Rail transit vehicle interior part assembly method, device and electronic equipment

Abstract

The application provides a rail transit vehicle interior part assembly method, device and electronic equipment, and belongs to the technical field of rail transit assembly manufacturing. The method comprises the following steps: obtaining three-dimensional point cloud data of an installation environment corresponding to a to-be-assembled interior part; transforming the three-dimensional point cloud data and a standard workpiece three-dimensional model to the same reference coordinate system; solving a target pose under the same reference coordinate system; generating a target contour according to a real boundary; detecting a to-be-removed region of the standard workpiece three-dimensional model existing relative to the target contour in the target pose state; determining a reserved region corresponding to the interior part, and obtaining process attribute information for processing the interior part; generating a process model based on the to-be-removed region, the reserved region, the target contour and the process attribute information; and automatically repairing and adjusting a standard workpiece entity of the interior part to obtain a target interior part. The technical scheme can improve the assembly efficiency and assembly quality consistency of the rail transit vehicle interior part.

Description

Technical Field

[0001] This invention relates to the field of rail transit equipment manufacturing technology, and provides a method, apparatus and electronic equipment for assembling interior parts of rail transit vehicles. Background Technology

[0002] In the field of rail transit equipment manufacturing, the driver's cab of a rail transit vehicle is typically one of the areas with the most complex geometry, the highest assembly precision requirements, and the greatest difficulty in process implementation during the overall vehicle manufacturing process. The driver's cab interior includes not only exterior components such as fiberglass interior trim, composite material roof panels, decorative covers, sealing plates, and inspection covers, but also functional components such as the frame, mounting bases, connecting flanges, air ducts, cable channels, and bolted fixing structures. These structures collectively constitute the driver's cab interior system. After assembly, the driver's cab interior components must not only meet visual consistency requirements but also comprehensively satisfy aerodynamic shape coordination, sound insulation, heat insulation and flame retardancy, and passenger space comfort. Therefore, high requirements are usually placed on the edge gaps, surface differences, surface continuity, and hole matching between interior components.

[0003] Currently, the common practice in assembling driver's interior trim is for assembly workers to transport standard parts to the installation location on-site for trial assembly. If interference is found, lines are manually drawn along the edges of the workpiece. The workpiece is then removed and locally ground and trimmed using a handheld grinder, edge trimmer, or cutting tool, before being transported back to the vehicle for trial assembly. If interference still exists, the cycle of "trial assembly—marking—grinding—trial assembly again" is repeated until it is basically installed. For parts that require drilling, grooving, or a smooth transition with surrounding curved surfaces, local adjustments are also necessary based on experience. This traditional method is inefficient; a complex end-of-line interior trim component often requires repeated transport, trial assembly, and adjustments, with a single part taking several hours or even longer. Summary of the Invention

[0004] This invention provides a method, apparatus, and electronic device for assembling interior components of rail transit vehicles, in order to solve the problem of low efficiency in traditional manual assembly methods.

[0005] This invention proposes a method for assembling interior components of rail transit vehicles, comprising: acquiring three-dimensional point cloud data of the installation environment corresponding to the interior component to be assembled; the installation environment including an installation cavity for installing the interior component; importing a three-dimensional model of a standard workpiece corresponding to the interior component, and transforming the three-dimensional point cloud data and the three-dimensional model of the standard workpiece to the same reference coordinate system; solving the target pose of the three-dimensional model of the standard workpiece in the installation environment based on the three-dimensional model of the standard workpiece and the three-dimensional point cloud data in the same reference coordinate system; extracting the true boundary of the installation cavity from the three-dimensional point cloud data, and generating a target contour based on the true boundary; detecting the area to be removed relative to the target contour in the target pose state of the three-dimensional model of the standard workpiece; determining the area to be retained corresponding to the interior component, and acquiring the process attribute information for processing the interior component; generating a process model based on the area to be removed, the area to be retained, the target contour, and the process attribute information; and automatically fitting the standard workpiece entity of the interior component according to the process model to obtain a target interior component that fits the installation cavity.

[0006] According to an embodiment of the present invention, based on a standard workpiece 3D model and 3D point cloud data, the target pose of the standard workpiece 3D model in the installation environment is solved, including: determining assembly constraints; the assembly constraints include at least one of the following constraints: installation constraints, boundary fit relationships, surface difference control requirements, surface continuity requirements, and installation direction accessibility requirements; based on the standard workpiece 3D model and 3D point cloud data, the target pose of the standard workpiece 3D model in the installation environment that meets the assembly constraints is solved.

[0007] According to one embodiment of the present invention, generating a target contour based on real boundary information includes: biasing the real boundary information based on at least one of preset assembly gap, sealing strip thickness, sealing compensation amount, adhesive layer thickness, thermal expansion compensation amount, adhesive layer compression amount, edge rounding transition requirements, or local process rules to generate the target contour.

[0008] According to one embodiment of the present invention, the installation environment includes an installation cavity and a surrounding associated structure; the surrounding associated structure includes a common reference feature for coordinate transformation; transforming the three-dimensional point cloud data and the standard workpiece three-dimensional model to the same reference coordinate system includes: transforming the three-dimensional point cloud data and the standard workpiece three-dimensional model to the same reference coordinate system according to the common reference feature.

[0009] According to one embodiment of the present invention, automatic fitting of a standard workpiece entity for interior trim parts based on a process model includes: generating a local machining path for the area to be removed based on the area to be removed; automatically fitting the standard workpiece entity according to the local machining path; the fitting includes one or more operations such as local cutting, trimming, drilling, grooving, chamfering or grinding.

[0010] According to one embodiment of the present invention, generating a local machining path for a region to be removed includes: dynamically adjusting at least one of the following machining parameters included in the local machining path based on at least one of the cutting depth, edge normal, material type, local thickness, rate of curvature change, and tool interference risk of the region to be removed: machining strategy, feed rate, path step distance, or tool axis posture.

[0011] According to an embodiment of the present invention, after identifying the area to be removed based on the process model, the method further includes: identifying the area to be removed by at least one of color labeling, layer labeling, attribute field labeling, manufacturing feature labeling, and independent modeling labeling of the cutting body.

[0012] According to one embodiment of the present invention, the process model is at least one of the following: solid model, surface model, mesh model, point cloud model with attributes, independent cut-off volume model, and three-dimensional model with product manufacturing information.

[0013] This invention also provides an assembly device for interior components of rail transit vehicles. The device includes: an environmental perception module for acquiring three-dimensional point cloud data of the installation environment corresponding to the interior component to be assembled; the installation environment includes an installation cavity for installing the interior component; a model import and coordinate unification module for importing a three-dimensional model of a standard workpiece corresponding to the interior component and transforming the three-dimensional point cloud data and the standard workpiece three-dimensional model to the same reference coordinate system; a virtual assembly module for solving the target pose of the standard workpiece three-dimensional model in the installation environment based on the standard workpiece three-dimensional model and the three-dimensional point cloud data under the same reference coordinate system; a geometric difference analysis module for extracting the true boundary of the installation cavity from the three-dimensional point cloud data and generating a target contour based on the true boundary; and detecting the area to be removed relative to the target contour in the target pose state of the standard workpiece three-dimensional model; a process model generation module for determining the retention area corresponding to the interior component and acquiring process attribute information for processing the interior component; generating a process model based on the area to be removed, the retention area, the target contour, and the process attribute information; and a processing execution module for automatically fitting the standard workpiece entity of the interior component according to the process model to obtain a target interior component that fits the installation cavity.

[0014] The present invention also provides an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the program to implement the assembly method of interior parts for rail transit vehicles as described above.

[0015] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the assembly method for interior parts of rail transit vehicles as described above.

[0016] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the assembly method for interior parts of rail transit vehicles as described above.

