A global nurbs parameter domain based spatial intelligent world modeling method and system

Abstract

The application belongs to the technical field of space intelligent modeling, and relates to a space intelligent world modeling method based on a global NURBS parameter domain, comprising the following steps: data input and data preprocessing; consistent mapping and management of multi-Patch geometric objects under the global NURBS parameter domain; generation and optimization of a control point matrix and a weight matrix for geometric expression based on the global NURBS parameter domain; automatic smooth transition of multi-Patch geometry in a spliced area is realized based on a continuity maintenance mechanism of curvature and normal constraint; NURBS surface subdivision operation is performed according to a curvature change rate and a geometric error threshold; global solving is performed on the control point matrix of all Patch, and a space intelligent world model that is continuously differentiable and has consistent parameters in a global range is output. The method significantly improves the continuity and controllability of geometric modeling. The application also provides a space intelligent world modeling system based on the global NURBS parameter domain.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of spatial intelligence and three-dimensional spatial intelligence modeling, and is applied to the industries of digital twinning, virtual reality, smart city and generative spatial understanding, and specifically relates to a spatial intelligence world modeling method based on a global NURBS parameter domain. BACKGROUND

[0002] At present, the three-dimensional world modeling of spatial intelligence systems mainly relies on discrete geometric expression forms such as mesh, voxel or implicit radiance field (NeRF). Although these methods have made certain progress in local scene reconstruction and visual rendering, they still have obvious limitations in large-scale, city-level and continuous spatial expression.

[0003] In recent years, generative artificial intelligence models such as NeRF and Marble model of LiFei-Fei team have been able to learn the implicit geometric structure of the world through neural networks, realizing the generation of three-dimensional spatial representation from multi-view or text input. However, the underlying geometric primitives of such models (such as voxel, point cloud or Gaussian sphere) are still discrete or probabilistic distribution-based expression, which has the following main problems in terms of continuous differentiability, geometric controllability and global consistency:

[0004] 1. Insufficient geometric continuity:

[0005] Current expressions based on volume density or Gaussian primitives can only achieve surface smoothness in a probabilistic sense, lack strict geometric continuity constraints, and are difficult to achieve G0 / G1 / G2 level controllable continuity, limiting the application of models in precise scenarios such as city modeling.

[0006] 2. Lack of global parameter uniformity:

[0007] Existing neural fields and Gaussian expressions rely on local coordinates or independent patch representations, which cannot establish a global unified parameter domain mapping, leading to scale drift and inconsistent deformation problems when large scenes or cross-region splicing are performed.

[0008] 3. Unclear learnable geometric parameters:

[0009] In Mesh or Gaussian Splatting expression, the geometric morphology is often implicit in the network weights, lacking explicit learnable geometric control parameters, which is not conducive to the explainable and adjustable spatial learning of AI models.

[0010] 4. Insufficient dynamic adaptive capability:

[0011] Existing models cannot achieve adaptive subdivision and reconstruction under different resolutions, viewing angles or task conditions, rendering and geometry updating still need manual resampling, and there is a lack of intelligent subdivision mechanism based on error tolerance.

[0012] 5. Smooth splicing of multi-patch scenes is difficult:

[0013] For city-level, complex building groups or natural terrain scenes, existing methods are difficult to maintain boundary continuity among multiple local models, and normal inconsistency, surface tearing or occlusion gaps often occur at the splicing place.

[0014] Therefore, although the existing spatial intelligent world modeling technology has made significant progress in visual perception and generative rendering, there are still obvious bottlenecks in continuous geometric expression, global parameter consistency, learnable controllability and adaptive generation capability. There is an urgent need for a geometric modeling method based on a global continuous parameter space, which can use a unified NURBS parameter domain as the underlying logic to realize continuous expression, curvature preservation, adaptive subdivision and AI-driven generation optimization of multi-patch world, so as to build a truly spatially intelligent world model (Spatially Intelligent World Model). SUMMARY

[0015] The present application provides a spatially intelligent world modeling method and system based on a global NURBS parameter domain, which is used to solve the problems of discrete geometric expression, insufficient continuity and non-uniform parameter domain in existing spatial intelligent modeling technology.

[0016] The present application adopts the following technical solutions:

[0017] A spatially intelligent world modeling method based on a global NURBS parameter domain, comprising:

[0018] Data input and data preprocessing;

[0019] Constructing a global NURBS parameter domain for spatially intelligent world modeling, and uniformly mapping and managing multi-patch geometric objects under the global NURBS parameter domain;

[0020] Generating and optimizing a control point matrix for geometric expression based on the global NURBS parameter domain;

[0021] A continuity preservation mechanism based on curvature and normal constraints is used to realize automatic smooth transition of multi-patch geometry in the splicing area;

[0022] Based on the adaptive subdivision and error control mechanism, NURBS surface subdivision operation is performed according to the curvature change rate and geometric error threshold;

[0023] Based on the global consistency optimization solution mechanism, the global solution is performed on the control point matrix of all patches, and the output is a spatial intelligent world model that is continuously differentiable and has consistent parameters in the global scope.

[0024] Preferably, the method further includes:

[0025] Based on a multi-scale hierarchical modeling and fusion mechanism, curvature and topological continuity are maintained between different levels through nested parameter domain mapping relationships.

[0026] Preferably, the global NURBS parameter domain construction process includes:

[0027] The point cloud data is normalized in the world coordinate system.

[0028] A two-dimensional parameter domain is established in the normalized space, and the point cloud coordinates are converted into parameter domain coordinates through an affine mapping function.

[0029] Automatically generate non-uniform node vectors based on point cloud density and curvature characteristics;

[0030] Based on the point cloud features, the parameter domain is divided into multiple patch regions, and a global parameter index table is established.

[0031] Preferably, generating and optimizing the control point matrix for geometric representation based on the global NURBS parameter domain includes:

[0032] In the global NURBS parameter domain, a control point parameter grid is generated according to a preset resolution or adaptive rules. Based on the mapping relationship between the parameter domain and spatial coordinates, the control points in the parameter domain are mapped to the corresponding spatial control point coordinates, thereby forming an initial control point matrix.

[0033] Perform global collaborative optimization on the initial control point matrix.

[0034] Preferably, the control point matrix, as an optimization variable, is subject to the combined effects of the following constraints:

[0035] Parameter domain continuity constraints are used to ensure that the boundary control point parameter positions of adjacent patches remain consistent in the global parameter domain.

