Implementation method and system of three-dimensional optical system

Abstract

The invention relates to the technical field of three-dimensional optics, in particular to an implementation method and system of a three-dimensional optical system. The method comprises the following steps: scanning a target object through an optical three-dimensional scanning device to obtain three-dimensional geometric data; collecting surface texture data of a target object to obtain surface texture data; constructing a three-dimensional model by using the three-dimensional geometric data and the surface texture data to obtain an initial three-dimensional digital model; carrying out defect area identification and marking on the initial three-dimensional digital model to obtain three-dimensional model defect area data; and performing simulation reduction according to the three-dimensional model defect area data to generate three-dimensional model defect prediction data. According to the method, the high-quality three-dimensional digital model can be efficiently generated through data collection, accurate modeling and automatic defect prediction and repair of the three-dimensional optical acquisition system, excellent user interaction experience is provided, and the method is suitable for the field of cultural relic repair.

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

[0003] Current 3D optical systems often show great 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 of the generated 3D model. In addition, for target objects with transparent, reflective, or high-contrast surfaces, traditional scanning equipment has difficulty accurately capturing their details. 3D modeling systems on the market can usually capture the complete shape of objects, but for areas that are damaged or missing, existing systems cannot automatically and effectively complete 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

[0005] 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.

[0006] To achieve the above object, a method for implementing a three-dimensional optical system includes the following steps: Step S1: Scanning the target object by 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: identifying and marking defective areas of the initial three-dimensional digital model to obtain defective area data of the three-dimensional model; performing simulation restoration based on the defective area data of the three-dimensional model to generate defect prediction data of the three-dimensional model; 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.

[0007] The present invention collects geometric data and surface texture data of the target object through an optical three-dimensional scanning device, and can accurately construct an initial three-dimensional digital model. This method not only improves the accuracy of modeling, but also obtains rich surface details. By identifying and marking the defective area of ​​the three-dimensional model, combined with the simulated restoration of the defective area data, the defective area can be accurately predicted, and high-quality three-dimensional model defect prediction data can be generated. This provides a reliable basis for subsequent repair and reconstruction. The defective part of the target object can be restored and reconstructed using the defect prediction data, and the integrity and authenticity of the model can be improved. Through surface smoothness optimization and detail enhancement, the visual effect of the three-dimensional model can be effectively improved, making it more delicate and realistic, and meeting the requirements of different application scenarios for model quality. The optimized three-dimensional digital model can be displayed and interacted in real time, supports the interactive operation between users and the three-dimensional model, enhances the usability and experience of the model, and is suitable for a variety of display platforms such as virtual reality and augmented reality. Through multi-platform adaptation, it is ensured that the three-dimensional model can be displayed smoothly on various devices, and the flexibility and universality of the system are improved.

[0008] 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: 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; use the 3D geometric data and surface texture data to construct a 3D model 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.

[0009] The present invention scans the target object through an optical three-dimensional scanning device, and the system can accurately obtain three-dimensional geometric data and surface texture data to ensure the high-precision construction of the initial three-dimensional digital model. This lays a solid foundation for subsequent modeling, repair and optimization. By identifying and marking the defective area of ​​the initial three-dimensional model and predicting and restoring it based on the defective area data, the system can accurately identify the defective part in the model and simulate the characteristics of the defective area. This function can greatly improve the accuracy and efficiency of model repair. Using the defect prediction data, the system accurately restores and reconstructs the initial three-dimensional digital model, thereby restoring the morphology and details of the defective area, ensuring that the final three-dimensional digital model is more complete and realistic. By optimizing the surface smoothness and enhancing the details of the reconstructed three-dimensional model, the system can improve the visual effect and detail expression of the model, further improve the accuracy and authenticity of the three-dimensional model, and make it more expressive during the display process. The system provides real-time display and interaction functions, allowing users to view and operate the three-dimensional model intuitively and flexibly. Through this function, users can not only observe the details of the model in depth, but also interact from different perspectives, enhancing the user experience and the flexibility of model display. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings: Figure 1 A schematic flow chart of the steps of the method for implementing the three-dimensional optical system of the present invention; Figure 2 for Figure 1 Detailed step flow diagram of step S1; Figure 3 for Figure 1 Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0013] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0014] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0015] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0016] 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: Step S1: Scanning the target object by 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: identifying and marking defective areas of the initial three-dimensional digital model to obtain defective area data of the three-dimensional model; performing simulation restoration based on the defective area data of the three-dimensional model to generate defect prediction data of the three-dimensional model; 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.