[0017] This invention provides a method, apparatus, and electronic device for assembling interior components of rail transit vehicles. First, it acquires three-dimensional point cloud data of the installation environment corresponding to the interior component to be assembled. Then, it imports a standard workpiece three-dimensional model of the interior component and transforms the three-dimensional point cloud data and the standard workpiece three-dimensional model to the same reference coordinate system. Under the same reference coordinate system, it solves the target pose of the standard workpiece three-dimensional model in the installation environment and generates a target contour based on the real boundary information and offset information. With the standard workpiece three-dimensional model in the target pose, it detects the areas to be removed relative to the target contour and generates a corresponding process model. Automatic fitting is achieved based on the process model to obtain a target interior component that can accurately fit the installation cavity. Compared to traditional manual assembly methods, this method eliminates the need for repeated manual handling, trial assembly, and adjustments, significantly improving the assembly efficiency of interior components in the driver's cab of rail transit vehicles. Furthermore, it establishes a complete data closed loop from environmental perception, virtual assembly, geometric difference recognition to process model generation and processing execution. By generating a target processing contour for the actual installation cavity, it avoids repeated trial assembly and experience-based fitting, significantly improving the first-time assembly success rate. Furthermore, by introducing fit clearance constraints, surface difference continuity constraints, and installation pose solving mechanisms, this method can make the final assembly gap more uniform and the surface difference more controllable, thus ensuring assembly accuracy and appearance consistency. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the assembly method for interior components of rail transit vehicles provided by the present invention; Figure 2 This is a schematic diagram of the software module structure of the rail transit vehicle interior parts assembly device provided by the present invention; Figure 3 This is a schematic diagram of the software module structure in a specific embodiment of the rail transit vehicle interior component assembly device provided by the present invention; Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention.

[0020] Figure label: 201. Environmental Perception Module; 2011. Data Preprocessing Module; 202. Model Import and Coordinate Unification Module; 203. Virtual Assembly Module; 204. Geometric Difference Analysis Module; 205. Process Model Generation Module; 2051. CAM Interface Module; 206. Machining Execution Module; 410. Processor; 420. Communication interface; 430. Memory; 440. Communication bus. Detailed Implementation

[0021] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0022] The driver's cab of high-speed trains, intercity trains, and other rail transit vehicles is typically one of the most geometrically complex, precision-critical, and technologically challenging areas in the entire vehicle manufacturing process. The cab interior includes not only exterior components such as fiberglass interior trim, composite material roof panels, decorative covers, sealing plates, and inspection covers, but also functional components such as the frame, mounting bases, connecting flanges, air ducts, cable channels, and bolted fixing structures. After assembly, the cab interior trim must not only meet visual consistency requirements but also aerodynamic shape coordination, sound insulation, heat insulation and flame retardancy, and passenger space comfort. Therefore, high requirements are typically placed on edge gaps, surface differences, surface continuity, and hole matching between interior trim components. In actual engineering, many mating positions require gaps controlled between 0.5mm and 2.0mm, surface differences controlled within 1.0mm, and even higher requirements for certain high-grade exterior areas.

[0023] However, during the manufacturing and assembly of high-speed trains, various manufacturing and installation errors are introduced in the preceding processes, such as car body welding, frame installation, fiberglass component positioning, seal bonding, and adjacent interior component installation. These errors are gradually transmitted along the structural and process chains, ultimately manifesting themselves in the assembly process of the closed-loop interior components at the end of the driver's cab. Especially when the part to be assembled is located at the position of the last sealing component or a later finishing accessory, the preceding structure is already fixed, and the installation space cannot be adjusted again, significantly increasing the difficulty of fitting the end part.

[0024] In related technologies, standard interior trim parts in inventory are usually manufactured according to theoretical design models, or, to adapt to on-site assembly needs, a certain amount of allowance is uniformly reserved at the edges. For example, to ensure that they can be finally installed on-site, the edges are often uniformly enlarged by 3mm to 8mm. However, the deviation between the actual installation cavity and the theoretical model is often not a simple translation or uniform scaling, but includes complex characteristics such as local distortion, boundary undulations, attitude coupling offsets, and non-uniform gap changes. Therefore, even if a allowance for modification is reserved, standard parts are still difficult to install directly.

[0025] The current industry practice involves assembly workers moving standard parts to the installation location on-site for trial assembly. If interference is found, lines are manually drawn on the edges of the workpiece. The workpiece is then removed and locally ground and trimmed using a handheld grinder, trimmer, or cutting tool, before being moved back to the vehicle for trial assembly. If interference persists, the cycle of "trial assembly—marking—grinding—trial assembly again" is repeated until the parts are basically assembled. For parts requiring drilling, grooving, or a smooth transition with surrounding curved surfaces, local adjustments are also necessary based on experience. This traditional method, relying on manual assembly, suffers from one or more of the following problems: Low efficiency: A complex end-of-life interior component often requires repeated handling, trial assembly, and repair, with a single component taking several hours or even longer.

[0026] Heavy reliance on human experience: Different workers have different understandings of interference positions, cutting allowances and surface difference control, resulting in poor consistency in assembly quality.

[0027] Limited precision: manual scribing is difficult to accurately reflect the differences in real three-dimensional boundaries, and the grinding process is prone to producing irregular edges, wavy contours and local overcuts, making it difficult to consistently guarantee uniform gaps and high-quality appearance.

[0028] It can easily lead to the scrapping of workpieces: once an excessive amount of material is removed from a certain part, the workpiece cannot be repaired, especially for fiberglass parts, honeycomb sandwich panels and high-cost composite material parts, resulting in significant scrapping losses.

[0029] High labor intensity and poor working environment: The interior space is small, and there is significant pollution from dust, noise and debris, which is not conducive to green manufacturing and occupational health management.

[0030] To address these issues, some solutions in related technologies have introduced digital measurement methods. For example, measurements are taken at the installation site first, and then assembly deviations are observed in software. However, these solutions do not create a truly usable data loop for manufacturing execution. Especially in confined spaces with multiple curved surfaces and mating boundaries, such as the driver's cab of a high-speed train, simply outputting a contour line or a deviation color chart cannot directly guide on-site processing. The object being processed is a solid workpiece with requirements for thickness, curvature, edge transitions, and hole positions; therefore, a solution is needed that can automatically convert measurement data into a "physical area to be processed" and an "executable process semantic model."

[0031] Specifically, the digital solutions in related technologies lack the following key aspects: After obtaining on-site data, how to perform virtual assembly of the standard workpiece model and the actual installation cavity under a unified coordinate system; how to automatically solve for the appropriate workpiece target pose while considering assembly gaps, sealant layers, hole relationships, and surface continuity; how to automatically identify which areas need to be removed, how much to remove, and in what direction to cut; and how to directly output these results as process models or machining instructions that can be recognized by CAM systems and CNC equipment.

[0032] Therefore, this technical field lacks a complete closed-loop adaptive assembly method and system that addresses the complex assembly scenarios of interior trim parts for high-speed train drivers, encompassing everything from on-site scanning to virtual assembly, from difference identification to processing output, and from workpiece repair to assembly verification.

[0033] In view of this, embodiments of the present invention propose a method, apparatus and electronic device for assembling interior parts of rail transit vehicles, in order to solve at least one of the following problems in the existing interior parts assembly process: low efficiency of manual repair, unstable assembly accuracy, easy over-cutting and scrapping, many on-site trial assembly times and difficulty in direct processing execution of digital solutions.

[0034] It should be noted that the method provided in this embodiment of the invention is not only applicable to the interior assembly of newly manufactured vehicles, but also to maintenance and repair scenarios. For example, if a train needs to replace a roof panel after 5 years of operation, the old empty space on the train can be scanned, compared with the new panels in the warehouse, and a processing model can be automatically generated to achieve the installation of spare parts without adjustment.

[0035] The following description, in conjunction with the accompanying drawings, details the assembly method and electronic equipment for interior components of rail transit vehicles proposed in the embodiments of the present invention.

[0036] The method for assembling interior components of rail transit vehicles proposed in this embodiment of the invention, such as... Figure 1 As shown, the method includes: Step 101: Obtain the 3D point cloud data of the installation environment corresponding to the interior parts to be assembled.

[0037] The mounting environment includes at least a mounting cavity for mounting interior trim components. Specifically, in some embodiments, the mounting environment may include the mounting cavity and surrounding associated structures.

[0038] A measurement benchmark is established in the installation environment corresponding to the interior parts to be assembled. Spatial geometric data of the installation cavity and its surrounding related structures are obtained using a 3D scanning device to generate 3D point cloud data of the installation environment.

[0039] Among them, the surrounding associated structure includes at least the adjacent plates that form a boundary mating relationship with the interior parts to be assembled, the mounting base or frame reference for defining the installation depth, and the common reference feature for coordinate transformation.

[0040] Boundary fit refers to the spatial position, contact state, and constraint method formed between the edge of the interior trim part to be assembled and the edge of its adjacent surrounding panels after assembly. For example, boundary fit can include one or more of the following types: Butt Joint, Lap Joint, Rabbit Joint, Flush, Clearance Fit, and Interference Fit.