[0036] Geometric smoothing constraints are used to suppress geometric abrupt changes caused by excessively dense or sparse distribution of local control points;

[0037] Error tolerance constraints are used to limit the adjustment range of control points so that the geometric accuracy meets the preset requirements.

[0038] Preferably, the mechanism for maintaining the continuity of curvature and normal constraints includes:

[0039] At the stitching boundary of adjacent patches, the corresponding curvature information and normal information are calculated based on the generated control point matrix, and the curvature difference and normal deviation at the boundary of adjacent patches are taken as continuity evaluation indicators; when the curvature difference or normal deviation is detected to exceed a preset threshold, a continuity optimization process is automatically triggered;

[0040] In the continuity optimization process, the control point coordinates in the control point matrix are taken as optimization variables, and the control point positions are iteratively adjusted by simultaneously constraining the curvature change and normal change at the boundary of adjacent patches.

[0041] Preferably, the NURBS surface subdivision operation according to the curvature change rate and the geometric error threshold comprises:

[0042] When the local curvature exceeds the threshold or the geometric error accumulation exceeds the threshold, a new control point layer is automatically inserted, and the NURBS surface is hierarchically refined.

[0043] Preferably, the multi-scale hierarchical modeling and fusion mechanism comprises:

[0044] A hierarchical NURBS expression system is adopted: the upper layer patch is responsible for expressing the city-level macro structure; the middle layer patch is responsible for the terrain and building group contour; and the lower layer patch describes the local geometric details; the curvature and topological continuity between layers are maintained through nested parameter domain mapping relationship.

[0045] A spatial intelligent world modeling system based on a global NURBS parameter domain comprises:

[0046] A parameter domain management module is configured to establish and maintain the mapping relationship of the global NURBS parameter domain;

[0047] A control point generation and optimization module is configured to automatically generate an initial control point matrix and a weight matrix according to input point cloud, curvature and semantic features;

[0048] A continuity constraint module is configured to detect and constrain the normal and curvature difference of adjacent patches to realize automatic continuity maintenance at the geometric boundary;

[0049] A subdivision and error control module is configured to perform an adaptive subdivision process based on error tolerance; when the local curvature or error exceeds a set threshold, the system automatically inserts a new control point and regenerates a node vector;

[0050] A global optimization and model output module is configured to uniformly solve and parameter coordinate the control point matrix of all patches and output the model.

[0051] Preferably, the system further comprises:

[0052] Multi-scale synchronization module: used to maintain parameter consistency and geometric smoothness between different levels, so that each scale model maintains nested continuity in topology and curvature.

[0053] The summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. The summary is not intended to identify key or essential features of the disclosure, nor is it intended to limit the scope of the disclosure.

[0054] Compared with existing spatial intelligent modeling techniques, the present application has obvious technical advantages and comprehensive effects in geometric continuity, parameter consistency, generation efficiency and system scalability, including:

[0055] 1. In terms of geometric continuity, the present application defines a multi-Patch geometric structure in a unified global NURBS parameter domain, and uses a curvature and normal constraint continuity optimization mechanism to realize automatic splicing of boundaries, so that adjacent surfaces are smoothly connected in the normal and curvature directions. This mechanism can realize G1 to G2 level surface continuity without human intervention, effectively eliminating the problems of fold lines, cracks and uneven surfaces at the splicing place of traditional mesh models, thereby realizing global consistent geometric expression in large scenes.

[0056] 2. In terms of modeling and optimization efficiency, the present application uses a parameterized control point matrix as the core representation of geometry, reducing the dependence on high-density point cloud or voxel data, making the modeling process more lightweight. Since the control points and weight parameters can be optimized in the form of analytical gradient, the present application is significantly superior to the implicit field method based on sampling in terms of iterative convergence speed and computational stability, and the overall modeling efficiency can be improved by about 30%, enabling continuous geometry generation and global optimization of city-level or complex scenes in a short time.

[0057] 3. In terms of geometric precision and structural expression capability, the global parameter domain structure of the present application can accurately control the surface shape, directly constrain the curvature distribution through the control point matrix, so that the generated results are superior to traditional methods in terms of surface smoothness, structural alignment and shape rationality. Compared with discrete models based on meshes, the world geometry generated by the present application is more consistent in continuity and differentiability, and is suitable for spatial intelligent scenes that require high-precision modeling and subsequent physical simulation.

[0058] 4. In terms of system structure and engineering application, the parameterized expression system of the present application is highly compatible with existing BIM modeling systems. The generated NURBS control point matrix can be directly exported as a standard industry format file, facilitating the connection with architectural design, structural analysis and digital twin platforms. This feature makes the generation and optimization process of spatial models more automated, without the need for complex post-processing operations, greatly reducing the cost and time overhead of manual modeling.

[0059] The method of the present application not only significantly improves the continuity and controllability of geometric modeling, but also has significant advantages in efficiency, compatibility and learnability. Compared with the prior art, the present application shows obvious technical progress and engineering application value in structure simplification, calculation acceleration, modeling automation and resource conservation. BRIEF DESCRIPTION OF DRAWINGS

[0060] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the several views.

[0061] Figure 1 A global NURBS parameter domain-based spatial intelligent world modeling method flowchart for an embodiment of the present application;

[0062] Figure 2 A global NURBS parameter domain-based urban-level three-dimensional scene modeling flowchart for an embodiment of the present application;

[0063] Figure 3 A global parameter domain definition and node vector construction flowchart provided for an embodiment of the present application;

[0064] Figure 4 A NURBS surface generation and geometric derivation module diagram provided for an embodiment of the present application;

[0065] Figure 5 A global NURBS parameter domain-based spatial intelligent world modeling system block diagram for an embodiment of the present application. DETAILED DESCRIPTION

[0066] Embodiments of the present disclosure will be described more fully hereinafter with reference to the accompanying drawings. While embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the present disclosure to those skilled in the art.

[0067] The term "comprising" and variations thereof as used herein are intended to cover a non-exclusive inclusion, i.e., "including, but not limited to". Unless specifically stated, the term "or" means "and / or". The term "based on" means "based, at least in part, on". The terms "one example embodiment" and "an embodiment" mean "at least one example embodiment". The term "another embodiment" means "at least one additional embodiment". The terms "first", "second", etc. can refer to different or same objects. Other explicit and implicit definitions can also be included below.