[0017] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a schematic diagram of a process flow 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: Step S1: Scanning the target object by 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; The embodiment of the present invention selects a structured light scanner, which is composed of a projector, a camera and a computer control system. The projector projects a series of light strips onto the surface of the target object, and calculates the spatial coordinates of each point on the surface of the object based on the deformation pattern captured by the camera. The structured light scanner projects a known pattern of light strips onto the surface of the object, obtains the deformation of the reflected light strips with the help of the camera, and calculates the three-dimensional coordinates of each point using triangulation. By scanning from multiple perspectives, the system can accurately obtain the three-dimensional geometric information of the surface of the object. By scanning the target object with a structured light scanning device, the system can obtain the three-dimensional geometric data of the surface of the target object. According to the size and shape of the target object, select a suitable scanning perspective and angle. During the scanning process, the object needs to be rotated or the scanner position adjusted multiple times to ensure that all angles of the object are covered to avoid the appearance of shadows or blocked areas. According to the complexity of the details on the surface of the object, select a suitable scanning accuracy setting. For fine textures or tiny features, the system can select a higher resolution for scanning to ensure that the accuracy meets the subsequent modeling requirements. After the scanning is completed, the obtained three-dimensional geometric data includes information such as the spatial coordinates and normal vectors of each point on the surface of the object. While acquiring 3D geometric data, it is necessary to collect the surface texture data of the target object so as to perform texture mapping during 3D modeling. Use a high-resolution camera device to scan the texture of the target object, usually using a directional light source to avoid shadow interference, so as to clearly capture the surface details of the object. In this process, the lighting angle and camera position need to be optimized to obtain a high-quality image of the surface texture of the object. The collected image is processed, including color correction, denoising, and texture stitching. Then, the surface texture image is accurately registered with the 3D geometric data using UV mapping technology to form complete 3D surface texture data. Surface texture data includes the color, gloss, and reflection of the object surface. The 3D model is constructed using the 3D geometric data and surface texture data. The collected 3D geometric data is processed, including removing noise and repairing holes or overlapping parts in the scan. The data is converted into a continuous 3D mesh model using a triangular meshing algorithm. The surface texture data is mapped to the processed 3D mesh model to ensure the correct registration of the texture with the geometric model. The texture mapping algorithm is used to enhance the details and authenticity of the texture. Finally, the 3D model is constructed by fusion of the 3D geometric data and the surface texture data to obtain the initial 3D digital model.

[0018] Step S2: identifying and marking defective areas of the initial three-dimensional digital model to obtain defective area data of the three-dimensional model; performing simulation restoration based on the defective area data of the three-dimensional model to generate defect prediction data of the three-dimensional model; The embodiment of the present invention identifies the defective area in the model. The defective area is usually manifested as an incomplete or missing part of the surface in the scanned data, such as a hole or discontinuous area caused by the limitation of the scanning angle, the reflection of the object surface or the damage of the object itself. The three-dimensional model is processed using a grid analysis algorithm. This includes calculating the distribution of the mesh faces and vertices of the model. For each mesh face, the system will calculate the connectivity of its adjacent faces and find the holes or damaged places. Common algorithms such as point cloud-based data integration, connectivity analysis, etc. can be used to identify discontinuous areas on the surface of the object. By performing a deep analysis of the grid model, the machine learning algorithm is used to automatically identify those parts that are different from the surrounding areas. Once the defective areas are identified, the system will mark these areas and assign different colors or shapes to them for subsequent processing. The marked data will be saved as a defective area data set as the basis for subsequent repair and restoration. The defective area data of the three-dimensional model is obtained, including the spatial position, shape, size and missing geometric information of the defective area. The defective area is predicted using an algorithm based on interpolation or fitting. In addition to the prediction of geometric shapes, 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, the texture synthesis technology is used to restore the missing texture information. Through the above-mentioned geometric and texture reconstruction, the simulated restoration data of the defective area of ​​the three-dimensional model is obtained. The data includes the reconstructed geometry of the defective area, texture details, and the restoration prediction of the integrity of the object. The three-dimensional model defect prediction data is obtained, including the restored geometry information, predicted texture and surface details of the defective area.

[0019] 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; The embodiment of the present invention loads the three-dimensional model defect prediction data into the system. The system docks and integrates these data with the defective area in the initial three-dimensional digital model, and prepares 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 method. For example, if the defective area is a large surface defect, the system can use the surrounding complete surface for surface fitting to create a smooth, seamless surface. For complex defective areas, mesh repair-based algorithms can be used for geometric repair. These algorithms will extend the mesh information of the surrounding area to the defective area, reconstruct the geometric surface of the missing part, and ensure a smooth transition with the surrounding mesh. For smaller defective areas, the system can use boundary completion, convex hull algorithm, Laplace smoothing and other methods to fill in the geometric data of the defective area. The reconstructed geometry will be compared with the original three-dimensional model to check the repair effect of the defective area. This process can be verified by accuracy inspection and comparison of the differences before and after model reconstruction. If necessary, the system can perform local optimization and detail adjustment. Through geometric reconstruction, a reconstructed three-dimensional digital model containing a 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 defect 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 smoothly transition between the defect area and the surrounding area to avoid visible stitching marks. For smaller defect 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 restoration effects. Through texture reconstruction, a reconstructed three-dimensional digital model with complete texture details is obtained.

[0020] 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; The embodiment of the present invention optimizes the mesh of the three-dimensional model by the Laplace smoothing algorithm to remove noise and irregular geometric surfaces. The algorithm adjusts the position of each vertex based on the neighborhood relationship of the mesh vertices so that its position is smoother and consistent with the surrounding vertices, reducing sharp edges and protruding points. Hierarchical smoothing is performed according to the local area and global morphology of the model. While retaining more details in the local detail area, the overall surface is strongly smoothed to reduce the unnatural feeling of the transition area. Combined with the adaptive smoothing algorithm, the system intelligently selects the smoothing intensity according to the specific conditions of the defect area and the reconstruction area. For areas with irregular surfaces, the smoothing algorithm will moderately reduce their impact, while for flat areas or seam areas, the smoothing process is enhanced to ensure seamless transition. Through smoothness optimization, a three-dimensional digital model with a smoother surface and natural transition is obtained. After the surface of the three-dimensional digital model is smoothed, the next step is to enhance the details of the model, especially the realism of the surface features and textures. A multi-scale detail enhancement algorithm is used to restore and enhance the tiny detail features of the surface. This process mainly uses surface concave and convex information to enhance the small-scale changes on the model surface to improve the realism of the model. 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 enhancing 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.

