Method and system for determining a dynamic visible scattering surface for millimeter wave radar human body digital twinning

Abstract

The application discloses a dynamic visible scattering surface determination method and system for millimeter wave radar human body digital twinning, and belongs to the technical field of digital human body modeling and radar echo simulation. The method takes a dynamic human body outer surface triangular mesh and radar space geometry as input, determines candidate surfaces through candidate scattering surface construction and orientation pre-screening, constructs a potential occlusion surface set by using deterministic projection overlap enumeration, identifies real occlusion surfaces after depth order judgment and recognition, performs geometric set operation on each occlusion shadow, and subtracts the set from the target surface to obtain a continuous visible sub-region; performs oriented triangulation on the continuous visible sub-region to maintain the parent surface around the sequence and normal consistency, performs hierarchical processing on extremely thin regions and records error budget, performs dynamic continuity gating on adjacent frames, and releases a dynamic visible scattering surface asset containing geometric information, attribute information and quality judgment indicators, to support coherent echo simulation of millimeter wave radar human body digital twinning.

Description

Technical Field

[0001] This invention relates to the fields of digital human body modeling, millimeter-wave electromagnetic scattering calculation, radar echo simulation, and digital twin technology, and in particular to a method and system for determining the dynamic visible scattering surface for millimeter-wave radar human body digital twins. Background Technology

[0002] Human digital twins play a crucial role in millimeter-wave radar vital sign modeling, respiration and heartbeat micro-motion simulation, algorithm regression testing, and multimodal sensor co-design. For millimeter-wave radar, the radar echo from the human body surface is not simply generated uniformly across the entire body surface, but rather formed by surface regions that meet illumination, unobstructed, and geometric orientation conditions. Therefore, accurately determining the visible scattering surface at each moment on a dynamic human body mesh is a prerequisite for establishing a reliable digital twin radar echo chain.

[0003] In existing technologies, one type of method discretizes the human body as a point scattering source or treats the entire body surface as an equivalent scattering surface, without explicitly handling the self-occlusion relationship between surface elements. This type of method can quickly obtain motion trends, but it is difficult to distinguish which surface areas are actually visible under a given radar viewpoint, and it is easy to introduce non-physical scattering contributions in occlusion-sensitive locations such as the sides of the torso, armpits, mastoid edges, and lower clavicle edges.

[0004] Another type of method uses the center of a triangular face, the centroid of a triangular face, or a small number of sampling points as representatives. It performs one or a small number of ray first intersection judgments on each face and directly classifies the entire face element as visible or invisible based on this. Although this method can reflect coarse-grained visibility, when the occlusion boundary cuts through the interior of the face element, even a tiny boundary movement can cause the state of the entire face element to change abruptly, thus producing visibility flicker in the dynamic sequence and causing non-physical jumps in coherence amplitude, range profile, and spectrum.

[0005] Another type of method first estimates the visible area ratio of the surface element, and then multiplies this ratio by an empirical reflection coefficient or an approximation of the radar cross section to correct the surface element contribution. Although this type of method is superior to pure binary determination, it often only outputs a single area ratio without retaining which specific area within the surface element is visible, and therefore cannot be stably connected to millimeter-wave coherent integral chains that require the preservation of spatial phase information.

[0006] Furthermore, if the set of occlusion surfaces is discovered using finite ray probes or finite quadrature points, changes in probe density may alter the discovered set of occlusion surfaces. If the original surface element order is not maintained for the clipped visible polygons, silent flipping will be introduced in subsequent calculations of normals, incident angles, or polarization. If extremely thin visible strips are directly deleted using a fixed area threshold, it may mask small but not negligible complex scattering contributions. Current techniques lack a unified method that simultaneously satisfies the requirements of continuous visible domain determination, reproducible occlusion enumeration, coherent integral compatibility, and dynamic stability gating.

[0007] Therefore, a dynamic visible scattering surface determination method for human digital twins of millimeter-wave radar is needed: it can not only determine whether each parent triangle is visible at each time, but also determine the continuous visible sub-regions inside the parent triangle; it can not only stably handle self-occlusion, boundary stripes and topological events, but also seamlessly integrate the results into coherent millimeter-wave echo integration, and complete the release judgment with explicit quality indicators and thresholds. Summary of the Invention

[0008] In view of this, the purpose of this invention is to provide a method and system for determining the dynamic visible scattering surface of a human body digital twin for millimeter-wave radar. This method addresses the problems of existing methods, such as visibility flicker caused by element-level binary determination, non-reproducible results due to limited probes, physical errors introduced by loss of normal after clipping, lack of error boundaries in extremely thin strips, and the inability of visibility results to directly serve coherent echo integration. Therefore, this invention proposes a method for determining the dynamic visible scattering surface of a human body digital twin for millimeter-wave radar.

[0009] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a method for determining the dynamic visible scattering surface of a human body in millimeter-wave radar digital twin, comprising the following steps: Acquire the dynamic outer surface mesh sequence and radar spatial geometric parameters of the human digital twin; Construct a set of candidate scattering surfaces according to preset rules; Through deterministic geometric calculations, all potential occlusion surfaces that have projection overlap with the target surface patch are enumerated; By performing a geometric union operation on the shadow regions of one or more real occluding surfaces on the target surface, and subtracting the shadow union from the target surface patch, a continuous visible sub-region on the target surface patch is obtained. Perform a meshing reconstruction process on the continuous visible sub-regions to maintain the original spatial orientation properties consistent with the original parent surface patch; Dynamic stability gating based on physical weights is applied to the visible scattering surface changes at adjacent time points; Publish a dynamic visible scattering surface asset that includes the geometric information of the continuous visible sub-regions and the quality judgment index of the dynamic stability gating.

