structures of different objects, the sensitivity of the analysis is adjusted to ensure that key details that meet expectations can be identified. In some areas with relatively smooth surfaces but complex textures, a texture analysis algorithm is used to assist in determining which areas should be marked as detail areas. These areas usually contain important texture or detail information, which determines the realism of the model. Using a trained deep learning model, key detail areas are automatically identified and marked by analyzing the surface features of the three-dimensional model. The key detail areas identified and marked by the above algorithm form the detail area marking data of the three-dimensional model.

[0125] Step S432: geometrically reconstructing the surface smoothed and optimized three-dimensional digital model according to the three-dimensional model detail region marking data to obtain a reconstructed detail region three-dimensional digital model;

[0126] The embodiment of the present invention reconstructs the geometric shape based on the marker data of the detail area of ​​the three-dimensional model, restores the original shape or optimized shape of the detail area, and ensures that the geometric structure of the detail area is more accurate and consistent with the characteristics of the actual object. Using the marker data of the detail area, focus on the local details of the model, and restore missing or incomplete geometric information through interpolation algorithms. Apply shape recovery technology to the detail area, such as shape generation from boundaries or constraint-based geometric recovery, to ensure that the geometric structure of the detail area is as consistent as possible with the characteristics of the original object. Apply surface fitting technology to the detail area, use the marker data and information of the adjacent areas to perform local surface fitting, so as to obtain a more accurate geometric shape. Adaptively adjust the reconstruction process according to the curvature or other geometric properties of the object surface. In view of the fact that there are features of different scales in the detail area, a multi-scale geometric analysis method is used to ensure that the reconstruction process takes into account both local details and global geometric consistency, avoiding detail loss or mismatch. Through geometric reconstruction, a reconstructed three-dimensional digital model of the detail area containing detail recovery information is obtained.

[0127] Step S433: amplifying the high-frequency features of the detail region and enhancing the clarity of local features of the reconstructed detail region three-dimensional digital model to obtain a detail-enhanced three-dimensional digital model.

[0128] This embodiment of the present invention reconstructs a 3D digital model of detail areas and further optimizes these areas using high-frequency feature amplification and local feature clarity enhancement techniques to improve the model's detail expressiveness and clarity. Frequency domain analysis is used to amplify high-frequency features in the detail areas of the 3D model. A discrete Fourier transform is used to enhance the high-frequency components in the detail areas, making these details more visually prominent. By combining the curvature and surface variations of the 3D mesh, high-frequency features in the detail areas are locally amplified to enhance the display of subtle surface variations, particularly complex textures and structures. A shape-preserving filter is applied to enhance local features in the detail areas, such as bumps and edges, removing blur and ensuring every detail is more clearly discernible. Parameter adjustment ensures that the detail areas are neither overly smoothed (losing realism) nor affected by noise, thereby enhancing the clarity and realism of local details. After high-frequency feature amplification and local clarity enhancement, a detail-enhanced 3D digital model is obtained.

[0129] The present invention identifies key detail areas in a smoothed and optimized three-dimensional digital model, allowing the system to accurately mark detail areas that require further processing. This process ensures that subsequent optimization and enhancement work can be focused on important areas, improving efficiency and reducing unnecessary processing. The geometric morphology of the three-dimensional model is reconstructed based on the detail area marking data, making the morphology of the detail area more refined and realistic. The reconstructed three-dimensional digital model of the detail area better conforms to the actual shape and features of the object, enhancing the accuracy and detail of the model. Amplifying high-frequency features and enhancing the clarity of local features in the reconstructed detail area can effectively enhance the expressiveness and visual effect of the details. This processing makes the details more prominent, and the model exhibits higher precision and realism, making it particularly suitable for fields requiring detailed display, such as digital cultural relic protection, industrial design, and virtual reality. Overall, through the identification, geometric reconstruction, and feature enhancement of key detail areas, this method can significantly improve the accuracy and visual effect of the three-dimensional model in detail presentation, especially for application scenarios that require highlighting details and textures.

[0130] Preferably, step S5 includes the following steps:

[0131] Step S51: converting the optimized three-dimensional digital model into a new format, and compressing and optimizing the converted format result to obtain real-time rendered three-dimensional model data;

[0132] The embodiment of the present invention converts the optimized three-dimensional digital model into a format suitable for real-time rendering according to the requirements of the target platform. These formats generally support properties such as texture, color, material, and are compatible with different rendering engines and display platforms. Use three-dimensional modeling software or a dedicated format conversion tool to accurately convert the data of the original three-dimensional digital model into the target format while maintaining high details. In response to the needs of real-time rendering, data compression technology is applied to reduce the file size and improve transmission and loading efficiency. Compression methods include but are not limited to vertex and facet-based geometric compression technology and texture mapping compression. Ensure that the compressed data can be quickly loaded during the rendering process without significantly losing visual effects, reduce memory usage, and improve real-time rendering performance. After completing the format conversion and compression optimization, the real-time rendering three-dimensional model data is output.