[0021] Step S5: Visualize the optimized three-dimensional digital model, and perform real-time display and interaction to obtain three-dimensional digital model display data.

[0022] The embodiment of the present invention uses real-time rendering technology to render the three-dimensional model, presenting realistic lighting effects and texture details. By setting different light source directions, intensities and material properties, the surface, color, reflection and other characteristics of the model are made more realistic. The system supports multi-view rotation, allowing users to view the three-dimensional model from different angles. Through mouse dragging, touch operation or keyboard shortcuts, users can freely adjust the viewing angle to view various sides and details of the model. In order to improve performance, the detail level technology is adopted to reduce the model detail level when viewing from a distance and increase the detail level when viewing from a close distance. This not only ensures the display effect, but also optimizes the loading speed. Through the visualization process, high-quality three-dimensional digital model display data is generated, which supports the display of the model on different display devices and adapts to different resolutions and performance requirements. Support users to interact through operations such as clicking, dragging, and zooming. For example, the user 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 have a more intuitive and immersive experience, and can freely view the three-dimensional digital model and interact with it.

[0023] The present invention collects geometric data and surface texture data of the target object through an optical three-dimensional scanning device, and can accurately construct an initial three-dimensional digital model. This method not only improves the accuracy of modeling, but also obtains rich surface details. By identifying and marking the defective area of ​​the three-dimensional model, combined with the simulated restoration of the defective area data, the defective area can be accurately predicted, and high-quality three-dimensional model defect prediction data can be generated. This provides a reliable basis for subsequent repair and reconstruction. The defective part of the target object can be restored and reconstructed using the defect prediction data, and the integrity and authenticity of the model can be improved. Through surface smoothness optimization and detail enhancement, the visual effect of the three-dimensional model can be effectively improved, making it more delicate and realistic, and meeting the requirements of different application scenarios for model quality. The optimized three-dimensional digital model can be displayed and interacted in real time, supports the interactive operation between users and the three-dimensional model, enhances the usability and experience of the model, and is suitable for a variety of display platforms such as virtual reality and augmented reality. Through multi-platform adaptation, it is ensured that the three-dimensional model can be displayed smoothly on various devices, and the flexibility and universality of the system are improved.

[0024] Preferably, step S1 comprises the following steps: Step S11: Scanning the target object at multiple angles by using an optical three-dimensional scanning device, and collecting surface geometric information to obtain three-dimensional 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 multi-angle surface texture information of the target object through a camera device 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.

[0025] 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: Step S11: Scanning the target object at multiple angles by using an optical three-dimensional scanning device, and collecting surface geometric information to obtain three-dimensional geometric data; 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 to generate a three-dimensional point cloud. By integrating the point cloud data scanned from multiple angles, complete three-dimensional geometric data is finally obtained.

[0026] 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; The embodiment of the present invention uses a point cloud-based noise reduction algorithm to remove noise points caused 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 closest 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 viewing angles 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.

[0027] Step S13: collecting multi-angle surface texture information of the target object through a camera device to obtain surface texture data; 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 surface of the target object 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 all surfaces of the target object are covered by rotating or moving the camera device. The camera device takes 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.

[0028] 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.

[0029] The embodiment of the present invention uses texture mapping technology to accurately match the surface texture data obtained from the multi-angle camera device with the high-precision three-dimensional geometric data. In the specific operation, the pixels of the texture image are mapped to the corresponding surface of the three-dimensional model through the spatial position relationship between the camera calibration data and the three-dimensional point cloud data. A UV coordinate system is generated on the three-dimensional geometric data, and the UV coordinates serve as the basis for texture mapping. Through the expansion algorithm, the surface of the object is divided into several areas, and the texture image is correctly mapped to these areas to ensure accurate texture fit. By fusing the three-dimensional geometric data and texture information, an initial three-dimensional digital model is generated.

[0030] The present invention scans the target object at multiple angles through an optical three-dimensional scanning device, and combines the surface geometry information collection to obtain high-precision three-dimensional geometry data of the object. This method can comprehensively and accurately capture the complex geometry of the object, ensuring the accuracy and details of the model. The three-dimensional geometry 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, and reduces the complexity of post-processing. The surface texture information is collected at multiple angles through the camera device, and 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 containing the geometry and surface texture of the object. This integration 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.

[0031] Preferably, step S2 comprises the following steps: Step S21: meshing the initial three-dimensional digital model to obtain meshing data of the initial three-dimensional model; Step S22: identifying and marking the surface defective area of ​​the initial three-dimensional digital model based on the initial three-dimensional model gridding data to obtain the defective area data of the three-dimensional model; Step S23: adjusting the defect boundary precision according to the defect area data of the three-dimensional model to obtain accurate defect area data of the three-dimensional model; Step S24: extracting geometric and texture features from the defective region data of the precise three-dimensional model to obtain feature data of the defective region of the three-dimensional model; Step S25: Using the defect area feature data of the three-dimensional model, a defect area restoration simulation is performed on the initial three-dimensional digital model to generate three-dimensional model defect prediction data.