[0010] Furthermore, the dynamic stability gating based on physical weights specifically includes at least one of the following: occlusion relationship variation gating based on shadow area weighting, occlusion relationship variation gating based on complex scattering contribution weighting, or occlusion relationship variation gating based on total target scattering energy.

[0011] Furthermore, the mesh reconstruction process is a triangulation process, and the original spatial orientation attributes include vertex wrap order and face normal. The vertex wrap order and face normal of the sub-region are the same as the vertex wrap order and face normal of its source parent surface patch.

[0012] Furthermore, the deterministic geometric calculation includes: establishing a local projection coordinate system with the radar position as the observation center, projecting the target surface patch and candidate scattering surfaces onto the same projection plane, and enumerating all potential occlusion surfaces based on the projection overlap relationship.

[0013] Furthermore, the dynamic visible scattering surface asset also includes: area information of the sub-surface patch, propagation distance information, phase reference information, incident angle information, distance gradient information, as well as the parent surface patch number from which the visible sub-region originates and the real occlusion surface index corresponding to the visible sub-region.

[0014] Furthermore, before performing mesh reconstruction processing on the continuous visible sub-regions to maintain the original spatial orientation attributes of their source parent surface patches, the process further includes: performing hierarchical processing on extremely thin visible sub-regions with areas smaller than a preset threshold; classifying extremely thin visible sub-regions that meet a preset precision threshold as resolvable fine strip regions and entering the mesh reconstruction processing flow; classifying extremely thin visible sub-regions that do not meet the preset precision threshold as sub-resolution fine strip regions and recording their area upper bound and complex contribution upper bound without performing forced mesh reconstruction.

[0015] Furthermore, the dynamic stability gating is also used to: output a failure record and not publish the dynamic visible scattering surface asset when the quality judgment index does not meet the preset threshold.

[0016] Furthermore, the radar spatial geometry parameters include the radar transmission and reception spatial location, radar operating wavelength or frequency, candidate surface component identifiers, and parent surface patch numbers used for identity tracking.

[0017] The present invention provides a dynamic visible scattering surface determination system for millimeter-wave radar-based human digital twins, used to perform the above-mentioned method, including: The data acquisition module is configured to acquire the dynamic outer surface grid sequence of the human digital twin and the radar spatial geometric parameters; The candidate surface construction module is configured to construct a set of candidate scattering surfaces according to preset rules; The occlusion enumeration module is configured to enumerate all potential occlusion surfaces that have projection overlap with the target surface patch through deterministic geometric calculations. The visible region generation module is configured to obtain continuous visible sub-regions on the target surface by performing a geometric union operation on the shadow regions of one or more real occluding surfaces on the target surface and subtracting the shadow union on the target surface sheet. The attribute-preserving reconstruction module is configured to perform a meshing reconstruction process on the continuous visible sub-regions to maintain the original spatial orientation attributes of their source parent surface patches. The stability gating module is configured to perform dynamic stability gating based on physical weights on the visible scattering surface changes at adjacent time points; The asset publishing module is configured to publish a dynamic visible scattering surface asset that includes the geometric information of the continuous visible sub-regions and the quality judgment index of the dynamic stability gating.

[0018] Furthermore, the dynamic visible scattering surface asset is a structured data volume, which includes at least: parent surface patch identification information, geometric description information of the visible sub-region, and the quality judgment index field of the dynamic stability gating. All three fields are indispensable and together constitute the asset.

[0019] The beneficial effects of this invention are as follows: This invention provides a method and system for determining the dynamic visible scattering surface of a human body in millimeter-wave radar digital twins. This method addresses the problems in existing methods, such as visibility flicker caused by element-level binary determination, non-reproducible results due to limited probes, non-physical errors introduced by loss of normal rotation after clipping, lack of error boundaries for extremely thin strips, and the inability of visibility results to directly serve coherent echo integration.

[0020] The principle of this invention is as follows: Using a dynamic human body outer surface triangular mesh and radar spatial geometry as input, the calculation range is first reduced through candidate scattering surface construction and orientation pre-screening; then, deterministic projection overlap enumeration is employed to transform the three-dimensional spatial occlusion relationship into geometric overlap determination on a two-dimensional projection plane, eliminating the uncertainty caused by finite sampling; next, the true occlusion surface is identified through depth order determination, and a geometric union operation is performed on each true occlusion shadow, subtracting the union from the target surface to obtain continuous visible sub-regions, thus advancing the visibility description from the surface element level to the continuous domain within the surface element; for continuous visible sub-regions... The method employs orientation-preserving triangulation to ensure that the vertex order and normals of the sub-triangles are consistent with the parent face, maintaining physical consistency for subsequent electromagnetic calculations. For extremely thin visible regions, it performs hierarchical processing of resolvable and sub-resolution stripes and records error budgets to control computational accuracy. Finally, it performs dynamic continuity gating based on area and complex contribution weighting on adjacent frames to distinguish between represented events and actual physical transitions. After successful gating, it publishes structured visible scattering surface assets containing geometric, attribute, and error information, directly supporting millimeter-wave coherent echo integration and subsequent calculations of scattering amplitude, phase, and range images. Compared with existing technologies, this method offers the following advantages: (1) This invention no longer simply classifies the entire parent triangle as visible or invisible, but outputs a continuous visible sub-region inside the parent triangle through geometric union and difference operations. This continuous sub-region preserves the complete boundary vertex order and the local coordinate continuity of the parent face, and can be directly used as the geometric domain of coherent millimeter-wave echo integration, solving the problem that existing visibility results cannot serve coherent integration.