[0068] The present application takes a global NURBS parametric domain as the underlying logic of geometric expression, organizes multi-Patch geometry structures under a unified global parameter space, and realizes continuous world-level geometry generation and optimization. The core idea of the whole method is to form an end-to-end spatial intelligent geometry modeling framework through parameter domain mapping, control point matrix constraint, curvature continuity preservation, adaptive subdivision optimization and global consistent solving, so that the spatial intelligent system can perform geometry optimization and generation in a unified parameter space, thereby establishing a world modeling framework with geometric continuity, AI learnability and dynamic adaptability.

[0069] In order to facilitate understanding, first, the technical terms involved in the embodiments of the present application are simply introduced:

[0070] NURBS: is the abbreviation of English Non-Uniform Rational B-Spline (i.e. non-uniform rational B-spline). NURBS is a method for accurately describing smooth curves and surfaces by mathematical formula (through control points and weights).

[0071] Patch: In geometry modeling, it is usually translated as surface patch or patch. Patch is the basic building unit of complex surface, which is a smooth and continuous surface segment defined by mathematical formula.

[0072] Control point: is the core parameter of defining NURBS geometry. They constitute a point set in space, and generate accurate mathematical surfaces through linear combination of weighted B-spline basis functions. Each control point contains three-dimensional coordinates and a weight value: coordinates determine the spatial position, and weight as a rational factor adjusts the "attractive force" strength of the point to the surface. This structure enables NURBS to accurately express free-form surfaces and conic curves (such as circular arcs) with a unified mathematical model. Control points are usually not located on the surface, but form an external control polygon or control grid. Moving control points can smoothly and predictably drive the local or overall deformation of the surface.

[0073] Control point matrix: is the core data structure in the mathematical definition of NURBS surface, which organizes all control points of the surface in u, v two parameter directions in a two-dimensional array form. The topology of this matrix directly determines the control mesh form of the NURBS surface: each row of the matrix corresponds to a sequence of control points under a u-direction isoparametric line, and each column corresponds to a control point sequence in the v-direction. This matrix arrangement not only provides a clear parameterized coordinate frame for the surface, making any surface point can be calculated through the corresponding basis function and control point subset indexed by (u, v) parameters, but also realizes the geometric basis of advanced operations on the surface (such as trimming, merging, and continuity matching). In industrial CAD systems, manipulating the control point matrix, i.e. adjusting the position and weight of its elements (control points) through interactive tools, becomes an efficient operation paradigm for editing the shape of the surface, which ensures that the entire surface can strictly maintain its mathematical smoothness and geometric continuity even after local modification.

[0074] Embodiment 1:

[0075] A global NURBS parameter domain-based spatial intelligent world modeling method is introduced below.

[0076] A global NURBS parameter domain-based spatial intelligent world modeling method, as shown in Figure 1 , comprises:

[0077] S100, data input and data preprocessing.

[0078] In one embodiment, the input data includes LiDAR point cloud, photogrammetry point cloud or other discrete spatial sampling data. Data preprocessing includes automatically performing noise removal, normal estimation and downsampling operations on point cloud data. Through spatial indexing construction, the efficiency of neighborhood query is improved, and a standardized point cloud file containing coordinates, normals and semantic information is obtained.

[0079] S200, constructing a global NURBS parameter domain for spatial intelligent world modeling, and uniformly mapping and managing multi-Patch geometric objects under the global NURBS parameter domain.

[0080] S200 is a global parameter domain establishment and mapping mechanism. The system defines a unified two-dimensional parameter domain as the basic coordinate system of the global geometric space in the initialization stage. Through affine mapping, the original point cloud coordinates are converted to the parameter domain space, and multiple logical patches are divided according to the point cloud distribution. All patches are generated under the same parameterization rule, realizing the expression structure of global geometric continuity and parameter consistency.

[0081] Different from the prior art which performs conventional parameterization on a single NURBS surface, the application is to construct a global NURBS parameter domain for spatial intelligent world modeling, and to perform consistent mapping and management on multiple Patch geometric objects under the global NURBS parameter domain.

[0082] The global NURBS parameter domain is a continuous, unified and extensible two-dimensional parameter space, and its parameter coordinate system is globally unique during the modeling process, and is used to carry the parameter mapping relationship of multiple geometric patches in space. Different from the method of defining a local (u, v) parameter domain for each surface in the conventional NURBS surface method, the application pre-establishes a global parameter domain, so that the parameter coordinates of different geometric patches are no longer independent of each other, but are subject to unified parameter scale, direction and boundary constraints.

[0083] After the global parameter domain is established, the application maps each geometric patch in space to the corresponding parameter sub-region in the global parameter domain.

[0084] Further, the global NURBS parameter domain of the application not only serves as a parameter basis for geometry generation, but also serves as a unified parameter coordinate system for subsequent control point matrix generation, curvature continuity constraint, adaptive subdivision and global consistency optimization. By organizing geometric information in the unified parameter domain, the spatial modeling process has good continuity, controllability and extensibility, and provides a bottom support for constructing a city-level or large-scale continuous spatial model.

[0085] Specifically, in the initialization phase, a unified two-dimensional parameter domain is first defined in the global coordinate system as a global NURBS parameter domain, which is the basic parameter coordinate system of the entire spatial intelligent world model. The global NURBS parameter domain is globally unique during the modeling process, and its parameter scale, direction and boundary rules are fixed and unchanged throughout the modeling process, and is used to provide a unified parameter basis for subsequent geometry generation, optimization and subdivision.

[0086] For city-level or complex world scenarios, the system divides the overall space into multiple logical Patch regions, and each Patch corresponds to a parameter sub-interval in the global parameter domain. Unlike the prior art, in which each Patch independently establishes a local parameter domain, all Patches in the present application share the same global parameter domain coordinate system, with only differences in parameter value ranges. The system maps the spatial coordinates of each Patch to the corresponding sub-interval in the global parameter domain through a pre-set parameter mapping rule, allowing adjacent Patches to naturally connect in the parameter domain.

[0087] During the mapping process, the Patch boundary parameters are uniformly aligned, ensuring that the boundary parameter ranges of adjacent Patches remain continuous and consistent in the parameter domain, thereby achieving parameter continuity without additional manual stitching during geometric splicing. Through this mapping mechanism, the scale drift, direction inconsistency, and cross-region splicing distortion problems caused by independent definition of local parameter domains in the prior art can be effectively avoided.

[0088] The global NURBS parameter domain established by S200 provides a unified parameter basis for subsequent control point matrix generation, continuity maintenance, adaptive subdivision, and global consistency optimization, enabling the multi-Patch geometric structure to be continuously expressed and cooperatively optimized within the same parameter space, thereby constructing a spatial intelligent world model with global consistency.