[0032] 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: Step S21: meshing the initial three-dimensional digital model to obtain meshing data of the initial three-dimensional model; The embodiment of the present invention adopts a 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 surface of the object. 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 obtained three-dimensional model mesh data includes vertex coordinates, boundary information and a mesh structure composed of facets. These data provide a basis for subsequent defect area identification, boundary adjustment and feature extraction. After meshing, the meshed data of the initial three-dimensional model is obtained.

[0033] Step S22: identifying and marking the surface defective area of ​​the initial three-dimensional digital model based on the initial three-dimensional model gridding data to obtain the defective area data of the three-dimensional model; The embodiment of the present invention analyzes the grid data of the initial three-dimensional digital model and applies an algorithm (such as point cloud-based density analysis, geometric difference detection, or texture consistency analysis) to identify surface defect areas. Defective areas usually appear as areas of surface discontinuity, missingness, or abnormality, 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.

[0034] Step S23: adjusting the defect boundary precision according to the defect area data of the three-dimensional model to obtain accurate defect area data of the three-dimensional model; The embodiment of the present invention identifies the boundary position of the defective area by analyzing the defective area data of the three-dimensional model. Then, a boundary smoothing algorithm (such as curve fitting, least squares method or weighted average method) is used to finely adjust 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. According to the boundary adjustment result, combined with the geometric features of the adjacent area, 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 defective boundary data forms accurate three-dimensional model defective area data.

[0035] Step S24: extracting geometric and texture features from the defective region data of the precise three-dimensional model to obtain feature data of the defective region of the three-dimensional model; The embodiment of the present invention extracts the shape features of the defective area by analyzing the defective area data 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 of the defective part, the surface texture, and its transition features with the surrounding intact areas. Detailed three-dimensional model defective area feature data is generated by extracting geometric and texture features.

[0036] Step S25: Using the defect area feature data of the three-dimensional model, a defect area restoration simulation is performed on the initial three-dimensional digital model to generate three-dimensional model defect prediction data.

[0037] The embodiment of the present invention uses 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 three-dimensional model. By analyzing the spatial structure and surface texture of the defective area, a suitable algorithm is selected to supplement the defective area. Based on the known complete area data, a similarity matching method or inverse modeling technology is used to generate a reasonable geometric shape in the defective area, and the defective part is filled according to the texture features of the surrounding area to form complete prediction data. After simulation and restoration processing, the three-dimensional model defect prediction data is generated.

[0038] The present invention can convert complex three-dimensional data into more operable 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 defect area identification and repair. Defective area identification and marking based on grid data can effectively locate the defective part 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 necessary detail 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 part can be accurately predicted and supplemented to ensure the integrity and realism of the three-dimensional model after repair.

[0039] Preferably, step S25 comprises 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; The embodiment of the present invention uses the feature data of the defective area of ​​the three-dimensional model as the input for constructing the interpolation model. Using these feature data, combined with the neighborhood information, the spatial structure and geometric morphology 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 features of the defective area. Through the interpolation algorithm, a preliminary geometric morphology 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 morphology and structure of the defective part to ensure geometric consistency with the surrounding area. The preliminary defective area interpolation model obtained by this step.

[0040] 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; The embodiment of the present invention takes the preliminary defective area interpolation model and the precise three-dimensional model defective 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, a shape matching technique 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 defective 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 defective area and the adjacent area to avoid morphological breaks or abrupt changes. Finally, the optimized defective area geometric morphology data is obtained.

[0041] Step S253: reconstructing the texture of the defective region according to the feature data of the defective region of the three-dimensional model and the optimized geometric data of the defective region to obtain texture restoration data of the defective region; The embodiment of the present invention inputs optimized defect area geometric morphology data and three-dimensional model defect area feature data. Texture mapping technology is used to accurately apply texture data to the geometric surface of the defect 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 defect area. In the texture reconstruction process, considering the transition between the texture of the defect 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 defect area is obtained.

[0042] Step S254: Comprehensively verify the optimized defect area geometry data and the defect area texture restoration data to obtain defect area restoration data; The embodiment of the present invention uses 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 is smoothly connected 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 results of geometric and texture verification, the existing deviations are automatically or manually adjusted. The geometry and texture mapping are readjusted based on the characteristics of the surrounding area, and accurate defect area restoration data is finally generated.

[0043] Step S255: Fusing the defect area restoration data with the initial three-dimensional digital model to obtain a defect prediction three-dimensional model; 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 shape 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 transition. 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 finally a defect prediction three-dimensional model is generated.

[0044] Step S256: Use the defect prediction three-dimensional model to perform defect area restoration simulation to obtain three-dimensional model defect prediction data.

[0045] The embodiment of the present invention determines the parameters of the restoration simulation according to the actual application scenario of the target object, including physical characteristics, geometric accuracy requirements and texture feature details. The geometric and texture fitting algorithms are used to merge the defect area feature data in the defect prediction three-dimensional model with the complete model to simulate and restore the final shape of the defect area. The consistency of the model is optimized by a step-by-step iteration method. Error analysis is performed on the restored three-dimensional model defect prediction data to ensure that the geometric shape and texture details are highly matched with the original data. The three-dimensional model defect prediction data containing the complete geometric shape and surface texture is output.