[0021] (2) This invention constructs a set of potential obstruction surfaces through deterministic enumeration of overlapping radar center projections, without relying on finite ray probes or random sampling points to discover obstruction surfaces. Under the same input, radar attitude, and geometric accuracy conditions, the set of potential obstruction surfaces is uniquely determined, ensuring the reproducibility of obstruction relationships and overcoming the fundamental defect of the finite probe method being sensitive to sampling density.

[0022] (3) The present invention performs continuous geometric clipping on the target parent triangle face of the shadow formed by the real occlusion surface. The occlusion boundary changes continuously inside the surface cell, which significantly reduces the visibility flicker when the occlusion boundary crosses the surface cell, reduces the non-physical high frequency components in the dynamic echo, and improves the stability of the visible domain description in the dynamic sequence.

[0023] (4) The present invention adopts a standard tolerance that only applies to the endpoint of the target surface itself for the self-occlusion determination, without relaxing the tolerance of the general target neighborhood. This not only eliminates the false determination of the self-occlusion of the target surface, but also prevents the real occlusion relationship from being covered by the relaxation of the global tolerance, thus taking into account both the robustness of numerical calculation and physical reality.

[0024] (5) In the present invention, orientation-preserving triangulation is used when triangulating the visible sub-region, so that the vertex order and surface normal orientation of all sub-triangle faces are consistent with their source parent face, avoiding silent errors caused by surface normal flipping in subsequent incident angle, polarization or scattering calculations, and ensuring the physical consistency of electromagnetic quantity calculation.

[0025] (6) This invention establishes a hierarchical processing mechanism for resolvable and sub-resolution fine stripes in extremely thin visible regions. For sub-resolution fine stripes, unstable numerical integration is not enforced, but their area upper bound and complex contribution upper bound are recorded. This avoids the uncontrollable errors caused by simply deleting patches according to a fixed area threshold, making the calculation accuracy quantifiable and auditable.

[0026] (7) The present invention performs dynamic temporal gating using physical quantities such as shadow area weighted variation, complex scattering contribution weighted variation and visible domain symmetry difference ratio, which can effectively distinguish between topological events and real visible domain transitions, avoid misjudging mesh representation changes as physical changes, and improve the engineering reliability of dynamic visible scattering surface asset publishing.

[0027] (8) The results published by this invention not only include the geometry of visible sub-faces and sub-triangles, but also include parent face identifiers, real occlusion surface indexes, ultra-thin region classifications, error budgets and dynamic continuity gating indices, which facilitates result traceability, regression testing and engineering integration with subsequent millimeter-wave echo simulation modules, and can directly support the calculation of scattering amplitude, phase and range images.

[0028] The above and other objects, advantages, and features of the present invention will be more fully set forth and demonstrated through the following detailed description of specific embodiments in conjunction with the accompanying drawings. Those skilled in the art, upon referring to the following detailed description and the accompanying drawings, will be able to better understand and realize the above advantages of the present invention. Other objects, features, and advantages of the present invention will become clearer after being described in detail in the detailed description section in conjunction with the accompanying drawings. Attached Figure Description

[0029] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following drawings are provided for illustration.

[0030] Figure 1 A flowchart illustrating the overall process for determining the dynamic visible scattering surface; Figure 2 The flowchart shows the process of radar center projection, potential obstruction surface enumeration, and depth order determination. Figure 3 A flowchart for continuous shadow clipping, visible sub-region restoration, and orientation-preserving triangulation; Figure 4 Flowchart for sub-resolution fine stripe classification, error budgeting, and phase integration interfaces; Figure 5A flowchart for dynamic continuity gating and release determination; Figure 6 A block diagram of the system for determining the dynamic visible scattering surface. Detailed Implementation

[0031] The present invention will be further described below with reference to the accompanying drawings and a specific embodiment. This embodiment sequentially describes the data object, the construction of the candidate scattering surface set, radar center projection and occlusion enumeration, continuous shadow clipping, orientation-preserving triangulation, ultra-thin region processing, phase integration interface, dynamic continuity gating, and system implementation. It should be noted that this embodiment is used to illustrate the technical concept and engineering implementation of the present invention and does not constitute a limitation on the scope of protection of the present invention.

[0032] Example 1

[0033] I. Overall Processing Flow and Data Objects like Figure 1 As shown, this is the overall processing flow of this embodiment. The dynamic human body mesh and radar parameters first enter the candidate scattering surface construction and orientation pre-screening, and then the radar center projection, deterministic occlusion enumeration, depth order determination, continuous shadow clipping and orientation-preserving triangulation are executed in sequence. After obtaining the visible sub-triangles, the fine strip region classification and dynamic continuity gating are performed. Only when the gating is passed can the dynamic visible scattering surface assets be published. Figure 1 middle, This represents the standard tolerance for the intersection of the target parent face with its own endpoint, preferably... ; This indicates that the general occlusion depth has a relatively high tolerance, and is preferably preferred. Since they affect different objects, they must not be combined into a single global ray tolerance.