[0089] S300, based on the global NURBS parameter domain, generates and optimizes the control point matrix for geometric expression.

[0090] In the present application, S300 is a control point matrix generation and global collaborative optimization mechanism, which includes: automatically generating control point matrices and weight matrices based on the global NURBS parameter domain division results, for describing the geometric features of surfaces. The control point positions and weights are optimized through algorithms or learning models, making the surface morphology smooth, continuous, and differentiable. This mechanism realizes the parametric reconstruction from discrete samples to continuous geometric models. Specifically:

[0091] After establishing and mapping the global NURBS parameter domain, the control point matrix for geometric expression is generated and optimized based on the global parameter domain. The control point matrix, as an explicit parametric representation of NURBS geometry, is the core geometric variable that can be calculated and optimized in the spatial intelligent modeling process of the present application.

[0092] In the spatial intelligent world modeling method based on the global NURBS parameter domain, the geometry generation and optimization of three-dimensional space are realized based on the NURBS surface model, and its mathematical expression form is as follows:

[0093] (1)

[0094] where, denotes the parameter coordinate corresponding three-dimensional space point is the control point coordinate in the control point matrix is the weight of the control point and respectively to and B-spline basis function; p, q are the order of the spline; n, m are the number of control point matrix in two directions respectively.

[0095] Formula (1) reflects the basic logic of the geometry generation of the present application: any point in three-dimensional space is generated by the weighted combination of all control points in the control point matrix. The influence weight of each control point is determined by the basis function and the weight . The numerator part represents the comprehensive contribution of each control point to the current parameter point, and the denominator is a normalization term to ensure the smoothness and numerical stability of the surface shape. The surface generated in this way is geometrically continuous and differentiable, locally controllable and globally consistent, and can achieve strict geometric continuity and smooth transition in the global parameter domain.

[0096] In the system of the present application, formula (1) is not only used for surface generation, but also as the core geometric representation mechanism of the entire space intelligent modeling.

[0097] Specifically, geometric continuity: through the smoothness of the basis function and the position constraint of the control point, G0, G1, G2 continuity between patches is realized; global parameter consistency: the parameter coordinates (u, v) are defined in the globally shared NURBS parameter domain, so that all geometric patches are generated in a unified coordinate system; learnability and optimizability: the control point matrix and the weight can be used as optimization variables and directly updated through gradient descent or AI models to realize automatic generation and adaptive optimization of geometry; hierarchical adaptive capability: when the local area curvature changes greatly or the error exceeds the threshold, new control points can be automatically interpolated in the parameter domain to realize hierarchical subdivision. This formula describes the generation method of geometric points, which constitutes the mathematical basis of the globally continuous and optimizable geometry system in the present application. Through the control point matrix optimization framework centered on this formula, the present application can realize consistent modeling and accurate optimization of local geometric patches to global scenes in a unified parameter space, providing a computable and differentiable geometric basis for high-precision generation and adaptive update of the spatial intelligent world model.

[0098] It needs to be particularly pointed out that: unlike the prior art based on a single surface or local Patch to directly construct the control point matrix, the control point matrix generation process of the present application is carried out under the constraint of the global NURBS parameter domain. Specifically, the present application constructs a regular or adaptive control point distribution structure in the parameter domain according to the mapping relationship between the parameter positions in the global NURBS parameter domain and the spatial geometry, and maps the control point distribution to the corresponding spatial control point matrix, so that the control point matrix has global consistency at the parameter domain level. Specifically:

[0099] First, the system generates a control point parameter grid in the global NURBS parameter domain according to a preset resolution or adaptive rule, and maps the control points in the parameter domain to the corresponding spatial control point coordinates according to the mapping relationship from the parameter domain to the spatial coordinates, thereby forming an initial control point matrix. The control point matrix is not only for a single Patch, but is generated under the constraint of the global parameter domain, so that the control points of different Patches have uniform distribution rules and index relationships at the parameter level.

[0100] Secondly, after the control point matrix is generated, the system performs global collaborative optimization processing on the control point matrix. Unlike the prior art which only adjusts the control points in a single surface or local Patch independently, the present application regards the control point matrices of multiple Patches as a unified optimization object and performs joint optimization under the constraint of the global parameter domain. Specifically, the control point matrix as an optimization variable is jointly affected by the following constraint conditions:

[0101] Parameter domain continuity constraint, used to ensure that the boundary control point parameter positions of adjacent Patches in the global parameter domain remain consistent;

[0102] Geometric fairing constraint, used to suppress geometric discontinuity caused by local control point distribution being too dense or too sparse;

[0103] Error tolerance constraint, used to limit the control point adjustment amplitude so that the geometric accuracy meets the preset requirements.

[0104] Through the above constraint conditions, the system realizes consistent distribution and collaborative optimization of the control point matrix at the global scale while maintaining the local geometric expression capability. The optimized control point matrix is not only used for current geometry generation, but also serves as the basis input for subsequent curvature continuity maintenance, adaptive subdivision, and global consistency solving, thereby enabling the geometric model to have iterative updating and adaptive optimization capability.

[0105] Therefore, although the NURBS geometry expression form adopted by the present application itself belongs to the prior art, by introducing the global parameter domain constraint and the cooperative optimization mechanism of the multi-Patch control point matrix, the control point matrix is converted from the traditional static modeling result to a globally optimizable geometry parameter set, which is significantly different from the conventional control point generation method in the prior art.

[0106] S400, a continuity maintaining mechanism based on curvature and normal constraints is established to realize automatic smooth transition of the multi-Patch geometry at the spliced region.

[0107] To realize automatic smooth transition of the multi-Patch geometry at the spliced region, the present application establishes a continuity maintaining mechanism based on curvature and normal constraints in S400. The mechanism is used to maintain stable geometry continuity of adjacent Patches under the global parameter domain without relying on manual repair. The mechanism includes: the system calculates the curvature and normal difference of the boundary of adjacent Patches to detect the geometric discontinuity region. By adjusting the control point position and optimizing the boundary constraint, the normal (G1) and curvature (G2) level continuity is automatically realized. The mechanism ensures smooth splicing between different Patches without visible joints or shape mutations.

[0108] Specifically, the system calculates the corresponding curvature information and normal information based on the generated control point matrix at the spliced boundary of adjacent Patches, and takes the curvature difference and normal deviation at the boundary of adjacent Patches as the continuity evaluation index. When the curvature difference or normal deviation is detected to exceed the preset threshold, the system automatically triggers the continuity optimization process.