[0041] Adjacent panels refer to other body or interior panels that are spatially adjacent, edge-to-edge, or overlap with the interior panel to be assembled in the final assembled state. For example, if the interior panel to be assembled is a door panel, then adjacent panels include the door sheet metal (internal metal frame), as well as the adjacent armrest panel, switch panel, etc. Another example is if the interior panel to be assembled is the central display screen trim frame; adjacent panels might include the dashboard body, air vent panel, passenger-side trim strip, etc.

[0042] A mounting base or frame reference refers to a physical limiting structure located deep within or inside a mounting cavity, used to control the depth to which interior trim pieces are installed, providing final, deterministic axial positioning. Specifically, a mounting base is typically a boss, small plane, or specific support structure protruding from the bottom or side wall of the mounting cavity. The function of the mounting base is to ensure that the interior trim piece sits on it when pushed into the cavity. The surface height of the mounting base determines the final position of the interior trim piece in the depth direction. Examples include the bottom support surface of a plastic clip or a slightly raised platform for attaching double-sided tape.

[0043] A frame reference refers to a part of the rigid frame structure of the vehicle body or dashboard, typically a plane, a stepped surface, or a locating pin hole. The frame reference serves as the final reference for the entire mounting system. Features on interior trim components (such as locating pins or support ribs) abut or insert into this frame reference, achieving precise positioning in the depth direction. Examples include a mounting bracket plane on a metal crossbeam inside the dashboard, or a locating hole on the A-pillar sheet metal.

[0044] Common datum features refer to geometric features existing in the scanned area that can be recognized by the measuring equipment and clearly defined and reproduced in different coordinate systems (such as measurement coordinate system, design coordinate system, machining coordinate system, and assembly robot coordinate system). Their function is to accurately transform the local point cloud coordinate system data obtained from scanning into the global design or assembly coordinate system, achieving data alignment and fusion. For example, common datum features can be one or more of the following: three non-collinear spherical sockets, conical holes, edge intersections, fixed datum holes (datum pin holes), specially machined datum planes or datum grooves, etc. The geometric information of common datum features (center point, plane normal, etc.) can be extracted through scanning and best-fitted with corresponding elements in the theoretical design model, thereby completing the coordinate transformation and ensuring that the measured data is consistent with the actual assembly environment and the theoretical design model in spatial position.

[0045] 3D scanning equipment includes one or more of the following: handheld laser scanners, structured light scanners, photogrammetry systems, and fixed scanners.

[0046] It should be noted that the acquisition range of the three-dimensional point cloud data of the installation environment should at least cover the boundary of the cavity to be installed and the surrounding area extending outward by a predetermined distance, so as to ensure that adjacent mating surfaces (such as stop surfaces) and installation references can be completely identified.

[0047] A stop surface refers to a stepped or staggered structure designed on the edge of a part (such as interior trim, panels, or housing). Its main function is to allow two parts to fit together or be positioned, restricting the movement of one part relative to another in a certain direction (e.g., up-down or left-right swaying). It can also conceal gaps, improve sealing, or enhance aesthetics. For example, at the contact point between a car door trim panel and the door frame, there is usually a raised step that fits into a corresponding recess in the door panel; the surface of that step is the stop surface.

[0048] A mating surface refers to the surface of two or more parts that comes into contact, fit together, or form a constraint relationship after assembly. Examples of mating surfaces include stop surfaces, lap surfaces, and positioning surfaces. In this embodiment of the invention, a mating surface refers to the contact surface between the interior trim part to be assembled and an adjacent panel (such as the side contact area), or the contact surface between the interior trim part and the mounting base, or any surface that comes into contact or is constrained by the frame or stop when closed or installed in place. The point cloud data of the mating surfaces is collected during scanning to analyze the actual mating clearance, surface difference, and whether there is interference or gaps in the contact area.

[0049] Step 102: Import the standard workpiece 3D model corresponding to the interior parts, and transform the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system.

[0050] The 3D model of a standard workpiece reflects the actual state of the standard workpiece in stock, including the allowance for repair and fitting, unmachined holes, burrs, or the boundary state of a semi-finished product.

[0051] The 3D model of a standard workpiece is predetermined. It can be obtained through one or more of the following methods: forward modeling from design drawings, reverse modeling from 3D scanning, or direct access from a standard parts library. During assembly, the pre-obtained 3D model of the standard workpiece is imported. The coordinate system used for modeling the 3D model of the standard workpiece is often different from the coordinate system of the 3D point cloud data, and needs to be transformed to the same reference coordinate system.

[0052] There is no single way to achieve coordinate system transformation. For example, coordinates can be uniformly registered using one or more of the following registration algorithms: common marker point registration, installation datum plane registration, structural feature registration, and vehicle coordinate system fusion registration.

[0053] For example, the registration of common marker points is implemented as follows: Around the perimeter of the cavity to be installed (such as adjacent plates, the skeleton reference surface, and the area to be scanned subsequently), attach three or more non-collinear highly reflective markers (usually circular or hemispherical) as common reference features. The positions of these highly reflective markers should be visible from different scanning positions. The 3D scanning equipment identifies these highly reflective markers and automatically fits the sphere center coordinates or circle center coordinates of each marker as a reference. Scanning the installation environment from different angles, since each scan frame contains common highly reflective markers, the software accompanying the 3D scanning equipment automatically stitches the point clouds from multiple frames into a unified measurement coordinate system using bundle adjustment. If the standard workpiece 3D model also has the same highly reflective markers preset, or if the standard workpiece 3D model does not have the same highly reflective markers, a connection can be established through the reference target base. The rigid transformation matrix between the two coordinate point sets of the 3D point cloud data and the standard workpiece 3D model can be calculated, transforming the entire 3D point cloud data into the coordinate system of the standard workpiece 3D model in one go.

[0054] When transforming 3D point cloud data and a standard workpiece 3D model to the same reference coordinate system using the common marker point registration method, the surrounding associated structures of the installation environment must include at least common reference features for coordinate transformation. Specifically, transforming the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system can be done based on common reference features (such as the aforementioned highly reflective marker points).

[0055] The registration method using mounting reference surfaces utilizes the product's inherent positioning surfaces to ensure that the workpiece's posture is consistent with the design reference. The specific implementation is as follows: From the scanned 3D point cloud data of the installation environment, extract the planar point clouds belonging to the mounting base or skeleton reference (e.g., three supporting boss surfaces, or large planes used for positioning). Then, fit the actual planes using the least squares method to calculate the actual normals and positions of these planes in space. Next, find the corresponding theoretical mounting reference surfaces in the standard workpiece's 3D model. Then, perform constraint solving: using the Iterative Closest Point (ICP) algorithm or its variants. Set constraints: only allow translation and rotation, so that the actual planes in the point cloud fit the theoretical planes. This registration method is mainly used to restrict the degrees of freedom of the part (usually used to determine the Z-axis height and the part's pitch and roll attitude).

[0056] The structural feature registration method utilizes the geometric features of the part itself for precise matching, eliminating the need for manual point mapping. The specific implementation is as follows: First, significant features are identified by recognizing common yet unique geometric features from the point cloud data and CAD model. Examples include: large-diameter holes, corners of mating surfaces, specific boss edges, or the contour lines of slots. Then, feature extraction and fitting are performed. For hole features, cylindrical surfaces or circular hole edges are extracted from the point cloud, and the center and axis are fitted. For edges, polyline edges are extracted from the point cloud. Next, feature matching and coordinate system establishment are performed using a three-point alignment (3-2-1 method) or a feature point pair registration algorithm. For example, the 3-2-1 method could involve selecting three coplanar points from both the standard workpiece model and the 3D point cloud data to determine the principal plane (constraining 3 degrees of freedom), then selecting two points to determine the direction (constraining 2 degrees of freedom), and finally selecting one point to determine the origin (constraining the last degree of freedom). Finally, fine-tuning and registration are performed: using feature matching as the initial value, the ICP algorithm is used for global fine-tuning to minimize the error between all fitted features.