[0046] The present invention can effectively fill the defective part in the three-dimensional model by constructing a preliminary defective area interpolation model and optimizing the geometric morphology, while maintaining consistency with the surrounding area. This ensures that the repaired three-dimensional model is more natural in structure and avoids errors and incoherence in the manual repair process. Texture reconstruction using defective area feature data and 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 can achieve 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 data conflicts or errors affecting the quality of the final model. By fusing the defective area restoration data with the initial three-dimensional digital model, the defective area can be accurately restored, and a high-quality defect prediction three-dimensional model can be finally generated. 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 the repair efficiency and reduce the loss of original data during the repair process.

[0047] Preferably, step S3 comprises 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; The embodiment of the present invention extracts significant geometric feature points from the three-dimensional model defect prediction data and the initial three-dimensional digital model to ensure that the basic points of alignment have high precision. An alignment algorithm based on iterative nearest point is used to adjust the poses of the two sets of data so that they are accurately aligned in space. A global error analysis is performed on the aligned model, the residual distribution is calculated, and the areas with deviations are adjusted to ensure the consistency of geometric shape and spatial position. The two sets of aligned data are fused to generate a complete three-dimensional digital defect prediction combined model.

[0048] Step S32: Generate a geometric grid of the defect area according to the three-dimensional digital defect prediction combined model, and fill the defect area to generate a three-dimensional digital model with a continuous grid structure; 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 geometry 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.

[0049] 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; The embodiment of the present invention extracts texture features of the defective area from the defect prediction data of the three-dimensional model, including color, gloss, texture pattern, etc. The texture information of the defective area is accurately projected to the corresponding position of the three-dimensional model of the continuous grid structure by using the extracted texture features through the texture mapping algorithm. The texture boundary is smoothed to avoid discontinuity or abruptness at the texture connection to ensure the natural transition of the overall texture effect. The texture details of the defective area are enhanced to simulate the texture characteristics of the real object, such as improving the resolution of the texture, adjusting the light and shadow effects, etc. After the texture effect is optimized, a repaired three-dimensional digital model is generated.

[0050] Step S34: performing geometry and texture consistency checks on the repaired three-dimensional digital model, and optimizing the repaired three-dimensional digital model based on the consistency check results to obtain a reconstructed three-dimensional digital model.

[0051] 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 matching degree 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, and the color, gloss and pattern connection of the texture are verified to be 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-tune the areas where the textures do 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, a reconstructed three-dimensional digital model is generated.

[0052] The present invention can accurately identify and locate the defect area by geometrically aligning the defect prediction data of the three-dimensional model with the initial three-dimensional digital model. The subsequent geometric grid generation and filling process ensures that the defect area is accurately restored, and the generated continuous grid structure three-dimensional digital model is more complete and natural, avoiding the improper connection or unnatural phenomenon in manual restoration. The texture effect optimization of the repaired three-dimensional model using the defect prediction data can effectively restore the visual effect of the defect area, ensure that the texture style of the repaired part is consistent with that of the original model, thereby improving the overall visual realism and integrity. By performing geometric and texture consistency checks on the repaired three-dimensional digital model, the inconsistency or error 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 geometric morphology 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.

[0053] Preferably, step S4 comprises the following steps: Step S41: Optimizing the geometric mesh of the reconstructed three-dimensional digital model to obtain a mesh-optimized three-dimensional digital model; 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 the mesh size, density, shape, and whether there is overlap or staggering. For areas where the mesh distribution is uneven, a mesh reconstruction algorithm is used to re-divide the mesh so that its density is uniform and the shape is regular. Redundant meshes are streamlined, unnecessary vertices or edges are deleted, and the model complexity 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, and the overall topological stability of the model is enhanced. 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.

[0054] Step S42: Surface optimization is performed on the mesh optimized three-dimensional digital model, and noise is removed to obtain a surface smooth optimized three-dimensional digital model; 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 the surface noise area. The Laplace smoothing algorithm or the bidirectional curvature constraint algorithm is used to adjust the surface vertex position, eliminate small-range fluctuations, and achieve smoothing of the surface transition. 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 consistency of the entire model surface is optimized to ensure that the smoothness of each part is coordinated and to avoid excessive or insufficient local smoothing. Finally, a surface smoothed and optimized three-dimensional digital model is obtained.

[0055] Step S43: performing detail restoration on the surface smoothing optimized three-dimensional digital model to obtain a detail enhanced three-dimensional digital model; 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 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 amplification algorithm is used to enhance the surface details of the model. Using detail mapping technology, the enhanced detail information is superimposed on the surface of the original model 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, 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.

[0056] Step S44: performing a global quality check on the detail-enhanced three-dimensional digital model to obtain three-dimensional digital model repair parameters; 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 geometry, texture consistency, and mesh quality of the model. The mesh quality of the model is assessed, including the shape of the triangular facets, the distribution of the 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 the surface is missing or abnormally protruding 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, the 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.

[0057] 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.