[0109] In the continuity optimization process, the system takes the control point coordinates in the control point matrix as the optimization variables, and iteratively adjusts the control point positions by simultaneously constraining the curvature change and normal change at the boundary of adjacent Patches. In this way, the normal direction at the spliced boundary is continuous, and the curvature mutation is further reduced, thereby realizing G1 to G2 level geometry continuity, while ensuring reasonable local geometry shape.

[0110] The continuity maintaining mechanism based on curvature and normal constraints not only acts on a single spliced boundary, but also uniformly processes the boundaries of multiple Patches under the global parameter domain constraint, so that the entire space model maintains continuous and smooth geometry structure in the global range. Through this mechanism, the problems of boundary polyline, normal jump and surface tearing commonly found in the existing multi-Patch modeling method can be effectively avoided, thereby constructing a stable and controllable continuous world geometry model.

[0111] The core optimization equation of the continuity maintaining mechanism based on curvature and normal constraints is defined as follows:

[0112] (2)

[0113] wherein: represents the control point coordinates of the control point matrix in the NURBS surface, and is an optimization variable; represents the curvature difference at the boundary of adjacent patches; represents the normal deviation of the boundary of adjacent patches; is a balance coefficient, used for weight adjustment between curvature continuity (G2 continuity) and normal continuity (G1 continuity).

[0114] By simultaneously minimizing the curvature difference and the normal deviation between adjacent patches, automatic smooth connection at the geometric spliced boundary is achieved.

[0115] When Δn=0, the surface is continuous in the boundary normal direction (G1 continuity);

[0116] When Δk=0, the curvature direction is continuous (G2 continuity).

[0117] By adjusting the coefficient λ, the system can dynamically balance between normal smoothing and curvature smoothing according to the scene requirements.

[0118] In the optimization process, the system takes the control point matrix and its weight as independent variables, calculates the deviation values of the curvature and the normal of the boundary of adjacent patches, and iteratively updates the control point positions by using numerical methods such as gradient descent or least squares solution. The optimization process aims to make the energy gradient of the surface continuous at the spliced boundary, avoiding the occurrence of broken lines, depressions or irregular mutations.

[0119] Through the above optimization mechanism, the spatial distribution of the control point matrix can be automatically adjusted, so that the multi-patch NURBS surface maintains G1 / G2 level geometric continuity in the global parameter domain, thereby forming a smooth and stable continuous world geometry structure.

[0120] S500, based on the adaptive subdivision and error control mechanism, automatically performs NURBS surface subdivision operation according to the curvature change rate and the geometric error threshold.

[0121] The present application realizes adaptive subdivision based on geometric complexity by setting the error tolerance ε. In complex or high curvature areas, the density of control points is increased, and in flat areas, the structure is simplified. This mechanism reduces the data amount while ensuring geometric accuracy, achieving a balance between modeling accuracy and computational efficiency.

[0122] When the local curvature exceeds the threshold or the geometric error accumulates too much, the system automatically inserts a new control point layer to perform hierarchical NURBS refinement on the NURBS surface. The subdivision process is completely completed in the global NURBS parameter domain, and the system can dynamically allocate the control point density according to the curvature change rate, thereby effectively reducing the data and calculation amount while ensuring the geometric accuracy, and realizing the adaptive balance of precision and efficiency.

[0123] S600, based on a global consistency optimization solving mechanism, performing global solving on control point matrices of all patches, and outputting a spatial intelligent world model which is continuously differentiable and parameter consistent in a global range.

[0124] After the local optimization of the internal control points of the patches is completed, global solving is performed on the control point matrices of all the patches. The optimization objectives include geometric continuity, parameter smoothness and overall energy constraint. Through global convergence update, consistent coordination and smooth transition of the entire world model in the parameter domain are realized.

[0125] Specifically, the control point matrices of all the patches are solved by a unified global optimizer, and the objective function comprehensively considers geometric continuity, parameter smoothness and energy constraint:

[0126] (3)

[0127] Wherein: represents a geometric continuity error term, represents a parameter domain smoothness term, is used to prevent the control points from being too dense. The optimizer adopts a multi-level solving strategy, first locally optimizes the internal control points of the patches, and then performs global synchronous update to realize globally convergent geometric consistent modeling.

[0128] Preferably, on the basis of completing the spatial intelligent world model in S600, according to specific application requirements, the spatial intelligent world model is further subjected to multi-scale expression, format export, system docking or application layer processing to improve the model applicability.

[0129] The spatial intelligent world modeling method based on the global NURBS parameter domain also includes:

[0130] S700, based on a multi-scale hierarchical modeling and fusion mechanism, the curvature and topological continuity are maintained through nested parameter domain mapping relationship between different levels, and seamless fusion of spatial intelligent world modeling across scales is realized.

[0131] The multi-scale hierarchical modeling and fusion mechanism adopts a multi-level NURBS expression mode to realize continuous modeling from the city level to the building level and then to the local structure. The upper layer Patch expresses the macro structure, and the middle and lower layers of Patch gradually refine the geometric details. The parameter domain and curvature are continuously maintained between different levels to realize a seamless fusion spatial intelligent modeling system.

[0132] To adapt to the spatial intelligent modeling needs at different scales, the application adopts a hierarchical NURBS expression system: the upper layer Patch is responsible for expressing the city-level macro structure; the middle layer Patch is responsible for the terrain and building group contour; and the lower layer Patch describes the local geometric details. The curvature and topological continuity are maintained between layers through nested parameter domain mapping relationships to realize seamless fusion modeling across scales.

[0133] The following takes the implementation of city-level three-dimensional scene modeling in the global NURBS parameter domain as an example to further illustrate the spatial intelligent world modeling method of the application:

[0134] This embodiment is used to realize continuous geometric expression of the city-level three-dimensional scene in the global NURBS parameter domain. NURBS (Non-Uniform Rational B-Spline) is used as the underlying geometric logic to realize parameterized reconstruction from discrete point cloud data to continuous and differentiable world models. As shown in Figure 2 , it includes:

[0135] S101, data preparation and preprocessing: input city-level LiDAR point cloud data to automatically perform noise removal, normal estimation and downsampling operations. The construction of a spatial index improves the efficiency of neighborhood queries to obtain a standardized point cloud file containing coordinate, normal and semantic information.