[0133] Step S52: performing interactive parameter configuration based on the real-time rendering 3D model data to obtain real-time rendering display configuration data;

[0134] The embodiment of the present invention selects a suitable interaction method based on the display platform and user needs. Determine the interactive functions that need to be configured. Set operational viewing parameters, such as camera position, viewing angle limit, dynamic zoom and rotation angle, to ensure that the user can freely adjust the viewing angle to view different parts of the three-dimensional model. Configure feedback effects for each interactive behavior, such as generating visual or sound feedback when the user clicks or drags the three-dimensional model to enhance the interactive experience. Set interactive hotspots in the model (for example, clicking on a part of the model can display more details or expand information), and bind them to the corresponding interactive functions. Optimize the rendering process according to the interaction requirements to ensure that the three-dimensional model can be displayed smoothly without freezes or delays during interaction. Configure relevant display parameters to ensure that the model rendering effect is stable and clear during the interaction process. Integrate the above-mentioned interactive parameters and display settings to generate the final real-time rendering display configuration data.

[0135] Step S53: Design interactive functions according to the real-time rendering display configuration data to obtain interactive display application data;

[0136] The embodiment of the present invention designs specific interactive functions based on real-time rendering and display configuration data. Designs a response mechanism for user input methods. Designs interactive interface elements to facilitate user interaction with the three-dimensional model. Designs information boxes or annotations for different areas or objects of the three-dimensional model, and the user displays relevant information or graphic materials by clicking on a certain part. Designs animation effects for specific operations (such as rotation, zooming in and out) to enhance the user experience. For example, when clicking on an object, the object's relevant data or rotation animation can be displayed. Provides immediate visual and auditory feedback. For example, when the user selects a certain part, the model color or highlight display changes, accompanied by corresponding sound effects. Integrate various interactive functions with interface elements to generate a complete interactive display application data file.

[0137] Step S54: loading the interactive display application data into the display platform, and performing multi-platform adaptation optimization to obtain three-dimensional digital model display data.

[0138] The embodiment of the present invention imports interactive display application data into the target display platform. This data includes three-dimensional models, interaction logic, UI elements and configuration files. According to the hardware and software environment of the target platform, adjust the rendering parameters, such as resolution, texture quality, lighting effects, etc., to ensure that the model can be presented efficiently and clearly. Analyze the performance and requirements of different platforms. Optimize the display data according to the processing capabilities and user interaction methods of different platforms. Through adaptive technology, adjust the size, display method and control method of the three-dimensional model and interactive elements. Compress and simplify model data to reduce rendering load and improve loading speed and smoothness. Use LOD technology to load models of different levels of detail on demand to optimize memory and processing power. Perform functional and performance tests on each platform to ensure that the display effect of the model is consistent on different devices and that user interaction is smooth. Adjust the display effect according to test feedback, fix compatibility or performance issues, and ensure smooth operation on multiple platforms. Finally, three-dimensional digital model display data that adapts to multiple platforms is obtained.

[0139] By converting the format and optimizing the compression of the optimized three-dimensional digital model, the system can generate efficient data suitable for real-time rendering. This process ensures that the three-dimensional model can be quickly loaded and rendered on different platforms, improving the display effect and interactive experience. Interactive parameter configuration is performed based on the real-time rendering three-dimensional model data, allowing users to adjust display parameters such as viewing angle, lighting, and detail display according to their needs. This function allows users to customize the experience, thereby improving the flexibility and adaptability of the model display. By designing interactive functions based on the real-time rendering display configuration data, the interactivity of the display is further enhanced. This allows users to not only view the three-dimensional model, but also explore and analyze the details of the model more intuitively through interactive functions, increasing the sense of participation and operability. The interactive display application data is loaded onto the display platform and multi-platform adaptation optimization is performed, so that the three-dimensional model can be displayed smoothly on different devices and systems. Whether on a computer, mobile device, or virtual reality platform, the system can provide stable and consistent display effects, improving the system's universality and cross-platform compatibility.

[0140] Preferably, the present invention further provides a three-dimensional optical system for executing the above-mentioned method for realizing the three-dimensional optical system, wherein the three-dimensional optical system comprises:

[0141] The 3D model initial construction module is used to scan the target object through an optical 3D scanning device to obtain 3D geometric data; collect surface texture data of the target object to obtain surface texture data; and construct a 3D model using the 3D geometric data and surface texture data to obtain an initial 3D digital model.