[0058] The embodiment of the present invention uses three-dimensional digital model repair parameters to make precise adjustments to problem areas in detail-enhanced three-dimensional digital models. A geometric repair algorithm is applied to make adjustments to geometric inconsistencies or defects found during 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, patch 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 effect 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.

[0059] The present invention optimizes the mesh structure of the three-dimensional digital model through geometric mesh optimization, reduces redundant parts and irregular shapes in the model, enhances the fineness and accuracy of the mesh, thereby ensuring higher geometric accuracy and laying a solid foundation for subsequent processing. The steps of surface optimization and noise removal 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 and avoids 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 expressing complex surface features or tiny details, it can better preserve the true appearance of the object and enhance the visual effect and perceived quality. The global quality check 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 geometric structure and surface texture of the model. This step helps to improve the stability and adaptability of the model and ensure that it can meet the needs of different application scenarios. Through adjustment and optimization based on restoration parameters, the final 3D digital model is not only improved in geometry and surface quality, but also has higher accuracy and visual consistency, making it suitable for high-demand application scenarios such as virtual reality, medical imaging, industrial inspection, etc.

[0060] Preferably, step S42 includes the following steps: Step S421: monitoring and marking the model surface noise points on the mesh-optimized three-dimensional digital model to obtain three-dimensional model noise point marking data; The embodiment of the present invention clarifies the definition of noise points. Usually, noise points appear as abnormal vertices or triangular faces 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 inconsistent with the neighborhood structure, have high local geometric fluctuations, or they do not match the topological structure of the surrounding area. Once potential noise points are detected, they are 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 finally form the three-dimensional model noise point marking data.

[0061] Step S422: removing noise points from the mesh-optimized three-dimensional digital model according to the three-dimensional model noise point marking data to obtain a noise-reduced three-dimensional digital model; The embodiment of the present invention utilizes the noise point marking 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, which are usually manifested as abnormal connection with the surrounding grid points or geometric mutations. For each noise point, the position of the noise point is corrected by calculating the geometric features of other points in its neighborhood and using this as a reference. Usually, weighted average or least squares method is 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 their surrounding areas. According to the noise point marking 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 is removed, the three-dimensional model grid structure is updated to obtain a denoised three-dimensional digital model.

[0062] Step S423: Optimizing the surface smoothness of the denoised three-dimensional digital model to obtain a preliminary smooth three-dimensional digital model; The embodiment of the present invention further optimizes the model surface on the basis of the denoised three-dimensional digital model 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 the weighted average method to reduce the surface high-frequency noise. Usually, through multiple iterations, the grid surface is made 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 smoothness 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 edge and detail features to avoid excessive smoothing resulting in loss of geometric features. After completion, a preliminary smoothed three-dimensional digital model is obtained.

[0063] Step S424: feature enhancement is performed on the preliminary smooth three-dimensional digital model to obtain a surface smoothed optimized three-dimensional digital model.

[0064] The embodiment of the present invention further restores and strengthens important geometric features, such as edges, corners, and uneven details, on the basis of the three-dimensional digital model after preliminary smoothing, so that it is more consistent with the shape and structure of the real object. A high-pass filter or sharpening algorithm is used to restore details on the smooth surface, and the details of the model are made clearer by highlighting the edge and texture information. Weighting is performed according to the local geometric features of the model surface to enhance the details of certain areas. For example, during the smoothing process, the key edges or parts with prominent structures are specifically enhanced to avoid excessive smoothing and loss of details. The surface fitting technology is used to supplement the details of the smooth area of ​​the model, reconstruct the complex geometric form, and enhance the detailed expression of the surface of the object. The surface features are enhanced by enhancing the texture details to make it more natural and realistic, especially for objects with textures or uneven surfaces. The local area features of the initially smoothed three-dimensional model are enhanced to retain details and strengthen edges and important structures. An appropriate enhancement algorithm is used to ensure that the geometric features of the model are restored without affecting the overall smoothing effect. Finally, a surface smoothed and optimized three-dimensional digital model is obtained.

[0065] The present invention monitors and marks the noise points on the model surface, and the system can effectively identify and remove unnecessary noise, thereby obtaining a de-noised three-dimensional digital model. The noise removal process greatly improves the accuracy of the model and avoids the influence of noise on subsequent processing or visualization effects. The surface smoothness of the three-dimensional digital model after noise reduction is optimized to ensure that the model surface is smoother. Optimizing the surface smoothness can reduce the irregularity of the model surface, make the surface more uniform, improve the visual effect and the accuracy of the model, especially in occasions where details need to be displayed. By performing feature enhancement on the three-dimensional digital model after preliminary smoothing, the key details and texture features of the model surface are enhanced. Feature enhancement can highlight the important structure and details of the model, so that the three-dimensional model is improved in terms of fineness and authenticity, and is particularly suitable for fields that require high-detail performance, such as industrial detection, virtual reality, etc. The whole process gradually improves the quality of the three-dimensional model through the continuous steps of noise removal, surface optimization and feature enhancement, ensuring that its details are more realistic, the structure is more accurate, the adaptability is stronger, and it can meet high-demand application scenarios.