[0034] Specifically, This is only used for comparing the depths of the candidate obstruction surface and the target parent surface as shown in equation (6), that is, to determine whether the candidate obstruction surface is closer to the radar than the target parent surface at the same projection position. During the determination, the depth of the candidate obstruction surface is first compared with... The sums are then compared with the depth of the target parent surface; only when the sum is strictly less than the depth of the target parent surface is the corresponding overlapping projection portion considered a true occlusion shadow. In this embodiment... for Therefore, a strict depth order determination is actually adopted. It is not used for identifying the endpoint intersection of the target parent face itself, nor is it used as the global tolerance for determining projection overlap.

[0035] The system input consists of a dynamic mesh sequence of the outer surface of a human body with fixed parent vertex identities and fixed parent triangle connection relationships, and the spatial position and attitude of a given monostation millimeter-wave radar. Let the... The outer surface of the frame is composed of The dynamic outer surface is composed of 3 parent triangles, then it is defined as: (1) In equation (1), Indicates the frame index. ; Indicates the index of the parent triangle face. ; Indicates the first The human body's outer surface; Indicates the first The first frame A parent triangle face; Indicates the total number of faces of the parent triangle; Indicates the total number of frames in the dynamic sequence; symbol It represents the union of sets.

[0036] Before calculation begins, all input records are processed for identity information, including radar attitude, mesh file summary, number of parent triangles, number of vertices, candidate component labels, operating frequency, and runtime parameter version. If any identity information is inconsistent with the pre-registered configuration, the system halts the publishing process.

[0037] II. Construction of Candidate Scattering Surface Sets and Orientation Pre-screening The candidate scattering surface set is derived from the parent triangular facets of the human body's outer surface, and internal organ faces and non-external scattering surfaces are filtered out according to preset component roles. Let the candidate scattering surface set be... It can be represented as: (2) In equation (2), Indicates the first Set of candidate scattering surfaces for a frame; , and The meaning is the same as that of equation (1); Indicates the first The component role marker of the parent triangle is taken as follows: when it belongs to the outer surface direct scattering component. ; This represents the area of ​​the parent triangle, in square meters. This represents the minimum effective area threshold, in square meters. This indicates a geometric validity flag, which is used when the parent face does not have a degenerate, missing, or illegal index. ;symbol This indicates that the elements in the set simultaneously satisfy the subsequent condition.

[0038] Calculate the face center for each candidate parent triangle. Unit legal direction and the unit line-of-sight vector pointing from the face center to the radar When the orientation condition is met, the parent face is included in the subsequent occlusion enumeration: (3) In equation (3), Indicates the first The parent triangle is in the... The dimensionless incident orientation coefficient of the frame; Indicates the external normal of the parent face unit; Represents the unit line-of-sight vector pointing from the center of the parent face to the radar position; symbol Represents the dot product of vectors; This indicates the maximum permissible angle of incidence, which can be expressed in radians or degrees, but the unit must be consistent within the same operation. This represents the cosine function. When... Less than If the parent face is facing away from the radar or is excessively squinting, it is directly determined that it will not participate in the direct scattering of the current frame.

[0039] III. Radar Center Projection and Enumeration of Potential Obstruction Surfaces like Figure 2 As shown, Figure 2 This represents the logic for determining projection and occlusion. Indicates the first The target parent triangle of the frame, Index the target parent face; Indicates the face center of the target's parent face; , and Let represent the three unit basis vectors of the local projected coordinate system, where Along the radar line of sight; This represents a two-dimensional projection operator in the local coordinate system corresponding to the target parent plane; This indicates that the candidate is obscuring the parent face. Index the candidate occluded surface; This represents the set of potential occlusion surface indices obtained after projection overlap; Represents two-dimensional local coordinates within the projection plane; and These represent the depths of the candidate occluding surface and the target parent surface along the line of sight at the same projected coordinates; Indicates candidate parent face For the target parent face The resulting two-dimensional occlusion shadow. The dashed branches in the diagram represent exceptions to the endpoints of the target surface; they only correct the intersection of the endpoints of the target surface itself and do not change the general depth order determination.

[0040] Specifically, the exception handling for the endpoint refers to: when determining the intersection point along the line of sight from the radar center to the target's parent face, if the parent face number of a certain intersection point is the same as the target's parent face number, and the absolute value of the distance difference between the intersection point and the target point along the ray direction does not exceed... If the intersection point is not found to be the endpoint of the target parent face, it will be identified as the endpoint of the target parent face itself and exempted from the list of occluded intersection points. The target face will not be considered occluded by itself based on this. This exception only applies to the target parent face itself; it does not apply to edge-adjacent faces, vertex-adjacent faces, or non-adjacent faces, and they will still be determined according to the normal depth order. In this embodiment... for Generally, the tolerance of the target neighborhood is maintained at 1. The two are independent of each other.

[0041] Among them, for any target parent triangle facing the pre-screened face According to radar position and the target father face to face Establish a local projected coordinate system. Let the local unit basis vector be... , and ,in Any point along the radar line of sight The local projected coordinates are defined as follows: (4) In equation (4), Represents any point in three-dimensional space; Point On the target parent side The corresponding two-dimensional projection coordinates on the projection plane, with the coordinate unit being meters; Indicates the center location of the radar; and These represent two orthogonal unit basis vectors in the projection plane; Represents the unit basis vector along the radar line of sight; Point Depth coordinates along the line of sight, in meters; square brackets represent a two-dimensional column vector consisting of two scalar components.