[0057] Vehicle coordinate system fusion and registration can be achieved through a combination of at least two of the three methods mentioned above. For example, first, a vehicle reference is introduced: while scanning the mounting cavity, existing RPS (Reference Point Positioning System) reference points or monitoring points (such as positioning holes on the chassis, door hinge mounting points, etc.) on the vehicle are simultaneously scanned. Then, multi-point fusion calculations are performed. If there are common marker points (i.e., common reference features), the local cavity point cloud is connected to a larger point cloud containing the vehicle reference using the common reference features. If the vehicle coordinate system corresponds to a theoretical coordinate value (such as the body frame drawing provided by the car manufacturer), the offset of the local coordinate system relative to the vehicle origin (usually located at the center of the front of the vehicle or the center of the front axle) is calculated by measuring several fixed features inside the vehicle body (such as the distance between the lower ends of the two A-pillars, the fixed point of the dashboard crossbeam). Next, a coordinate transformation script is written to multiply the processed point cloud data by the calculated transformation matrix, converting all coordinate values ​​into values ​​in the vehicle coordinate system.

[0058] Step 103: Under the same reference coordinate system, based on the standard workpiece 3D model and 3D point cloud data, solve the target pose of the standard workpiece 3D model in the installation environment.

[0059] Specifically, in some embodiments, the assembly constraints are first determined, and then the target pose of the standard workpiece 3D model in the installation environment that meets the assembly constraints is solved based on the standard workpiece 3D model and 3D point cloud data.

[0060] Assembly constraints include at least one of the following constraints: installation constraints, boundary fit relationships, surface difference control requirements, surface continuity requirements, and installation direction accessibility requirements.

[0061] For example, based on the installation constraints, boundary fit relationships, surface difference control requirements, surface continuity requirements, and installation direction accessibility requirements of the interior parts to be assembled, virtual assembly calculations are performed on the 3D model of the standard workpiece and the point cloud data of the installation environment to solve the target pose of the standard workpiece in the real installation environment.

[0062] Installation constraints refer to the restrictions on the degrees of freedom of a workpiece based on physical positioning features. For example, it requires that positioning features on a standard workpiece (such as locating pins, clips, and support surfaces) coincide with corresponding references extracted from the point cloud of the installation environment (such as positioning holes, mounting planes, and positioning grooves on the frame) in spatial coordinates. In this way, through point-to-point, surface-to-surface, or shaft-to-hole mating, the final installation position and orientation of the workpiece are forcibly determined, thus restricting the workpiece's translational and rotational degrees of freedom.

[0063] Surface difference control requirements refer to the quantification and directional specification of the degree of flatness between the workpiece and the surfaces of adjacent parts. For example, surface difference control requirements may specify whether the surface of the interior part is higher, flush, or slightly lower than the surface of the adjacent panel at the mating boundary, and the allowable deviation range (e.g., -0.5mm to +0.5mm). In this way, as a rigid geometric constraint in virtual assembly, by optimizing the workpiece's pose, the Z-axis height difference of the mating edge is forcibly controlled within this target range to ensure aesthetic appearance and aerodynamic performance.

[0064] Surface continuity requirements refer to the mathematical smoothness requirements for the light and shadow flow at the junction of a workpiece and its adjacent parts. For example, a surface continuity requirement might require that the surfaces of two parts at their contact boundary mathematically satisfy a specific level of continuity, typically at least tangential continuity. In this way, the virtual assembly algorithm adjusts the workpiece pose to make the slopes (first derivatives) of the surfaces at the joint more consistent, thereby eliminating visual creases or light distortions and achieving a smooth transition.

[0065] The installation orientation accessibility requirement refers to the process feasibility constraint that ensures a standard workpiece can be placed into the mounting cavity along a predetermined path without interference. For example, the installation orientation accessibility requirement restricts the workpiece's assembly trajectory (e.g., it must be downward along the Z-axis, or translated first and then rotated). When solving for the target pose, it is required that the bounding box of the workpiece model does not overlap with the surrounding point cloud model (vehicle structure) throughout the entire motion from the starting point to the ending point. Thus, the installation orientation accessibility requirement prevents the calculated final pose from being perfect in a static state but actually unable to be installed due to obstruction by surrounding structures, ensuring that the pose is physically feasible in the assembly process.

[0066] It should be noted that solving the target pose of the standard workpiece 3D model in the installation environment can be achieved by one or more of the following algorithms: the virtual assembly module uses an improved ICP algorithm, a constrained least squares fitting algorithm, a feature-based iterative optimization algorithm, a multi-objective optimization algorithm, or an intelligent optimization algorithm to solve the target pose that meets the installation conditions.

[0067] For example, when a high-quality initial pose can be obtained, and registration mainly relies on geometric shapes (especially freeform surfaces), an improved ICP algorithm can be used to solve for the target pose. The specific implementation is as follows: Standard ICP minimizes distance by iteratively finding the nearest point between two point clouds. However, standard ICP treats all points equally, which can easily lead to the reference surface being "suspended" or incorrect force on the stop. In this embodiment of the invention, a weighted calculation is performed based on standard ICP: different weights are assigned to different types of points. For example, the weights of locating pins / holes and mounting planes are set to the highest (to ensure accurate positioning), followed by appearance surfaces (to ensure smoothness), while non-mating surfaces (such as heat dissipation fins) have a weight of 0. When calculating errors, directional constraints are used to constrain not only the distance between points but also the consistency of normals (to ensure fit). For example, for the mounting reference surface, it is only allowed to move in the normal direction, and tangential sliding is prohibited. Exemplarily, the specific solution process can first give an initial pose, then match the nearest point and distinguish the point type, then solve the weighted optimization problem, then update the pose, and iterate until convergence.

[0068] In application scenarios where explicit geometric feature pairs (such as face-to-face mating or hole-to-axis alignment) serve as hard constraints, a constrained least-squares fitting algorithm can be used to solve for the target pose, transforming the virtual assembly problem into an optimization problem with equality or inequality constraints. The specific solution method involves constructing an objective function, determining hard constraints, and finding solutions that satisfy these constraints. For example, the objective function can be constructed by minimizing the sum of squares of all mating features (such as the distances from the point cloud of a standard workpiece's 3D model to the point cloud of the target contour). Hard constraints may include planar mating constraints, axial alignment constraints, and surface difference and gap range constraints. During the solution process, the Lagrange multiplier method or Singular Value Decomposition (SVD) is used (for point-pair constraints) to directly find the closed-form solution or exact numerical solution that satisfies all hard constraints and minimizes the objective function.

[0069] In applications where the scanned data is noisy or only a few key features (such as stops and latches) are available, feature-based iterative optimization algorithms can be chosen to solve for the target pose. This optimizes the extracted high-dimensional semantic features (such as planes, cylinders, spheres, and lines) rather than massive point clouds. The specific solution method is as follows: First, feature extraction is performed: geometric primitives (such as the equations of three mounting planes and the center coordinates of a positioning hole) are extracted from the point cloud. Then, feature association is performed: the theoretical features (F...) on the standard part are associated... c1 ,F c2 ) and the actual features obtained from the scan (F) s1 ,F s2A one-to-one correspondence is established. Next, an iterative solution is performed: the distance error between features (such as the angle error between planes, the distance error between circle centers) is calculated under the current pose. Then, R and T are iteratively adjusted using the Gauss-Newton method to minimize the sum of errors for all feature pairs. Here, R (Rotation matrix) represents the rotation matrix, and T (Translation vector) represents the translation vector.

[0070] When multiple potentially conflicting objectives need to be addressed simultaneously (e.g., ensuring a tight fit between the reference surface and uniform gaps in the exterior), a multi-objective optimization algorithm can be used to solve for the target pose. Instead of pursuing the optimal value for a single metric, it seeks a set of optimal pose solutions. The specific solution method is as follows: First, the problem is decomposed into multiple sub-objectives: objective f1 = the fitting error of the mounting base reference surface (the smaller the better); objective f2 = the uniformity of the gap between the mating surfaces (the more uniform the better); objective f3 = the sum of squares of the differences in the exterior surfaces (the smaller the better). During the solution process, evolutionary algorithms such as NSGA-II (Nondominated Sorting Genetic Algorithm II with an elitist strategy) and MOEA / D (Multi-objective Evolutionary Algorithm based on Decomposition) are used to search within the solution space, ultimately outputting a Pareto front. For example, two options can be provided: perfect fit but slightly larger gaps in the exterior, and perfect exterior but slightly floating reference surface.