[0142] The defect area identification and prediction module is used to identify and mark the defect area of ​​the initial three-dimensional digital model to obtain the defect area data of the three-dimensional model; simulate and restore the defect area data of the three-dimensional model to generate the defect prediction data of the three-dimensional model;

[0143] A three-dimensional model restoration and reconstruction module is used to restore and reconstruct the initial three-dimensional digital model according to the three-dimensional model defect prediction data to obtain a reconstructed three-dimensional digital model;

[0144] The model optimization and detail enhancement module is used to optimize the surface smoothness of the reconstructed three-dimensional digital model and enhance the details to obtain an optimized three-dimensional digital model;

[0145] The three-dimensional model visualization display module is used to visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

[0146] The present invention uses an optical 3D scanning device to scan a target object. The system accurately acquires 3D geometric and surface texture data, ensuring high-precision construction of the initial 3D digital model. This lays a solid foundation for subsequent modeling, restoration, and optimization. By identifying and marking defective areas in the initial 3D model and performing predictive restoration based on the defective area data, the system can accurately identify the defective parts in the model and simulate the features of the defective areas. This function can significantly improve the accuracy and efficiency of model restoration. Utilizing the defect prediction data, the system accurately restores and reconstructs the initial 3D digital model, restoring the morphology and details of the defective areas, ensuring the final 3D digital model is more complete and realistic. By optimizing the surface smoothness and enhancing the details of the reconstructed 3D model, the system improves the model's visual quality and detail, further enhancing the accuracy and realism of the 3D model and making it more expressive during presentation. The system provides real-time display and interactive features, allowing users to intuitively and flexibly view and manipulate the 3D model. This feature allows users to not only deeply observe the model's details but also interact from different perspectives, enhancing the user experience and the flexibility of model presentation.

[0147] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is not limited by the above description. Therefore, it is intended that all changes that fall within the meaning and scope of the equivalent elements of the application documents are included in the present invention.

[0148] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for realizing a three-dimensional optical system, characterized in that: The following steps are involved: Step S1: Scanning a target object with an optical 3D scanning device to obtain 3D geometric data; collecting surface texture data of the target object to obtain surface texture data; constructing a 3D model using the 3D geometric data and the surface texture data to obtain an initial 3D digital model; Step S2: Identify and mark the defective areas of the initial three-dimensional digital model to obtain the defective area data of the three-dimensional model; perform simulation restoration based on the defective area data of the three-dimensional model to generate the three-dimensional model defect prediction data; wherein step S2 includes the following steps: Step S21: meshing the initial three-dimensional digital model to obtain meshed data of the initial three-dimensional model; Step S22: identifying and marking surface defect areas of the initial three-dimensional digital model based on the initial three-dimensional model gridding data to obtain defect area data of the three-dimensional model; Step S23: adjusting the defect boundary precision according to the defect area data of the 3D model to obtain accurate defect area data of the 3D model; Step S24: extracting geometric and texture features from the defective area data of the precise three-dimensional model to obtain feature data of the defective area of ​​the three-dimensional model; Step S25: Using the feature data of the defective area of ​​the 3D model, a defective area restoration simulation is performed on the initial 3D digital model to generate 3D model defect prediction data; wherein step S25 includes the following steps: Step S251: constructing a preliminary interpolation model of the defective region based on the feature data of the defective region of the three-dimensional model to obtain a preliminary interpolation model of the defective region; Step S252: optimizing the defect area geometry of the preliminary defect area interpolation model using the defect area data of the accurate three-dimensional model to obtain optimized defect area geometry data; Step S253: reconstructing the texture of the defective area according to the feature data of the defective area of ​​the three-dimensional model and the optimized geometric data of the defective area to obtain texture restoration data of the defective area; Step S254: Comprehensively verifying the optimized defect area geometry data and the defect area texture restoration data to obtain defect area restoration data; Step S255: Fusing the restored data of the defect area with the initial three-dimensional digital model to obtain a defect prediction three-dimensional model; Step S256: performing defect area restoration simulation using the defect prediction three-dimensional model to obtain three-dimensional model defect prediction data; Step S3: restoring and reconstructing the initial three-dimensional digital model according to the three-dimensional model defect prediction data to obtain a reconstructed three-dimensional digital model; Step S4: Optimizing the surface smoothness of the reconstructed three-dimensional digital model and performing detail enhancement to obtain an optimized three-dimensional digital model; Step S5: Visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

2. The method for realizing a three-dimensional optical system according to claim 1, wherein: Step S1 includes the following steps: Step S11: Scanning the target object at multiple angles using an optical 3D scanning device, and collecting surface geometric information to obtain 3D geometric data; Step S12: performing noise reduction processing on the three-dimensional geometric data, and aligning and fusing the noise reduction processing results to obtain high-precision three-dimensional data; Step S13: collecting surface texture information of the target object from multiple angles using a camera to obtain surface texture data; Step S14: Mapping the surface texture data to high-precision three-dimensional data, and constructing a three-dimensional model to obtain an initial three-dimensional digital model.