[0042] With target parent face Based on this, perform projection on all other candidate parent triangles. If a candidate parent face... If the projected bounding box or projected polygon overlaps with the projection of the target parent face, then it is included in the set of potential occlusion faces. : (5) In equation (5), Indicates the index of the candidate parent triangle face other than the target parent face; Indicates the target parent face In the The set of potential occlusion surface indices for a frame; Represents the parent face The two-dimensional projection region in the target parent plane projection coordinate system; This indicates that the candidate occluding face cannot use the same parent face index as the target parent face; Represents the intersection of sets; This represents the empty set. The set of potential occlusion surfaces is preferably accelerated using a spatial index structure, but the final result is still determined by the projection overlap relationship.

[0043] IV. Depth Order Determination and Real Occlusion Surface Recognition Not all potentially occluding surfaces will actually occlude the target parent surface. For any local coordinate within the projected overlap region... A candidate occluding surface constitutes true occlusion only when its depth at a given location is strictly less than the depth of the target parent surface. (6) In equation (6), Represents the parent face In the Frame to target parent face The resulting two-dimensional occlusion shadow area; Represents two-dimensional local coordinates within the projection plane; Indicates the candidate occlusion surface in the projected coordinates The depth along the radar line of sight; Indicates the depth of the target parent surface at the same projected coordinates; This indicates the general depth comparison tolerance, in meters, with the preferred option being... ;symbol This means that the candidate blocking surface must be closer to the radar than the target's parent surface.

[0044] For cases where the ray endpoint is located on the target's parent surface itself, to eliminate misjudgments due to target surface self-occlusion, this embodiment employs a specification tolerance that only applies to the endpoint of the target surface itself. The preferred value is When the intersection point is identified as the endpoint of the target parent face itself, only... Modify the normalization endpoint determination without changing the general target neighborhood tolerance in equation (6). This avoids masking the true occlusion relationship by relaxing the global tolerance.

[0045] V. Continuous Shadow Clipping and Visible Sub-region Restoration like Figure 3 As shown, Figure 3 This represents the logic of continuous shadow clipping and orientation-preserving triangulation. This indicates that the shadow is occluded in two dimensions. The occluded region after backmapping to the local 3D coordinates of the target parent face; the large union symbol in the figure represents the index range. All true occlusion shadows within the target face are geometrically unioned, denoted by the function name GeomUnion in the text (7); the set difference symbol in the figure indicates the difference from the target parent face. The union of the shadows is subtracted from the sum, and is represented by the function name GeomDiff in the text (7); This represents the resulting continuous visible sub-region; This indicates the number of sub-triangles obtained after triangulation of the visible sub-region; Index of the sub-triangle face; Indicates the first There are 1 visible sub-triangles. The two input boxes in the figure provide the target parent face and the set of true occlusion shadows, respectively. The two are combined in the shadow union and geometric difference steps; before output, the local coordinates relative to the parent face must be kept in the same order and normal.

[0046] Specifically, the union of the two-dimensional occlusion shadows of all real occlusion surfaces on the projection domain of the target parent surface is obtained, and then subtracted from the target parent surface to obtain the continuous visible sub-region: (7) In equation (7), Indicates the target parent face In the The continuous visible sub-region of a frame; the function name GeomUnion represents the range of indices given after the semicolon. All real occlusion shadows within the set are subjected to a geometric union operation; the function name GeomDiff indicates the geometric difference operation of subtracting the next set from the previous set; the semicolon is used to separate the shadow expression being merged from its index range; This indicates the two-dimensional shaded area. The 3D occlusion shadow region after backmapping to the 3D local coordinates of the target parent surface; represents the set of potential occlusion face indices for traversing the target parent face; the meanings of the other symbols are the same as in equations (5) and (6).

[0047] when When the set is empty, the target parent face is not visible in this frame; when When it is a simply connected polygon, it means that the target parent face retains only one continuous visible region; when When there are multiple polygons, it means that the target parent face is divided into multiple unconnected visible regions by different occlusion shadows. All visible regions retain the corresponding parent face number, boundary winding order, and local two-dimensional coordinates.

[0048] VI. Triangulation and Grading of Extremely Thin Regions To connect to the subsequent electromagnetic integral chain, this embodiment triangulates each visible polygon into a set of sub-triangular faces, and ensures that the vertex order of each sub-triangular face is consistent with that of the parent face. Let the parent face... Derived The triangular facets are ,but: (8) In equation (8), Represents the parent face In the The number of visible sub-triangles obtained after frame triangulation; Indicates the sub-triangle index. ; Indicates from the parent face Derived A visible subtriangle; the function name GeomUnion indicates that the triangle within the parentheses is... to Perform a geometric union operation on all listed visible subtriangles; symbol Indicates indexing by sub-triangle face The intermediate terms are omitted sequentially; the result should restore the complete and continuous visible sub-regions of the parent face.

[0049] After triangulation, visible regions with very small areas or poor numerical conditions are classified into sub-regions. Let the area of ​​each sub-region be... Its conservative upper bound for the complex value of the radar region of interest is The sub-resolution fine stripe region can then be identified using the following formula: (9) In equation (9), Indicates the first The area of ​​each sub-region, in square meters; This represents the upper bound of the conservative contribution of the sub-region to the complex value of the radar region of interest. Its dimensions are consistent with the complex echo representation used, and in the relative scattering model, it can be a dimensionless relative complex amplitude. This represents the area classification threshold, in square meters. This represents the upper bound threshold of the complex contribution, and its dimensions are... Consistency; Symbol Indicates no greater than. Subregions satisfying equation (9) can be classified as sub-resolution stripe regions, and their cumulative area and cumulative contribution upper bound are recorded; preferably, the cumulative area is... Not higher than The total budget.