[0071] In applications where the objective function is complex, non-convex, and non-differentiable (e.g., containing logical judgments such as "whether it collides with a certain structure"), intelligent optimization algorithms can be used to solve for the target pose, simulating the collective intelligence behavior in nature. This eliminates the need for gradient calculations and is insensitive to initial values. Specific solution methods include: for example, using Particle Swarm Optimization (PSO): each pose parameter (6 degrees of freedom, i.e., 3 rotations and 3 translations) is treated as a particle. The particle swarm travels through the solution space, continuously recording the individual best pose and the collective best pose. The final "collective best" is the target pose that maximizes the scores across all constraints. Alternatively, a Genetic Algorithm (GA) can be used: the pose is encoded as a "chromosome." Through selection, crossover, and mutation operations, natural evolution is simulated to select the best pose genes best suited to the real installation environment.

[0072] Step 104: Extract the true boundary of the installation cavity from the 3D point cloud data, and generate the target contour based on the true boundary.

[0073] The true boundary of the installation cavity is extracted from the 3D point cloud data of the installation environment, and the true boundary is offset based on preset assembly gaps, sealing compensation, adhesive layer thickness, process offset, or partitioning process rules to generate a target contour. The target contour may include a target assembly contour and / or a target machining contour.

[0074] Offsetting the real boundary to generate the target contour does not mean that the repaired interior parts perfectly fit the scanned real boundary of the mounting cavity. Instead, based on various assembly process requirements, it involves indenting or expanding the real boundary by a certain amount to calculate a virtual ideal contour line. This calculated virtual contour line is the target assembly contour or target machining contour. The target assembly contour is used to control the final assembled clearance surface difference. The target machining contour guides the cutting distance of the tool / robot during machining.

[0075] The true boundary of the installation cavity can be extracted from the 3D point cloud data. Specifically, one or more of the following extraction methods can be used: curvature change detection, normal change detection, point cloud slice fitting, edge density change recognition, and spline curve reconstruction.

[0076] For applications where boundaries exhibit significant curvature (such as rounded corners or irregularly shaped holes), or where there is a large difference in elevation between the part surface and the background, the curvature abrupt change detection method can be used. Its basic principle is that in three-dimensional space, boundary points are typically points with drastic curvature changes. The specific extraction method is as follows: For the preprocessed point cloud, calculate the principal curvature and Gaussian curvature of each point; set a curvature change threshold, and when the curvature change of a point exceeds the threshold, it is marked as a candidate boundary point; using region growing or clustering algorithms, extract a spatially continuous set of boundary points and filter out isolated noise points. The curvature abrupt change detection method extracts true boundaries, is sensitive to weak boundaries, and can extract fine features.

[0077] For applications where the planar orientations on both sides of a boundary differ significantly (such as at the edge of a step, or the intersection of a vertical wall and the bottom surface), normal vector change detection can be used to extract the true boundary. In smooth regions, the normal vectors of adjacent points change gradually; however, at the boundary, the normal vector direction changes abruptly (e.g., from horizontal to vertical). The specific extraction method is as follows: estimate the normal vector for each point (e.g., through local plane fitting); calculate the angle or dot product between the normal vectors of the current point and its neighbors; when the angle exceeds a preset threshold (e.g., >30°), the point is determined to be a boundary point. Using normal vector change detection for true boundary extraction is computationally efficient and performs well in extracting planar intersections and step edges.

[0078] For application scenarios where cavity boundaries are obstructed or data is missing, making direct extraction unstable, a point cloud slicing fitting method can be used to extract the true boundary. This method uses a set of parallel planes (slices) to "cut" the point cloud, reducing the 3D problem to a 2D contour extraction problem. The specific extraction method is as follows: Generate a set of equally spaced slicing planes along the main direction of the installation cavity (e.g., longitudinal or transverse of the driver's cab); extract the point cloud near each slicing plane (within its bandwidth) and project it onto the slicing plane; perform 2D sorting and spline interpolation on the projected points on each slice to generate the slice contour; connect the extreme points or feature points of all slice contours to reconstruct the 3D boundary. Using the point cloud slicing fitting method for true boundary extraction offers strong noise resistance and can handle situations with missing data.

[0079] For applications where there are significant differences in point cloud density on both sides of a boundary (e.g., sparse points at the far edge and dense points near the edge during scanning), an edge density change recognition method can be used to extract the true boundary. The boundary is typically a region where the point cloud density undergoes a sudden change. The specific extraction method is as follows: Establish a voxel grid or octree, and calculate the point cloud density of each local region; calculate the density gradient between adjacent voxels; extract regions where the density gradient exceeds a threshold and the distribution of neighboring points is discontinuous as boundary candidates. Using the edge density change recognition method for true boundary extraction is insensitive to geometric abrupt changes and is suitable for detecting physical edges caused by the scanning viewpoint.

[0080] For applications requiring the generation of smooth, continuous target contour spline curves for reconstruction, given a discrete set of boundary points, spline curve reconstruction methods can be used to extract the true boundary. The discrete boundary points extracted by these methods are then fitted into high-precision B-spline (B-Spline) or NURBS (Non-Uniform Rational B-Spline) curves. The specific extraction method is as follows: The previously detected candidate boundary points are sorted in spatial order; a least-squares fitting or approximation algorithm is used to generate a spline curve that passes through or approximates these points; smoothing parameters (such as the number of control points and smoothness weights) are set to remove local jitter caused by noise. The final output of this spline curve reconstruction method can be directly used for subsequent bias operations (biasing the true boundary) and CAM (Computer-Aided Manufacturing) processing.

[0081] The target contour is generated based on the real boundary. Specifically, it can be generated by biasing the real boundary information based on at least one of the following: preset assembly gap (or design gap), sealing strip thickness, sealing compensation amount, adhesive layer thickness, thermal expansion compensation amount, edge rounding transition requirements, or local process rules (or zoning process rules).

[0082] Preset assembly clearance is a non-contact safety distance artificially reserved between interior trim parts and adjacent panels (or cavity boundaries). The purpose of preset assembly clearance is to prevent interference caused by manufacturing tolerances, deformation, or assembly errors, while also allowing space for adjustments and thermal expansion / contraction. For example, a clearance of 0.5mm can be set.

[0083] The thickness of the sealing strip is its original cross-sectional height in a free state (not installed, not under pressure). The thickness of the sealing strip serves as a basic input parameter for calculating the sealing compensation amount. For example, the original thickness of the strip is 5mm.

[0084] The sealing compensation amount is a pre-set offset value on the target profile to accommodate the compressed sealing strip. Its purpose is to ensure that the strip is compressed to the appropriate degree after installation, generating sufficient sealing pressure without causing excessive interference that would prevent installation. For example, the sealing compensation amount can be calculated by multiplying the sealing strip thickness by the target compression ratio. For instance, with a sealing strip thickness of 5mm and a target compression ratio of 25%, the compensation amount is approximately 1.25mm.

[0085] The adhesive layer thickness is the final designed thickness of the adhesive between the interior trim and the mounting substrate after curing. The purpose of the adhesive layer thickness is to ensure bond strength. When using adhesive bonding processes, the target contour needs to be recessed inwards to allow for this thickness, providing space for adhesive application. For example: 0.5mm.

[0086] Thermal expansion compensation is an allowance added to the material profile to compensate for dimensional changes caused by temperature variations (such as high temperatures inside a car in summer or low temperatures in winter). The purpose of thermal expansion compensation is to prevent deformation due to compression at high temperatures or excessive gaps at low temperatures caused by thermal expansion and contraction. For example, material thermal expansion coefficient × part length × expected temperature difference = thermal expansion compensation.

[0087] Adhesive layer compression is the thickness of a liquid or paste adhesive layer that is compressed under assembly pressure. The adhesive layer thickness is the final, stable thickness after curing; while adhesive layer compression is the difference between the applied adhesive thickness and the final cured thickness. For example, adhesive thickness = adhesive layer thickness + adhesive layer compression.

[0088] The requirement for rounded corners is to ensure that the processed contour edges have a specific rounded shape (R value) to avoid sharp corners and edges. The purpose of rounded corners is to maintain aesthetics, safety (preventing cuts), avoid stress concentration, or create conditions for subsequent edge wrapping / painting. It should be noted that the requirement for rounded corners is not an offset amount, but rather a requirement for modifying the shape of the contour (e.g., "all edges are rounded with R2 fillets").

[0089] Local process rules are applied to specific local areas (such as corners, near fasteners, or near welds) using offset amounts or special treatment rules that differ from the global rules, based on experience or specific requirements. The purpose of local process rules is to adapt to the unique characteristics of local structures, stresses, or assembly. For example, the offset at inner corners needs to be increased by 0.3mm to prevent crushing. Another example is that offset is prohibited near locating pins to ensure positioning accuracy.