[0066] Preferably, step S43 includes the following steps: Step S431: identifying key detail areas of the surface smoothed and optimized three-dimensional digital model to obtain three-dimensional model detail area marking data; The embodiment of the present invention identifies and marks detail areas in a surface smoothing optimized three-dimensional digital model that are critical to the final visual effect or function. By calculating the geometric features of the curvature, normal direction, edge sharpness, etc. of the three-dimensional model surface, the area where the surface changes more significantly is identified. For example, a higher curvature value usually indicates that the area is a detail area, such as a sharp edge, a hole, or a protrusion. Based on different surface features, an adaptive algorithm is used to dynamically adjust the recognition criteria. For the geometric structure 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 fidelity of the model. Using a trained deep learning model, by analyzing the surface features of the three-dimensional model, key detail areas are automatically identified and marked. The key detail areas identified and marked by the above algorithm form detail area marking data of the three-dimensional model.

[0067] Step S432: geometrically reconstructing the surface smoothed optimized three-dimensional digital model according to the three-dimensional model detail area marking data to obtain a reconstructed detail area three-dimensional digital model; The embodiment of the present invention reconstructs the geometric morphology according to 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 the missing or incomplete geometric information through the interpolation algorithm. Apply shape recovery technology to the detail area, such as shape generation from the boundary 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 the information of the adjacent area to perform local surface fitting, so as to obtain a more accurate geometric morphology. According to the curvature or other geometric properties of the object surface, the reconstruction process is adaptively adjusted. In view of the features of different scales in the detail area, a multi-scale geometric analysis method is used to ensure that the reconstruction process simultaneously considers local details and global geometric consistency to avoid detail loss or mismatch. Through geometric morphology reconstruction, a reconstructed three-dimensional digital model of the detail area containing detail recovery information is obtained.

[0068] Step S433: amplifying the high-frequency features of the detail region of the reconstructed three-dimensional digital model, and enhancing the clarity of the local features to obtain a detail-enhanced three-dimensional digital model.

[0069] The embodiment of the present invention is based on reconstructing a three-dimensional digital model of a detail area, and further optimizes the detail area by using high-frequency feature amplification and local feature clarity enhancement technology to improve the detail expression and clarity of the model. Through frequency domain analysis, high-frequency feature amplification is applied to the detail area of ​​the three-dimensional model. The discrete Fourier transform is used to enhance the high-frequency components in the detail area to make these details more prominent visually. Combined with the curvature and surface changes of the three-dimensional grid, the high-frequency features of the detail area are locally amplified to enhance the display effect of small surface changes, especially the complex parts of texture and structure. A shape-preserving filter is applied to enhance the local features of the detail area, such as concave-convex textures, surface edges, etc., to remove the blur effect and ensure that every detail is clearer and more discernible. By adjusting the parameters, it is ensured that the detail area is neither too smooth to lose the sense of reality, nor the influence of noise can be avoided, thereby improving the clarity and sense of reality of the local details. After high-frequency feature amplification and local clarity enhancement, a detail-enhanced three-dimensional digital model is obtained.

[0070] The present invention identifies key detail areas of the three-dimensional digital model after surface smoothing optimization, and the system can accurately mark the detail areas that need further processing. This process ensures that the subsequent optimization and enhancement work can be concentrated on the important parts, improves efficiency and reduces unnecessary processing. The geometric morphology of the three-dimensional model is reconstructed according to the detail area marking data, so that the morphology of the detail area is more refined and realistic. The reconstructed three-dimensional digital model of the detail area is more consistent with the actual shape and characteristics of the object, and the accuracy and meticulousness of the model are enhanced. The high-frequency feature amplification and local feature clarity enhancement of the reconstructed detail area can effectively improve the expressiveness and visual effect of the details. This processing makes the details more prominent, and the model shows higher accuracy and realism, which is particularly suitable for fields that require fine display, such as digital cultural relics protection, industrial design and virtual reality. Overall, through the identification, geometric reconstruction and feature enhancement of key detail areas, the method can significantly improve the accuracy and visual effect of the three-dimensional model in detail presentation, especially for those application scenarios that need to highlight details and textures, the effect is particularly prominent.

[0071] Preferably, step S5 comprises the following steps: Step S51: converting the optimized three-dimensional digital model into a new format, and compressing and optimizing the format conversion result to obtain real-time rendering three-dimensional model data; 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 attributes 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 size of the file 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, output real-time rendering three-dimensional model data.

[0072] Step S52: performing interactive parameter configuration based on the real-time rendering three-dimensional model data to obtain real-time rendering display configuration data; The embodiment of the present invention selects a suitable interaction method according to the display platform and user needs. Determine the interactive functions that need to be configured. Set operable viewing angle 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. According to the interactive requirements, optimize the rendering process 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.

[0073] Step S53: Design interactive functions according to the real-time rendering display configuration data to obtain interactive display application data; 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 users can display 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 you click on an object, you can display the object's relevant data or a rotation animation. Provides immediate visual and auditory feedback. For example, when a 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.

[0074] 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.

[0075] The embodiment of the present invention imports interactive display application data into the target display platform. This data includes a three-dimensional model, interactive 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, light and shadow 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 power and user interaction mode of different platforms. Through adaptive technology, adjust the size, display mode and control mode of the three-dimensional model and interactive elements. Compress and simplify the model data, reduce the rendering load, and improve the loading speed and fluency. 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 on different devices is consistent and the user interaction is smooth. Adjust the display effect according to the test feedback, fix compatibility or performance issues, and ensure smooth operation of multiple platforms. Finally, obtain three-dimensional digital model display data that adapts to multiple platforms.