[0050] VII. Phase Integration Interface and Visible Scattering Contribution Construction like Figure 4 As shown, Figure 4This represents the interface logic for fine-strip region classification, error budgeting, and phase integration. Indicates the first The area of ​​each sub-region; This represents the upper bound of the conservative contribution of the sub-region to the complex value of the radar region of interest. and These represent the area threshold and the upper bound threshold for complex contribution, respectively. Normal regions and resolvable stripes enter the robust triangulation and phase integration branches; subresolution stripes are not forced to perform numerically unstable integration, but only accumulate the area and the upper bound for complex contribution. and Represents two local coordinates within the sub-triangular face; This represents the one-way geometric distance from the radar to that local point; Indicates a one-way reference distance; and This represents the first-order gradient of the distance relative to two local coordinates; This represents the coherent scattering complex contribution of the visible subtriangle; Indicates the amplitude coefficient; Represents the complex scattering coefficient; Represents the imaginary unit; Indicates wave number; Indicates the operating wavelength; This represents an area infinitesimal element. Both classification branches in the diagram ultimately converge to the phase integration interface. The subresolution branch only provides an error budget and does not provide unverified numerical integral values.

[0051] A phase integration interface is established for the visible sub-triangles with non-sub-resolution. Let the local coordinates of the sub-triangles be... The one-way geometric distance of the radar to a local point on the sub-triangle face is expressed as follows under the linear approximation: (10) In equation (10), Indicates that the radar reaches the sub-triangle. The one-way geometric distance of a local point, in meters; This indicates the one-way reference distance to the selected reference point, in meters. and These represent two local coordinate variables within the plane containing the subtriangle; and Representing the geometric distance of a one-way trip relative to and The first-order gradient of , whose unit is meters per local coordinate unit; symbol This indicates that a first-order linear approximation is used.

[0052] Based on equation (10), the coherent scattering complex contribution of the sub-triangle can be written as: (11) In equation (11), Represents the sub-triangle In the The complex contribution of coherent scattering of the frame; This represents the amplitude coefficient, which includes distance attenuation, antenna or system normalization factors; Represents the complex scattering coefficients that are related to the incident angle, material, and polarization; Represents the imaginary unit; Represents the free-space wavenumber, with units of radians per meter; Represents pi; This indicates the radar's operating wavelength, measured in meters; because For one-way distance, the round-trip propagation phase of a single-station radar is written as... ; Represents a complex exponential function; This indicates that a double area integral is performed on the subtriangular region; The area element is represented in square meters. The integration preferably employs analytical linear phase integration, rather than simply relying on dense numerical sampling.

[0053] VIII. Dynamic Continuity Gating and Release Judgment like Figure 5 As shown, Figure 5 This represents the dynamic continuity gating and release decision logic. and This represents a consecutive visible sub-region between two adjacent frames; This represents the area of ​​the symmetrical difference between the visible domains of two frames; Represents the area of ​​the union of the visible regions of two frames; Indicates the proportion of symmetric differences in the visible domain; Indicates the occlusion pair index; and These represent area weight and complex contribution weight, respectively; This indicates a change in the occlusion pairing status. and These represent the changes in occlusion relationships for area-weighted and complex contribution-weighted methods, respectively. The system also simultaneously checks the sub-resolution cumulative error budget, checkpoint signal closure residuals, and repeatability determinism; it determines whether a diamond-shaped branch leads to asset release, or if not, it enters a failure log and stops release.

[0054] Dynamic continuity gating is applied to the visible scattering surface results between adjacent frames. Let the first frame be... Frame and the The visible domain symmetric difference area between frames is The area of ​​the union of the visible domains is Therefore, the domain-symmetric difference ratio is defined as: (12) In equation (12), This represents the dimensionless visible field symmetry difference ratio between two adjacent frames. Indicates the first Frame and the The area of ​​the visible domain symmetry difference, in square meters; Represents the area of ​​the union of the visible regions of two frames, in square meters; subscript Indicates symmetric difference, subscript This represents the union of sets. When... When this occurs, it should be handled according to the pre-defined airspace rules, and division should not be performed directly.

[0055] Let the set of occlusion pairs between adjacent frames be... The shadow area weights corresponding to each element are: Its pairing state change indicator is The corresponding complex contribution weight is The area-weighted change and the complex contribution-weighted change can be expressed as follows: (13) In equation (13), This represents the pairing index in the joint occlusion pairing set of two adjacent frames; This represents the dimensionless change in occlusion relationship weighted by the shaded area; This represents the dimensionless variation in the occlusion relation weighted by the complex scattering contribution; Indicates the first Each occlusion pair has a non-negative area weight; Indicates the first The non-negative complex contribution magnitude weights of each occlusion pair; superscript Indicates area weight, superscript Indicates the weight of complex contributions; This is an indicator of paired state changes; when a change occurs, it is taken as... Otherwise take ;symbol This represents the summation of all paired indices. When any denominator is zero, the corresponding change should be set to zero or marked as inapplicable according to the predefined empty pairing rules.