[0090] Offset processing of the true boundary can specifically involve shifting the extracted true boundary curve (or surface) inward or outward along its normal direction by a comprehensive offset calculated from one or more process parameters, thereby generating a new curve, i.e., the target contour. Process parameters include at least one of the following: preset assembly gap, sealing strip thickness, sealing compensation, adhesive layer thickness, thermal expansion compensation, adhesive layer compression, edge fillet transition requirements, or local process rules.

[0091] Specifically, the direction of the offset must first be determined. The offset direction includes inward offset, outward offset, and mixed offset. In this embodiment of the invention, the inward offset direction is set as positive, and the outward offset direction is set as negative. For example, if the preset assembly gap is +0.5mm (required clearance → inward offset), the sealing compensation is +1.0mm (leaving space for compression of the adhesive strip → inward offset), the adhesive layer thickness is +0.3mm (leaving space for adhesive application → inward offset), and the thermal expansion compensation is +0.2mm (reserving space for thermal expansion → inward offset), then the total offset amount d = 0.5 + 1.0 + 0.3 + 0.2 = 2.0mm.

[0092] Step 105: Detect the area to be removed in the standard workpiece 3D model relative to the target contour in the target pose state.

[0093] Geometric difference analysis is performed between the 3D model of the standard workpiece in the target pose and the target contour to identify the solid regions in the standard workpiece that exceed the target contour, and the solid regions are determined as the removal areas to be processed (removed areas).

[0094] In some embodiments, an insufficient size warning is triggered if a local dimension deficiency, material shortage, or inability to meet the minimum safety margin is detected.

[0095] Specifically, for example, geometric difference analysis can be performed using either method one or method two as follows: Method 1: Spatial region determination based on model.

[0096] For example, if both the standard workpiece model and the target contour are closed solid models (or can generate closed cutting bodies), solid Boolean operations can be used. Define the standard workpiece model as the workpiece body, and define the final ideal part generated by extruding or generating the target contour as the target body. Region to be removed = Workpiece body - Target body (Boolean subtraction operation).

[0097] First, generate the target body: stretch or sweep the target assembly contour along the installation direction to form a closed "target entity". This entity represents the space that the workpiece should occupy after repair.

[0098] Then, coordinate alignment is performed to ensure that both the workpiece and the target are in the target pose solved by S3. Next, Boolean subtraction is performed, using the CAD kernel or the kernel of other design software or models to calculate the workpiece-target ratio. The results are extracted; all non-empty parts in the Boolean operation result are the areas to be removed during processing.

[0099] Method 2: Deviation calculation based on point cloud.

[0100] When the model is a curved mesh or difficult to generate closed solids, a point cloud distance field can be used for judgment. The target contour is used as a reference surface, and the signed distance of each point on the standard workpiece surface relative to this surface is evaluated. A positive distance value indicates that the point exceeds the target contour. The specific implementation is as follows: Sampling Standard Workpiece: Uniformly sample the surface of a standard workpiece model to generate a dense point cloud P. part Then calculate the nearest distance: for P part For each point in the target contour surface, calculate its signed distance d' to the nearest point on the surface and define its direction: the normal of the contour surface is set to point outwards from the workpiece. If d' > 0, it means the point is outside the surface, indicating excess material (needs to be removed); if d' < 0, it means the point is inside the surface, indicating insufficient material (requires a warning). Next, generate a difference map: display the distance values ​​using color mapping (e.g., red indicates excess, blue indicates insufficient), and then extract the area to be processed: extract all points where d' > the threshold (e.g., 0.1 mm), perform clustering and boundary extraction to form the area to be removed.

[0101] Optionally, after identifying the area to be removed based on the process model, the area to be removed can be marked, for example, the area with missing material can also be marked. In addition to color marking, marking methods can also include layer differentiation, attribute metadata marking, or generating a separate cut-off body model file to indicate the area to be processed. The area to be processed includes the area to be removed. Specifically, in some embodiments, at least one of the following methods can be used to identify the area to be removed: color marking, layer marking, attribute field marking, manufacturing feature label marking, and independent modeling marking of the cut-off body.

[0102] Color annotation is the process of rendering and displaying the geometric surfaces or solid areas of the region to be removed using specific colors. For example, in CAD (Computer-Aided Design) / CAE (Computer-Aided Engineering) software, RGB color values ​​are set for each surface or solid of the region to be removed, such as red (indicating that it needs to be cut), orange (indicating that it needs to be semi-finished), and blue (indicating that it needs to be retained).

[0103] Layer annotation involves storing the geometric objects of the area to be removed in a separate layer. Layer names typically have a clear process meaning. For example, creating a layer named Layer_Removal_Area or Layer_RoughCut allows you to place all the geometric features (faces, edges, volumes) of the area to be removed into that layer. CAM (Computer-Aided Manufacturing) software can then configure filtering and processing strategies based on layers.

[0104] Attribute field annotations are custom attribute data (key-value pairs) attached to the geometric objects of the area to be removed, storing process-related values ​​or states. For example, attributes can be added to the faces or volumes of the area to be removed in the underlying data of the CAD model, such as: "MachiningAllowance" = 0.5 (finishing allowance 0.5mm), "ProcessType" = "Roughing" (roughing), "ToolDiameter" = 10.0 (recommended tool diameter 10mm), "Priority" = 1 (priority machining area), etc. Attribute field annotations can provide CAM systems with more refined machining command parameters, enabling semi-automatic or fully automatic programming.

[0105] Manufacturing feature labeling involves classifying and labeling the areas to be removed according to their feature types (such as slots, holes, bosses, steps, freeform surfaces, etc.). For example, it automatically identifies the geometry of the area to be removed and labels it with a predefined feature type, such as FeatureType = "Slot", FeatureType = "Hole", FeatureType = "Step", FeatureType = "Freeform", etc. The CAM system can automatically match machining strategies based on the feature type (e.g., trochoidal milling for slots, drilling cycles for holes). Manufacturing feature labeling enables feature-based intelligent machining path planning, significantly reducing manual programming intervention.

[0106] Independent modeling and annotation of the cutting body involves modeling the area to be removed as a separate 3D entity (the cutting body), storing and identifying it separately from the original workpiece model. For example, `RemovalBody.stp` can be extracted from a standard workpiece model using Boolean operations; this entity only contains the portion to be cut off. The cutting body and the original workpiece model maintain a spatial pose relationship (usually within the same reference coordinate system). Independent modeling and annotation of the cutting body enables lightweight data transfer; only the cutting body model needs to be transmitted, not the complete workpiece. Furthermore, machining path generation is simpler; CAM software can directly generate paths targeting the cutting body. After machining, the completeness of the removal can be verified by checking if the cutting body has been removed. Independent modeling and annotation of the cutting body can be used for lightweight data exchange and direct CAM path driving, making it suitable for workshop network transmission or direct interface with CNC systems.

[0107] Step 106: Determine the retention area corresponding to the interior trim parts and obtain the process attribute information for processing the interior trim parts; generate a process model based on the area to be removed, the retention area, the target contour, and the process attribute information.

[0108] The reserved area refers to the part that cannot be cut during the repair process and must be kept as is.

[0109] Standard workpiece 3D models inherently contain some unmachinable areas. In some embodiments, these unmachinable areas can be automatically extracted through feature recognition and retained as reserved areas. For example, unmachinable areas may include one or more of the following: positioning features (e.g., locating pins, locating holes, snap-fit ​​seats, mounting bosses, etc.), functional features (screw posts, reinforcing ribs, limit blocks, etc.), mating features (male / female edges of stop surfaces, lap surfaces, etc.), and appearance surfaces (e.g., Class A surfaces, visible surfaces, etc.). These feature areas are the features that need to be retained within the reserved areas.

[0110] Alternatively, in some other embodiments, the retention area can be derived from the results of the geometric difference analysis. For example, the geometric difference analysis in step S105 above will produce two types of areas: the area to be removed (the portion of the standard workpiece that extends beyond the target contour and needs to be cut off) and the naturally retained area (the portion of the standard workpiece that lies inside the target contour and does not need to be cut off). Naturally retained area = Total standard workpiece - Area to be removed (Boolean subtraction).