[0136] The input data of this embodiment includes LiDAR (Light Detection and Ranging) point cloud, photogrammetry point cloud or other discrete spatial sampling data. The point cloud data is only used as the original spatial sampling input to provide scene geometric shape, normal direction and local curvature information. The application does not directly perform geometric modeling on the point cloud, but uses it as a geometric constraint condition to convert the discrete data into continuous geometric expression through parameterized mapping and control point fitting.

[0137] Specifically, the system establishes a two-dimensional parameter domain in the global coordinate system, projects the point cloud data to the parameter domain through an affine mapping function, and generates a control point matrix and a weight matrix based on the point cloud distribution density and curvature characteristics. In this way, the LiDAR point cloud is only used as the input sample source, and the NURBS parameter domain is used as the underlying geometric expression structure.

[0138] The input data is city-level LiDAR point cloud (LAS format) with an average density of about 10 points / m2 The system first performs the following preprocessing steps: noise filtering: using statistical outlier rejection algorithm (radius 1.0 m, neighborhood point number 20); normal estimation: using PCA method to estimate the normal of each point; downsampling: using voxel filtering (voxel size 0.2 m) for uniform sampling; spatial index construction: index is established by KD-Tree to speed up subsequent neighborhood curvature calculation. The point cloud file output by preprocessing contains spatial coordinates, normal and semantic label information.

[0139] S201, global parameter domain definition and node vector construction: a unified two-dimensional NURBS parameter domain is established in the global coordinate system, and the point cloud is mapped into a parameterized space representation. The system performs coordinate normalization, parameter mapping, node vector generation and patch division to ensure global continuity and differentiability.

[0140] A unified NURBS parameter domain is established in the global coordinate system, which is used to map discrete point cloud spatial data into continuous two-dimensional parameter representation. As shown in Figure 3 , step S201 includes:

[0141] S2011, global coordinate normalization: the input LiDAR point cloud is processed in the world coordinate system for coordinate normalization, eliminating the difference in scale of different scenes, providing a unified calculation basis for subsequent parameterization mapping.

[0142] S2012, parameter domain mapping: a two-dimensional parameter domain (u, v) is established in the normalized space, and the point cloud coordinates are converted into two-dimensional parameter domain (u, v) coordinates through an affine mapping function. Each point cloud sample obtains a unique parameter representation, realizing the unified mapping from geographic coordinates to parameter space.

[0143] S2013, node vector generation: non-uniform node vectors are automatically generated according to the point cloud density and curvature characteristics, which are used to control the surface resolution. The nodes are more densely distributed in high curvature areas, and the node spacing is larger in flat areas, so as to ensure smoothness while improving efficiency.

[0144] S2014, patch region division and splicing relationship: the parameter domain is divided into multiple patch regions according to the point cloud characteristics, and a global parameter index table is established. Adjacent patches share boundary nodes to realize seamless connection at the parameter continuous and geometric splicing place, ensuring global consistency.

[0145] Specifically:

[0146] First, the input LiDAR point cloud is represented in WGS-84 or UTM coordinates in the world coordinate system, with a large numerical range and inconsistent magnitude. To ensure the stability of subsequent parameter calculation, the system first performs global normalization to normalize the original spatial coordinates to [0, 1] 3The normalized space is calculated as follows:

[0147]

[0148] Wherein: X, Y, Z: respectively represent the three-dimensional space coordinate components of any sampling point in the LiDAR point cloud in the original world coordinate system (WGS-84 or UTM coordinate system), which correspond to the original coordinate values of the point in the X-axis, Y-axis and Z-axis directions. X', Y', Z': respectively represent the normalized coordinate components obtained by globally normalizing the original space coordinates X, Y, Z, whose value range is mapped to [0, 1], which is used for subsequent parameter domain mapping and numerical calculation. Xmax, Ymax, Zmax: respectively represent the maximum coordinate values of the current input point cloud data set in the X-axis, Y-axis and Z-axis directions, which are used to determine the upper bound of the spatial range of the point cloud in the corresponding coordinate axis direction. Xmin, Ymin, Zmin: respectively represent the minimum coordinate values of the current input point cloud data set in the X-axis, Y-axis and Z-axis directions, which are used to determine the lower bound of the spatial range of the point cloud in the corresponding coordinate axis direction.

[0149] Through this step, the point cloud data is mapped to a unified scale of the spatial range, providing a calculation basis for subsequent parameterization mapping.

[0150] Secondly, the system defines a two-dimensional parameter domain in the normalized space and establishes the corresponding relationship between the point cloud coordinates and the parameter coordinates through an affine mapping function: wherein, is a linear transformation coefficient, which is automatically calculated from the size range of the global scene. Through this mapping function, each point cloud sample obtains a unique (u, v) coordinate in the two-dimensional parameter domain, realizing the unified conversion from geographic coordinates to parameter space.

[0151] Subsequently, the system adaptively divides the parameter domain according to the point cloud distribution density and curvature gradient, forming a number of rectangular subdomains (i.e. Patch regions). Each Patch corresponds to a local geometric element in the urban space (such as building facade, road, land parcel or vegetation area). The Patches are continuously arranged in the parameter space, with seamless connection of the boundary parameter ranges, thereby maintaining global continuity and differentiability in the entire parameter domain range. For each Patch, the node vectors U, V are automatically generated according to the number of control points and the selected NURBS order (this embodiment uses cubic NURBS, i.e. p=q=3) as follows:

[0152]

[0153] The non-uniform distribution of node vectors reflects the difference in the geometric complexity of the point cloud. In areas with high curvature or dense data, the node spacing is automatically reduced to support high-resolution local surface expression; while in flat or sparse areas, the node spacing is increased to reduce computational redundancy and improve overall computational efficiency.

[0154] The automatic generation of node vectors uses the "curvature weighted distribution method". The system calculates the local curvature average k(u, v) in each parameter direction, and according to the weighted function: , is the curvature adjustment coefficient, which dynamically adjusts the node spacing to adaptively match the surface resolution to the scene geometric complexity, thereby significantly improving the modeling efficiency and accuracy while maintaining geometric continuity.

[0155] After the node vector generation is completed, the system establishes a global parameter index table, recording the parameter range of each Patch When the boundary parameter intervals of adjacent Patches overlap, the system automatically binds the corresponding node index to ensure the parameter continuity and node consistency at the splicing boundary. Through this "parameter domain sharing" mechanism, the entire city-level scene is expressed globally and consistently in the parameter space, and there is no normal jump or parameter break problem even when splicing across regions, thereby providing a unified geometric basis for subsequent curvature-driven continuity optimization and adaptive subdivision.