[0056] Specifically, an occlusion pair is determined by the target parent face number and the candidate occlusion face number, and its identity remains consistent across adjacent frames according to the same ordered pairing. If the pair exists in the deterministic projection overlap enumeration result of the previous frame but not in the next frame, or not in the previous frame but exists in the next frame, then its state is determined to have changed, and the pair is set to... Set to 1; if the pair exists in both frames or does not exist in either frame, then let The value is set to 0. Therefore, changes correspond to the addition or removal of pairs in adjacent frames, rather than solely to an increase or decrease in the total number of pairs. Area weights and complex contribution weights are further used to measure the actual shadow area and complex scattering contribution carried by these pair changes, to avoid misinterpreting representation events caused by changes in zero area, extremely thin area, or contact combinations as real physical transitions.

[0057] At least judge at the time of release , , The system considers the upper bound of the cumulative contribution of sub-resolution stripes and the closure residual of checkpoint signals. If any mandatory threshold is not met, the system saves the failed frame number, target parent face number, relevant occlusion face number, observation value, and threshold value, and stops publishing. In a preferred implementation, the general target neighborhood tolerance can be fixed at [value missing]. The endpoint specification tolerance is fixed as And control the upper bound of the cumulative contribution of sub-resolution fine stripes within Within.

[0058] IX. System Modules and Data Output like Figure 6 As shown, Figure 6 This section describes the system modules and data flow. The input and identity recording module outputs verified dynamic meshes and radar parameters; the candidate scattering surface construction module and orientation pre-screening module output valid target parent surfaces; the projection overlap enumeration module and depth order determination module output true occlusion surfaces and their shadows; the continuous shadow clipping module and orientation-preserving triangulation module output visible sub-triangles; the fine stripe region classification module outputs integrable regions and error budgets; the dynamic continuity gating module outputs pass or fail status; the asset publishing module only publishes dynamic visible scattering surface assets in the pass status, and only saves failure records in the fail status. Data transmitted between modules retains frame indexes, parent surface indexes, and input identities to ensure... Figure 6 The data flow within is traceable.

[0059] The input and identity recording module is used to read the human dynamic mesh, radar parameters and version identity; the candidate scattering surface construction module is used to perform the processing related to equation (2); the orientation pre-screening module is used to perform the processing related to equation (3); the projection overlap enumeration module is used to perform the processing related to equations (4) and (5); the depth order determination module is used to perform the processing related to equation (6); the continuous shadow clipping module is used to perform the processing related to equation (7); the orientation-preserving triangulation module is used to perform the processing related to equation (8); the ultra-thin region classification module is used to perform the processing related to equation (9); the dynamic continuity gating module is used to perform the processing related to equations (12) and (13); the asset publishing module is used to output the final dynamic visible scattering surface asset and failure records.

[0060] The system output data preferably includes: parent triangle facet number, child triangle facet vertex coordinates, child triangle facet local coordinates, facet normal, visible area, projected occlusion pairing, true occlusion facet index, stripe classification, phase integral reference value, dynamic continuity gating result, error budget, and version identity. This output can be used for geometric visibility auditing and as read-only input to a millimeter-wave coherent echo simulation chain.

[0061] Example 2

[0062] This embodiment uses a dynamic mesh sequence of an upper torso digital twin as an example. The number of candidate outer parent triangles is 6223, the dynamic sequence length is 200 frames, and the radar is located at a fixed position in front of the tested human body. Only the outer surface is considered as a candidate scattering surface, excluding internal components such as the lungs, heart, and diaphragm. The target neighborhood tolerance is generally fixed at [value missing]. The target surface itself has its own endpoint specification tolerance. Dynamic continuity gating primarily uses area and complex contribution weighted indicators, rather than relying solely on changes in the number of original occlusion pairs as the sole criterion.

[0063] In this embodiment, after replacing the finite probe with deterministic projection overlap enumeration to discover the occlusion surface, the coherence amplitude, range-dimensional inverse fast Fourier transform (Range-IFFT) energy, region of interest complex value, and range image morphology error of medium and high resolution under the condition of a common visible domain are significantly reduced. This indicates that the main source of error has changed from the original inconsistency of the binary visible domain to a controllable higher-order numerical error. Auditing of the neighborhood of dynamic topological representation events shows that even if the number of original occlusion pairs changes significantly, as long as the changes in area weighting and complex contribution weighting remain at a low level, they can still be interpreted as representation events rather than physical jumps.

[0064] In this implementation, the visible sub-region is divided into normal regions, resolvable striped regions, and sub-resolution striped regions. For sub-resolution striped regions, unstable integrals are skipped only when their cumulative complex contribution upper bound is below the budget; once the cumulative budget exceeds the limit, the local resolution accuracy must be increased or the current frame must be deemed not to pass the publication gating. This maintains engineering computability while avoiding the replacement of explicit error control with unauditable empirical deletion rules.

[0065] This method is applicable to dynamic visible scattering surface determination, coherent echo simulation, range image generation, vital sign modeling, and algorithm regression testing in millimeter-wave radar human digital twins. For different human meshes, different radar attitudes, or different operating frequency bands, the candidate scattering surface rules, orientation thresholds, spatial index structure, stripe classification thresholds, and gating combinations can be adjusted, but they should be fixed before formal operation and saved along with the output assets.

[0066] The visible scattering surface determination results in this method are engineering electromagnetic modeling assets and are not directly equivalent to medical diagnostic conclusions, absolute scattering cross section calibration results, or clinically regulated results. For accurate modeling under specific conditions such as absolute power, absolute signal-to-noise ratio, or specific subject clothing, further integration of material parameters, measured antenna patterns, system constants, and measured calibration data is required.