3. The method for realizing a three-dimensional optical system according to claim 1, wherein: Step S3 includes the following steps: Step S31: geometrically aligning the three-dimensional model defect prediction data with the initial three-dimensional digital model to obtain a three-dimensional digital defect prediction combined model; Step S32: generating a geometric grid of the defect area according to the three-dimensional digital defect prediction combined model, and filling the defect area to generate a three-dimensional digital model with a continuous grid structure; Step S33: optimizing the texture effect of the defective area of ​​the continuous grid structure three-dimensional digital model using the three-dimensional model defect prediction data to obtain a repaired three-dimensional digital model; Step S34: performing geometry and texture consistency verification on the repaired three-dimensional digital model, and optimizing the repaired three-dimensional digital model based on the consistency verification result to obtain a reconstructed three-dimensional digital model.

4. The method for realizing a three-dimensional optical system according to claim 1, wherein: Step S4 includes the following steps: Step S41: performing geometric mesh optimization on the reconstructed three-dimensional digital model to obtain a mesh-optimized three-dimensional digital model; Step S42: performing surface optimization on the mesh-optimized three-dimensional digital model and removing noise to obtain a surface-smoothed optimized three-dimensional digital model; Step S43: performing detail restoration on the surface smoothed and optimized three-dimensional digital model to obtain a detail enhanced three-dimensional digital model; Step S44: performing a global quality check on the detail-enhanced three-dimensional digital model to obtain three-dimensional digital model restoration parameters; Step S45: adjusting and optimizing the detail-enhanced three-dimensional digital model based on the three-dimensional digital model restoration parameters to obtain an optimized three-dimensional digital model.

5. The method for realizing a three-dimensional optical system according to claim 4, wherein: Step S42 includes the following steps: Step S421: monitoring and marking noise points on the surface of the mesh-optimized three-dimensional digital model to obtain three-dimensional model noise point marking data; Step S422: removing noise points from the mesh-optimized 3D digital model according to the 3D model noise point marking data to obtain a denoised 3D digital model; Step S423: Optimizing the surface smoothness of the denoised three-dimensional digital model to obtain a preliminary smooth three-dimensional digital model; Step S424: performing feature enhancement on the preliminary smooth three-dimensional digital model to obtain a surface smoothed optimized three-dimensional digital model.

6. The method for realizing a three-dimensional optical system according to claim 4, wherein: Step S43 includes the following steps: Step S431: identifying key detail areas of the surface smoothed and optimized three-dimensional digital model to obtain detail area marking data of the three-dimensional model; Step S432: geometrically reconstructing the surface smoothed and optimized three-dimensional digital model according to the three-dimensional model detail region marking data to obtain a reconstructed detail region three-dimensional digital model; Step S433: amplifying the high-frequency features of the detail region and enhancing the clarity of local features of the reconstructed detail region three-dimensional digital model to obtain a detail-enhanced three-dimensional digital model.

7. The method for realizing a three-dimensional optical system according to claim 1, wherein: Step S5 includes the following steps: Step S51: converting the optimized three-dimensional digital model into a new format, and compressing and optimizing the converted format result to obtain real-time rendered three-dimensional model data; Step S52: performing interactive parameter configuration based on the real-time rendering 3D model data to obtain real-time rendering display configuration data; Step S53: Design interactive functions according to the real-time rendering display configuration data to obtain interactive display application data; Step S54: loading the interactive display application data into the display platform, and performing multi-platform adaptation optimization to obtain three-dimensional digital model display data.

8. A three-dimensional optical system, characterized in that: A method for implementing a three-dimensional optical system according to claim 1, wherein the three-dimensional optical system comprises: The 3D model initial construction module is used to scan the target object through an optical 3D scanning device to obtain 3D geometric data; collect surface texture data of the target object to obtain surface texture data; and construct a 3D model using the 3D geometric data and surface texture data to obtain an initial 3D digital model. The defect area identification and prediction module is used to identify and mark the defect area of ​​the initial three-dimensional digital model to obtain the defect area data of the three-dimensional model; simulate and restore the defect area data of the three-dimensional model to generate the defect prediction data of the three-dimensional model; A three-dimensional model restoration and reconstruction module is used to restore and reconstruct the initial three-dimensional digital model according to the three-dimensional model defect prediction data to obtain a reconstructed three-dimensional digital model; The model optimization and detail enhancement module is used to optimize the surface smoothness of the reconstructed three-dimensional digital model and enhance the details to obtain an optimized three-dimensional digital model; The three-dimensional model visualization display module is used to visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

Images