[0111] Alternatively, in some other embodiments, protection rules are forcibly defined from the process rule base. The built-in process knowledge base defines some protection rules, such as minimum wall thickness protection rule: cutting is prohibited in areas with a wall thickness of <2mm; functional feature protection rule: cutting is prohibited within 3mm of all positioning holes; appearance area protection rule: any cutting operation is prohibited in the A-side area; structural strength protection rule: cutting is prohibited at the root of the reinforcing rib. Following these rules during processing is equivalent to defining reserved areas through rules.

[0112] Process attribute information refers to a series of parameters that guide how processing equipment performs its operations. The process attribute information for processing interior parts can be obtained in the following ways: Process attribute information can be obtained by matching from the process knowledge base. The system has a built-in process knowledge base that automatically matches and recommends parameters based on material type, processing characteristics, and equipment capabilities. For example, process attribute information can include at least one of the following attribute categories: tool type, tool diameter, spindle speed, feed rate, depth of cut, and processing strategy. For example, the tool type can be a ball end mill, a flat end mill, or a drill bit, and the corresponding process parameters can be matched based on the feature type, such as a ball end mill for curved surfaces and a flat end mill for flat surfaces. As another example, the tool diameter can be Φ6 or Φ10, and the process knowledge base can pre-match the corresponding tool diameter based on the radius of curvature and size of the region.

[0113] It should be noted that if the standard workpiece 3D model is a model with PMI information, some process attributes have already been preset. For example, a surface is marked with a roughness requirement of Ra 1.6, a hole is marked with a tolerance requirement of H7, and an edge is marked with a chamfer requirement of C0.5. These attributes can be read directly from the model without resetting.

[0114] In other embodiments, process attribute information can be obtained from a processing experience database. After obtaining the target interior part, the interior part can be verified, and the verification results can be stored in the process database as feedback data. When similar parts are processed again, the system can query the successful parameters of similar features in the historical processing records, and automatically correct the process attributes of the current operation based on the previous deviation detection results (e.g., if the surface roughness exceeds the standard, the feed rate will be automatically reduced next time).

[0115] Optionally, for areas without preset or special requirements, process engineers can also manually input or modify process attributes in the process model generation interface. For example, this area can be changed to use a Φ3 ball end mill, and the feed rate can be reduced to F=500, etc.

[0116] After determining the area to be retained and obtaining the process attribute information, a corrected interior trim process model is automatically generated based on the area to be removed, the area to be retained, the target contour, and the process attributes related to processing. The process model contains at least one or more of the geometric model information and processing semantic information of the interior trim.

[0117] The essence of generating a process model is to integrate all the previous analysis results—the areas to be removed (where to cut), the areas to be retained (where not to cut), the target contour (what contour to cut to), and the process attribute information (what parameters to use for cutting)—into a three-dimensional model with semantic information that can be directly recognized by the CAM system.

[0118] It should be noted that, in this embodiment of the invention, the process model is an intermediate model generated by comparing a standard workpiece 3D model with 3D point cloud data of the on-site installation environment. There are various ways to implement this; for example, the process model can be a solid model, a surface model, or a point cloud model with attached machining attributes. For instance, in one embodiment, the process model may take one or more of the following forms: solid model, surface model, mesh model, point cloud model with attributes, independent cut-off volume model, or 3D model with PMI information.

[0119] A 3D model with PMI (Product and Manufacturing Information) is a model that adds product manufacturing information to a 3D geometric model. This model can not only express "what the shape is", but also tell downstream processes "how to manufacture and how to inspect".

[0120] A solid model is a complete, closed, volumetric three-dimensional digital model that clearly distinguishes between the "inside" and the "outside".

[0121] Surface models consist only of surface sheets, without thickness or internal volume, resulting in small data volume; they are suitable for representing complex appearance surfaces; however, they cannot perform volume calculations.

[0122] The mesh model is an approximate curved surface composed of a large number of small triangles (or quadrilaterals), which is directly generated from the scan data.

[0123] Attributed point cloud models are based on the original point cloud, with additional information (normal, color, label, etc.) added to each point, while retaining the most original measurement information.

[0124] Independent cut-off volume model only represents the part that needs to be cut off, does not include the geometry of the complete part, has a very small data volume, and only focuses on "excess material"; it needs to be used in conjunction with the original model.

[0125] For example, the specific steps for generating a process model are as follows: Constructing the basic geometric model: Perform Boolean operations between the standard workpiece 3D model and the area to be removed to generate the corrected geometry. Corrected geometry = Standard workpiece model - Area to be removed (Boolean subtraction). If a solid model is used: directly perform the Boolean subtraction operation to obtain a new closed solid. If a mesh model is used: delete the triangular faces of the area to be removed using a trimming algorithm and repair the opening boundaries. Output the corrected geometry (excluding process semantics).

[0126] Overlay reserved region labels (protection constraints): The reserved region is attached to the process model as a non-machinable marker. For example, color annotation can be used to render the reserved region surface as red (prohibiting machining); layer annotation can be used to place the reserved region in a separate layer (such as LAYER_PROTECT); attribute field annotation can be used to add the attribute Machinable=False to each face of the reserved region; manufacturing feature labeling can be used to annotate FeatureType=ProtectionRegion, etc.

[0127] Overlaying the target contour as the machining boundary: The target contour defines the final boundary line of the cut and needs to be clearly identified in the process model. As a cutting path, the contour line directly serves as the boundary constraint of the tool path, and the tool must not cross this line; as an extrusion guide line, the target contour is extruded along the mounting direction into a virtual cutting body, used for the contour machining strategy in CAM.

[0128] Next, attach process attributes (machining semantic information): bind the parameters obtained from the process knowledge base to the corresponding geometric features in the form of PMI or attributes. For example, the process attributes are tool type and diameter, and the bound object is the area to be removed. The corresponding process parameters (machining semantic information) could be Tool=BallEnd_D6. Another example is that the process attributes are spindle speed and feed, and the bound object is the area to be removed. The process parameters are S=12000, F=800, and so on.

[0129] Then, the final process model file is generated. The above geometric and semantic information is output as a file in a specified format. For example, the file format type can be one or more of the following: STEP AP242 (Standard for the Exchange of Product model data - Application Protocol 242), attributed STL (StereoLithography) mesh, independently cut volume model, native CAD format, etc.

[0130] Step 107: Automatically repair the standard workpiece entity of the interior trim according to the process model to obtain the target interior trim that fits the installation cavity.

[0131] Specifically, the automatic fitting of standard workpiece entities for interior trim parts based on the process model can be achieved by generating local machining paths for the areas to be removed, and then automatically fitting the standard workpiece entities according to these local machining paths. Fitting includes one or more operations such as local cutting, trimming, drilling, grooving, chamfering, or grinding.

[0132] The local machining path automatically adjusts the machining strategy, feed rate, path step distance, or tool axis posture based on the cutting depth, edge normal, material type, local thickness, rate of curvature change, and tool interference risk of the area to be machined. In some embodiments, generating a local machining path for the area to be removed can specifically involve dynamically adjusting at least one of the following machining parameters (objects) included in the local machining path based on at least one of the cutting depth, edge normal, material type, local thickness, rate of curvature change, and tool interference risk of the area to be removed: machining strategy, feed rate, path step distance, or tool axis posture. In other words, when planning the local machining path, the cutting depth, edge normal, material type, local thickness, rate of curvature change, and tool interference risk of the area to be machined are analyzed in real time, and the machining strategy, feed rate, path step distance, and tool axis posture are automatically adjusted based on these attributes.

[0133] That is, the local machining path in the embodiments of the present invention can be dynamically adjusted. The machining parameters are automatically and in real time adjusted according to the actual situation of the current cutting position (such as material hardness, surface steepness, whether the tool will hit the workpiece), instead of using a fixed parameter to cut from beginning to end.

[0134] The depth of cut indicates how much material needs to be removed. Larger depths require reduced feed rates and layered cutting. Edge normal indicates the angle and orientation of the cutting edge. It's used to calculate the tool axis posture to avoid overcutting or contact with machined surfaces. Material types include ABS plastic and aluminum alloys. Different materials determine the spindle speed and feed rate. Local thickness indicates the wall thickness at that location on the part. This prevents cutting through (e.g., if the wall thickness is 2mm, the depth of cut cannot exceed 2mm). Rate of curvature change indicates the severity of surface curvature. In areas of severe curvature, the step distance is reduced to ensure a smooth surface. Tool interference risk indicates whether the tool holder or tool tip will collide with the workpiece during machining. It's used to automatically tilt the tool axis to avoid collisions.