[0067] The above description is merely one specific embodiment of the present invention. Those skilled in the art can adjust or equivalently replace the candidate scattering surface extraction rules, projection methods, spatial indexing, clipping algorithms, triangulation strategies, fine-strip error budgeting, and dynamic continuity gating combinations without departing from the core ideas of the present invention; all such adjustments should fall within the protection scope of the present invention.

Claims

1. A method for determining the dynamic visible scattering surface of a human body in millimeter-wave radar digital twin, characterized in that, Includes the following steps: Acquire the dynamic outer surface mesh sequence and radar spatial geometric parameters of the human digital twin; Construct a set of candidate scattering surfaces according to preset rules; Through deterministic geometric calculations, all potential occlusion surfaces that have projection overlap with the target surface patch are enumerated; By performing a geometric union operation on the shadow regions of one or more real occluding surfaces on the target surface, and subtracting the shadow union from the target surface patch, a continuous visible sub-region on the target surface patch is obtained. Perform a meshing reconstruction process on the continuous visible sub-regions to maintain the original spatial orientation properties consistent with the original parent surface patch; Dynamic stability gating based on physical weights is applied to the visible scattering surface changes at adjacent time points; Publish a dynamic visible scattering surface asset that includes the geometric information of the continuous visible sub-regions and the quality judgment index of the dynamic stability gating.

2. The method for determining the dynamic visible scattering surface of a human body for millimeter-wave radar digital twins according to claim 1, characterized in that, The dynamic stability gating based on physical weights specifically includes at least one of the following: occlusion relationship variation gating based on shadow area weighting, occlusion relationship variation gating based on complex scattering contribution weighting, or occlusion relationship variation gating based on total target scattering energy.

3. The method for determining the dynamic visible scattering surface of a human body for millimeter-wave radar digital twins according to claim 1, characterized in that, The mesh reconstruction process is a triangulation process. The original spatial orientation attributes include vertex wrap order and face normal. The vertex wrap order and face normal of the sub-region are the same as the vertex wrap order and face normal of its source parent surface patch.

4. The method for determining the dynamic visible scattering surface of a human body for millimeter-wave radar digital twins according to claim 1, characterized in that, The deterministic geometric calculation includes: establishing a local projection coordinate system with the radar position as the observation center, projecting the target surface patch and candidate scattering surfaces onto the same projection plane, and enumerating all potential occlusion surfaces based on the projection overlap relationship.

5. The method for determining the dynamic visible scattering surface of a human body for millimeter-wave radar digital twins according to claim 1, characterized in that, The dynamic visible scattering surface asset also includes: area information of the sub-surface patch, propagation distance information, phase reference information, incident angle information, distance gradient information, as well as the parent surface patch number from which the visible sub-region originates and the real occlusion surface index corresponding to the visible sub-region.

6. The method for determining the dynamic visible scattering surface of a human body for millimeter-wave radar digital twins according to claim 1, characterized in that, Before performing mesh reconstruction processing on the continuous visible sub-regions to maintain the original spatial orientation attributes of their source parent surface patches, the process further includes: performing hierarchical processing on extremely thin visible sub-regions with areas smaller than a preset threshold; classifying extremely thin visible sub-regions that meet a preset precision threshold as resolvable fine strip regions and entering the mesh reconstruction processing flow; classifying extremely thin visible sub-regions that do not meet the preset precision threshold as sub-resolution fine strip regions and recording their area upper bound and complex contribution upper bound but not performing forced mesh reconstruction.

7. The method for determining the dynamic visible scattering surface of a human body for millimeter-wave radar digital twins according to claim 1, characterized in that, The dynamic stability gating is also used to: output a failure record and not publish the dynamic visible scattering surface asset when the quality judgment index does not meet the preset threshold.

8. The method for determining the dynamic visible scattering surface of a human body for millimeter-wave radar digital twins according to claim 1, characterized in that, The radar spatial geometry parameters include the radar transmission and reception spatial location, radar operating wavelength or frequency, candidate surface component identifiers, and parent surface patch numbers used for identity tracking.

9. A dynamic visible scattering surface determination system for millimeter-wave radar human digital twins, used to perform the method as described in any one of claims 1 to 8, characterized in that, include: The data acquisition module is configured to acquire the dynamic outer surface grid sequence of the human digital twin and the radar spatial geometric parameters; The candidate surface construction module is configured to construct a set of candidate scattering surfaces according to preset rules; The occlusion enumeration module is configured to enumerate all potential occlusion surfaces that have projection overlap with the target surface patch through deterministic geometric calculations. The visible region generation module is configured to obtain continuous visible sub-regions on the target surface by performing a geometric union operation on the shadow regions of one or more real occluding surfaces on the target surface and subtracting the shadow union on the target surface sheet. The attribute-preserving reconstruction module is configured to perform a meshing reconstruction process on the continuous visible sub-regions to maintain the original spatial orientation attributes of their source parent surface patches. The stability gating module is configured to perform dynamic stability gating based on physical weights on the visible scattering surface changes at adjacent time points; The asset publishing module is configured to publish a dynamic visible scattering surface asset that includes the geometric information of the continuous visible sub-regions and the quality judgment index of the dynamic stability gating.

10. The dynamic visible scattering surface determination system for millimeter-wave radar human digital twins according to claim 9, characterized in that, The dynamic visible scattering surface asset is a structured data volume, which includes at least: parent surface patch identification information, geometric description information of visible sub-regions, and the quality judgment index field of the dynamic stability gating. All three fields are indispensable and together constitute the asset.

Images