[0076] The present invention converts the format of the optimized three-dimensional digital model and optimizes the compression, so that 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, so that users can adjust the 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 model display. By designing interactive functions according to 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 into 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, a mobile device, or a virtual reality platform, the system can provide a stable and consistent display effect, improving the universality and cross-platform compatibility of the system.

[0077] 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: 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; use the 3D geometric data and surface texture data to construct a 3D model 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.

[0078] The present invention scans the target object through an optical three-dimensional scanning device, and the system can accurately obtain three-dimensional geometric data and surface texture data to ensure the high-precision construction of the initial three-dimensional digital model. This lays a solid foundation for subsequent modeling, repair and optimization. By identifying and marking the defective area of ​​the initial three-dimensional model and predicting and restoring it based on the defective area data, the system can accurately identify the defective part in the model and simulate the characteristics of the defective area. This function can greatly improve the accuracy and efficiency of model repair. Using the defect prediction data, the system accurately restores and reconstructs the initial three-dimensional digital model, thereby restoring the morphology and details of the defective area, ensuring that the final three-dimensional digital model is more complete and realistic. By optimizing the surface smoothness and enhancing the details of the reconstructed three-dimensional model, the system can improve the visual effect and detail expression of the model, further improve the accuracy and authenticity of the three-dimensional model, and make it more expressive during the display process. The system provides real-time display and interaction functions, allowing users to view and operate the three-dimensional model intuitively and flexibly. Through this function, users can not only observe the details of the model in depth, but also interact from different perspectives, enhancing the user experience and the flexibility of model display.

[0079] 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, so it is intended to include all changes within the meaning and scope of the equivalent elements of the application documents in the present invention.

[0080] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be 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 will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for realizing a three-dimensional optical system, characterized in that: The following steps are involved: Step S1: Scanning the target object by 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: identifying and marking defective areas of the initial three-dimensional digital model to obtain defective area data of the three-dimensional model; performing simulation restoration based on the defective area data of the three-dimensional model to generate defect prediction data of the three-dimensional model; 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, characterized in that: Step S1 includes the following steps: Step S11: Scanning the target object at multiple angles by using an optical three-dimensional scanning device, and collecting surface geometric information to obtain three-dimensional 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 multi-angle surface texture information of the target object through a camera device 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, characterized in that: Step S2 includes the following steps: Step S21: meshing the initial three-dimensional digital model to obtain meshing data of the initial three-dimensional model; Step S22: identifying and marking the surface defective area of ​​the initial three-dimensional digital model based on the initial three-dimensional model gridding data to obtain the defective area data of the three-dimensional model; Step S23: adjusting the defect boundary precision according to the defect area data of the three-dimensional model to obtain accurate defect area data of the three-dimensional model; Step S24: extracting geometric and texture features from the defective region data of the precise three-dimensional model to obtain feature data of the defective region of the three-dimensional model; Step S25: Using the defect area feature data of the three-dimensional model, a defect area restoration simulation is performed on the initial three-dimensional digital model to generate three-dimensional model defect prediction data.

4. The method for realizing a three-dimensional optical system according to claim 3, characterized in that: 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 region according to the feature data of the defective region of the three-dimensional model and the optimized geometric data of the defective region to obtain texture restoration data of the defective region; Step S254: Comprehensively verify the optimized defect area geometry data and the defect area texture restoration data to obtain defect area restoration data; Step S255: Fusing the defect area restoration data with the initial three-dimensional digital model to obtain a defect prediction three-dimensional model; Step S256: Use the defect prediction three-dimensional model to perform defect area restoration simulation to obtain three-dimensional model defect prediction data.

5. The method for realizing a three-dimensional optical system according to claim 1, characterized in that: 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: Generate a geometric grid of the defect area according to the three-dimensional digital defect prediction combined model, and fill 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 checks on the repaired three-dimensional digital model, and optimizing the repaired three-dimensional digital model based on the consistency check results to obtain a reconstructed three-dimensional digital model.

6. The method for realizing a three-dimensional optical system according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Optimizing the geometric mesh of the reconstructed three-dimensional digital model to obtain a mesh-optimized three-dimensional digital model; Step S42: Surface optimization is performed on the mesh optimized three-dimensional digital model, and noise is removed to obtain a surface smooth optimized three-dimensional digital model; Step S43: performing detail restoration on the surface smoothing 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 repair 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.

7. The method for realizing a three-dimensional optical system according to claim 6, characterized in that: Step S42 includes the following steps: Step S421: monitoring and marking the model surface noise points on 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 three-dimensional digital model according to the three-dimensional model noise point marking data to obtain a noise-reduced three-dimensional 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: feature enhancement is performed on the preliminary smooth three-dimensional digital model to obtain a surface smoothed optimized three-dimensional digital model.

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

9. The method for realizing a three-dimensional optical system according to claim 1, characterized in that: Step S5 includes the following steps: Step S51: converting the optimized three-dimensional digital model into a new format, and compressing and optimizing the format conversion result to obtain real-time rendering three-dimensional model data; Step S52: performing interactive parameter configuration based on the real-time rendering three-dimensional 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.

10. 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; use the 3D geometric data and surface texture data to construct a 3D model 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