[0156] S301, control point matrix and weight matrix generation: automatically generate control point matrix and weight matrix within each Patch to define the initial surface structure. The number of control points is adaptively adjusted according to the curvature characteristics, and the generated matrix file is used for subsequent surface splicing and optimization calculation.

[0157] Specifically: within each parameter Patch, the system automatically generates a control point matrix and a weight matrix . Wherein: the control point grid size is: default 40x40 (high curvature areas can be increased to 50x50) Each control point records its coordinates and local normal. The weight initialization formula is: . Wherein is the curvature variance of the control point neighborhood. The control point matrix and weight matrix are stored as files for subsequent continuity optimization and AI training module calls.

[0158] S401, NURBS surface generation and splicing: call the geometric operation module (such as libnurbs or OpenNURBS) to perform surface solving and splicing. Through basis function interpolation and node management, local support and global smoothness are achieved, and G1 / G2 level continuous connection is maintained between adjacent Patches.

[0159] In the spatial intelligent world modeling system of the application, NURBS surface generation and splicing NURBS (non-uniform rational B-spline) geometric calculation are realized by a NURBS geometric operation module at the bottom of the system, such as Figure 4 The NURBS geometric operation module comprises:

[0160] A base function evaluation submodule is configured to calculate the numerical results of B-spline base functions and generate continuous and differentiable shape functions on the parameter domain, which are used for surface weighted interpolation.

[0161] A surface evaluation and derivative calculation submodule is configured to calculate surface point coordinates according to the control point matrix and weights, and to calculate the first derivative, the second derivative and the normal vector of the surface, so as to provide input for subsequent curvature-driven optimization and continuity constraints.

[0162] A node vector management submodule is configured to realize interpolation and adaptive subdivision of node vectors, and to automatically insert new nodes to improve local geometric accuracy when the local curvature gradient exceeds a threshold.

[0163] A curvature and continuity calculation submodule is configured to calculate the curvature difference and normal deviation of the surface at the splicing boundary, support G0, G1 and G2 condition detection, and output constraint terms for control point optimization.

[0164] A geometric derivation submodule is configured to export the generated NURBS model to realize seamless data exchange with BIM and digital twin systems.

[0165] Based on a general mathematical calculation library (such as libnurbs, OpenNURBS or equivalent geometric kernel), numerical solutions and surface generation are performed on the control point matrix, node vector and weight parameters. Each patch is defined as a continuous and differentiable function in the parameter domain, and the interpolation property of the node vector ensures local support and global smoothness. At the adjacent patch, the system automatically adjusts the boundary nodes and weights to make the splicing boundary meet the G1 (normal continuity) and G2 (curvature continuity) conditions, thereby forming an overall continuous world geometry model.

[0166] S501, continuity optimization of curvature and normal constraints: calculate the curvature difference and normal deviation of the adjacent patch boundary, and perform energy minimization with the control point position as the optimization variable. The system automatically updates the boundary control points to realize seamless splicing and smooth transition of the surface, and enhances the global geometric stability.

[0167] Specifically, after the patch is spliced, the curvature difference Δk and the normal deviation Δn of the adjacent patch boundary are calculated to further improve the overall smoothness and numerical stability, and the energy minimization is performed with the control point position as the optimization variable: where: λ is the balance coefficient (value 0.5-1.0). The optimization is achieved by gradient descent, learning rate 0.01, convergence threshold The optimization result automatically updates the control point matrix and boundary parameters to achieve seamless splicing between multiple patches.

[0168] S601, adaptive subdivision and error control: when the local curvature changes exceed the threshold, the patch is automatically subdivided and the new control point matrix is interpolated. The node vector is updated synchronously by B-spline interpolation rule to ensure that the high curvature area still maintains curvature continuity and accuracy consistency.

[0169] Specifically: when the local curvature changes exceed the preset threshold, the system automatically interpolates the new control point matrix and updates the basis function node vector to maintain the consistency of the surface parameterization. The subdivision adopts the B-spline interpolation rule, and the boundary control points are updated in real time to ensure that the G2 continuity is still maintained in the high curvature area.

[0170] S701, global consistency solving and model output: after all patch optimization is completed, the system performs global consistency solving to smooth the control point distribution and parameter coordinates. Finally, the continuous and differentiable, globally consistent spatial intelligent city-level modeling result is achieved.

[0171] Specifically: after all patch level optimization is completed, the system performs a global consistency solving to smooth the control point distribution and re-normalize the parameter coordinates. The output model includes the following formats: STEP / IGES file: for direct loading on BIM platform; JSON file: containing parameter domain, node vector, control point and weight information, which can be used for digital twin system, simulation engine or AI training. The generated model is continuous and differentiable in the global range, and the parameter is consistent, which can realize smooth and seamless spatial intelligent modeling in city-level scenarios.

[0172] Embodiment 2

[0173] Corresponding to the spatial intelligent world modeling method based on global NURBS parameter domain provided in Embodiment 1 of the present application, Embodiment 2 of the present application also provides a spatial intelligent world modeling system based on global NURBS parameter domain, as shown in Figure 5 The main functional modules include:

[0174] Parameter domain management module: used for establishing and maintaining the mapping relationship of the global NURBS parameter domain. This module defines a unified two-dimensional parameter coordinate system, maps the input point cloud or spatial sampling data into the global parameter domain, and provides a unified geometric coordinate basis for subsequent control point generation and surface splicing.

[0175] Control point generation and optimization module: used to automatically generate initial control point matrix and weight matrix according to input point cloud, curvature and semantic features. The system performs control point position optimization in this module, so that the surface has continuous and differentiable geometry expression in the global parameter domain, and is locally controllable.

[0176] Continuity constraint module: used to detect and constrain the normal and curvature difference of adjacent patches, and realize automatic continuity maintenance at the geometric boundary. This module adjusts the boundary control point parameters iteratively to ensure smooth connection of the model at G1 (normal) and G2 (curvature) levels.

[0177] Subdivision and error control module: used to perform adaptive subdivision process based on error tolerance ε. When the local curvature or error exceeds the set threshold, the system automatically inserts new control points and regenerates node vectors to realize local refinement and smooth compensation of high-precision areas.

[0178] Global optimization and model output module: used to solve and parameter coordinate the control point matrix of all patches and output the model. This module considers geometric continuity, parameter smoothness and energy constraint to realize global consistent optimization of the overall structure, and the output model is continuously differentiable and parameter consistent in the global range.

[0179] Preferably, it further comprises:

[0180] Multi-scale synchronization module: used to maintain parameter consistency and geometric smoothness between different levels. This module uses multi-level NURBS expression mechanism to make each scale model keep nested continuity in topology and curvature, so as to realize unified modeling and seamless fusion across scales.

[0181] The above has described the embodiments of the present disclosure, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those skilled in the art without departing from the scope and spirit of the described embodiments. The choice of terms used in the present invention is intended to best explain the principles, practical application or technical improvement in the market of the embodiments, or to enable other ordinary skilled in the art to understand the disclosed embodiments.

Claims

1. A globally NURBS parametric domain based spatial intelligent world modeling method, characterized in that, The method comprises the following steps: data input and data preprocessing; constructing a global NURBS parameter domain for spatial intelligent world modeling, and uniformly mapping and managing multi-Patch geometry objects under the global NURBS parameter domain; generating and optimizing a control point matrix for geometric expression based on the global NURBS parameter domain; implementing automatic smooth transition of multi-Patch geometry in a spliced region based on a continuity maintenance mechanism of curvature and normal constraints; performing NURBS surface subdivision operation according to a curvature change rate and a geometric error threshold based on an adaptive subdivision and error control mechanism; performing global solving on control point matrices of all patches based on a global uniform optimization solving mechanism, and outputting a spatial intelligent world model which is continuously differentiable and parameter consistent in a global range; wherein the global NURBS parameter domain construction process comprises the following steps: performing coordinate normalization processing on point cloud data in a world coordinate system; establishing a two-dimensional parameter domain in the normalized space, and converting point cloud coordinates into parameter domain coordinates through an affine mapping function; automatically generating a non-uniform node vector according to point cloud density and curvature characteristics; dividing the parameter domain into multiple Patch regions according to point cloud characteristics, and establishing a global parameter index table; the continuity maintenance mechanism of curvature and normal constraints comprises the following steps: at the spliced boundary of adjacent patches, the corresponding curvature information and normal information are calculated based on the generated control point matrix, and the curvature difference and normal deviation at the boundary of adjacent patches are taken as continuity evaluation indexes; when the curvature difference or the normal deviation is detected to exceed a preset threshold, a continuity optimization process is automatically triggered; in the continuity optimization process, the control point coordinates in the control point matrix are taken as optimization variables, and the control point positions are iteratively adjusted by simultaneously constraining the curvature change and the normal change at the boundary of adjacent patches.

2. The method of claim 1, wherein, The method further comprises the following steps: based on a multi-scale hierarchical modeling and fusion mechanism, the curvature and topological continuity are maintained among different levels through nested parameter domain mapping relationships.

3. The method of claim 1, wherein, the step of generating and optimizing a control point matrix for geometric expression based on the global NURBS parameter domain comprises the following steps: a control point parameter grid is generated in the global NURBS parameter domain according to a preset resolution or an adaptive rule, and the control points in the parameter domain are mapped into corresponding spatial control point coordinates according to the mapping relationship from the parameter domain to the spatial coordinates, so as to form an initial control point matrix; the initial control point matrix is subjected to global collaborative optimization processing.

4. The method of claim 1 or 3, wherein, the control point matrix is taken as an optimization variable, and is subjected to the following constraint conditions: a parameter domain continuity constraint for ensuring that the boundary control point parameter positions of adjacent patches in the global parameter domain remain consistent; a geometric fairing constraint for inhibiting geometric mutation caused by excessively dense or sparse local control point distribution; an error tolerance constraint for limiting the control point adjustment amplitude so that the geometric precision meets a preset requirement.

5. The method of claim 1, wherein, the step of performing NURBS surface subdivision operation according to a curvature change rate and a geometric error threshold comprises the following steps: when the local curvature exceeds the threshold or the geometric error accumulation exceeds the threshold, a new control point layer is automatically inserted to perform hierarchical refinement on the NURBS surface.

6. The method of claim 2, wherein, the multi-scale hierarchical modeling and fusion mechanism comprises the following steps: The layered NURBS expression system is adopted: the upper layer patch is responsible for expressing the city-level macro structure; the middle layer patch is responsible for the terrain and building group contour; and the lower layer patch describes local geometric details; and the curvature and topological continuity are maintained between layers through nested parameter domain mapping relationship.

7. A globally NURBS parametric domain based spatial intelligent world modeling system characterized by, The system comprises: a parameter domain management module, which is used for establishing and maintaining the mapping relationship of the global NURBS parameter domain; wherein the construction process of the global NURBS parameter domain comprises: performing coordinate normalization processing on the point cloud data in the world coordinate system; establishing a two-dimensional parameter domain in the normalized space, and converting the point cloud coordinates into parameter domain coordinates through an affine mapping function; automatically generating a non-uniform node vector according to the point cloud density and curvature characteristics; dividing the parameter domain into a plurality of patch regions according to the point cloud characteristics, and establishing a global parameter index table; a control point generation and optimization module, which is used for automatically generating an initial control point matrix and a weight matrix according to the input point cloud, curvature and semantic characteristics; a continuity constraint module, which is used for detecting and constraining the normal and curvature difference of adjacent patches, and realizing automatic continuity maintenance at the geometric boundary based on the continuity maintenance mechanism of the curvature and normal constraint; wherein the continuity maintenance mechanism of the curvature and normal constraint comprises: at the splicing boundary of adjacent patches, the corresponding curvature information and normal information are calculated based on the generated control point matrix, and the curvature difference and normal deviation at the boundary of adjacent patches are taken as the continuity evaluation indexes; when the curvature difference or the normal deviation exceeds a preset threshold, the continuity optimization process is automatically triggered; in the continuity optimization process, the control point coordinates in the control point matrix are taken as the optimization variables, and the control point positions are iteratively adjusted by simultaneously constraining the curvature change and the normal change at the boundary of adjacent patches; a subdivision and error control module, which is used for performing an adaptive subdivision process based on the error tolerance; when the local curvature or error exceeds a set threshold, a new control point is automatically inserted and a node vector is regenerated; a global optimization and model output module, which is used for uniformly solving the control point matrix of all patches and parameter coordination and model output.

8. The spatially intelligent world modeling system of claim 7, wherein, The system further comprises: a multi-scale synchronization module, which is used for maintaining the parameter consistency and geometric smoothness between different levels, so that the models of different scales maintain nested continuity in topology and curvature.

Images