[0135] Machining strategies can include selecting the toolpath mode, such as using parallel milling for flat areas and automatically switching to contour milling for steep areas. Feed rate refers to the speed at which the tool moves; for example, faster feed on straight lines (F2000) and automatic speed reduction at corners (F500) to prevent edge chipping. Path step distance represents the distance between two toolpaths; a larger step distance can be used in areas with gentle curvature (to improve efficiency), while automatically reducing the step distance in areas with severe curvature (to ensure accuracy). Tool axis posture represents the tilt angle of the tool in multi-axis machining; for example, to avoid the tool holder hitting the sidewall of a deep cavity, the tool is automatically tilted by 15 degrees during machining.

[0136] The aforementioned dynamic adjustment mechanism for local machining paths enables intelligent and adaptive machining. It automatically slows down and reduces the step distance on complex curved surfaces or thin walls to prevent overcutting or cutting through. On flat surfaces with large allowances, it automatically accelerates and increases the step distance, saving time and improving efficiency. Furthermore, it ensures safety by automatically avoiding interference areas to prevent the tool or spindle from damaging the workpiece.

[0137] Figure 1 The embodiments shown address the aforementioned deficiencies in related technologies by providing an adaptive assembly method, apparatus, and electronic equipment for interior components in areas such as the driver's cab of rail transit vehicles (e.g., high-speed trains). This solves numerous problems in current interior component assembly processes, including low efficiency of manual repair, unstable assembly accuracy, easy over-cutting and scrapping, numerous on-site trial assembly attempts, and the difficulty of directly implementing digital solutions for processing. Furthermore, the solutions proposed in these embodiments employ a technical approach that automatically generates process models based on the actual on-site installation environment and standard workpiece models. This enables the system to automatically identify the area to be removed (hereinafter referred to as the area to be removed), the target contour, and related process attribute information, and directly apply them to local CNC repair, thereby improving the first-time assembly success rate, reducing workpiece loss, and enhancing assembly consistency. Moreover, the solutions proposed in these embodiments essentially establish a universal adaptive repair framework applicable to both new vehicle assembly and maintenance / replacement scenarios, effectively connecting the standard parts inventory model with individualized on-site adaptation.

[0138] Specifically, the method for assembling interior components of rail transit vehicles proposed in this embodiment of the invention has at least the following technical effects: Achieving an automated closed loop from scanning to processing: The solution proposed in this embodiment of the invention does not only collect on-site geometric data, but establishes a complete data closed loop from environmental perception, virtual assembly, geometric difference identification to process model generation and processing execution; Significantly improve the first-time assembly success rate: By generating target machining contours for actual installation cavities, repeated trial assembly and experience-based repairs are avoided, thereby improving the first-time installation qualification rate of end interior parts; Ensuring assembly accuracy and appearance consistency: By introducing fit clearance constraints, surface difference continuity constraints, and installation pose solving mechanisms, the final assembly gap can be more uniform and the surface difference more controllable; Reduce reliance on manual labor and workpiece scrap rate: Replacing manual scribing and grinding with automatic difference analysis and local precision cutting can significantly reduce worker experience dependence, dust pollution, and the risk of overcutting and scrapping.

[0139] Compatible with standard parts inventory management model: The production organization method of "standard parts + partial repair" can take into account both the generality of inventory and the personalized adaptation needs on site, and reduce the manufacturing and management costs of spare parts.

[0140] It has good scalability: the solution can be extended to scenarios such as maintenance, replacement, and upgrade, and is suitable for both new construction and operation and maintenance businesses.

[0141] Optionally, in some embodiments, after step 101, the collected 3D point cloud data of the installation environment can be preprocessed. The data preprocessing specifically includes performing at least one of the following processing operations on the collected environmental data: denoising, outlier removal, stitching and registration, sparsification, resampling, coordinate transformation, boundary enhancement, and region segmentation.

[0142] Optionally, in some embodiments, in Figure 1 Based on the illustrated embodiment, after step 107, an assembly and verification step can be performed: the modified interior trim is assembled to the target installation position, and the assembly gap, surface difference, hole matching, edge continuity, and stress state between it and the surrounding related structures are verified. In some embodiments, the assembly result can be scanned again, and the verification result can be fed back to the process database to form a closed-loop optimization. For example, specifically, in one embodiment, in the assembly and verification step, a scan is performed again after assembly to obtain the deviation between the actual assembly result and the process model, and this deviation is used to correct the boundary offset rules, assembly constraint parameters, or process database of subsequent similar parts.

[0143] Optionally, in some embodiments, in step 105, when it is found that the local boundary of the standard workpiece is located inside the target machining contour and is less than the minimum safety margin threshold, a material shortage alarm, a machining prohibition command, or a compensation suggestion is output.

[0144] Optionally, in some embodiments, after step 106 generates the process model and before step 107, the following steps may also be performed: converting the process model into a data structure, intermediate model, or machining instructions that can be recognized by CNC machining software.

[0145] This invention also provides an assembly device for interior components of rail transit vehicles. The meanings of the technical terms used in each step of this device, their specific implementations, and their effects can be compared with the aforementioned assembly method for interior components of rail transit vehicles, and will not be repeated here. Figure 2 As shown, the device may include the following modules: The environment perception module 201 is used to acquire three-dimensional point cloud data of the installation environment corresponding to the interior parts to be assembled.

[0146] The model import and coordinate unification module 202 is used to import the standard workpiece 3D model corresponding to the interior parts and transform the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system.

[0147] The virtual assembly module 203 is used to solve the target pose of the standard workpiece 3D model in the installation environment based on the standard workpiece 3D model and 3D point cloud data under the same reference coordinate system.

[0148] The geometric difference analysis module 204 is used to extract the true boundary of the mounting cavity from the three-dimensional point cloud data, generate the target contour based on the true boundary, and detect the area to be removed relative to the target contour in the target pose state of the standard workpiece three-dimensional model.

[0149] The process model generation module 205 is used to determine the retention area corresponding to the interior parts and obtain the process attribute information for processing the interior parts; based on the area to be removed, the retention area, the target contour and the process attribute information, a process model is generated.

[0150] The processing execution module 206 is used to automatically repair the standard workpiece entity of the interior part according to the process model to obtain the target interior part that fits the installation cavity.

[0151] In one embodiment, the virtual assembly module 203 is specifically used to: determine assembly constraints; the assembly constraints include at least one of the following constraints: installation constraints, boundary fit relationships, surface difference control requirements, surface continuity requirements, and installation direction accessibility requirements; and, based on the standard workpiece 3D model and 3D point cloud data, solve for the target pose of the standard workpiece 3D model in the installation environment that meets the assembly constraints.

[0152] In one embodiment, the contour extraction module 204 is specifically used to: bias the real boundary information based on at least one of the preset assembly gap, sealing strip thickness, sealing compensation amount, adhesive layer thickness, thermal expansion compensation amount, adhesive layer compression amount, edge rounded corner transition requirements, or local process rules to generate the target contour.

[0153] In one embodiment, the model import and coordinate unification module 202 is specifically used to: transform the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system based on common reference features.

[0154] In one embodiment, the processing execution module 207 is specifically used to: generate a local processing path for the area to be removed based on the area to be removed; and automatically repair the standard workpiece entity according to the local processing path; the repair includes one or more operations such as local cutting, trimming, drilling, grooving, chamfering, or grinding.

[0155] In one embodiment, when generating a local machining path for the area to be removed, the machining execution module 207 is specifically used to: dynamically adjust at least one of the following machining parameters included in the local machining path based on at least one of the cutting depth, edge normal, material type, local thickness, rate of curvature change, and tool interference risk of the area to be removed: machining strategy, feed rate, path step distance, or tool axis posture.

[0156] In one embodiment, the device further includes a labeling module for identifying the area to be removed after identifying the area to be removed based on the process model, using at least one of the following methods: color labeling, layer labeling, attribute field labeling, manufacturing feature labeling, and independent modeling labeling of the cutting body.

[0157] The following is a specific embodiment of the rail transit vehicle interior component assembly device provided by the present invention. In this specific embodiment, such as... Figure 3 As shown, the interior component assembly device for rail transit vehicles includes the following modules: The environmental perception module 201 is used to collect various environmental data, such as 3D point cloud data, image data, and reference information, of the installation environment (or installation area) corresponding to the interior parts to be assembled.

[0158] and Figure 2 Unlike the illustrated embodiment, this specific embodiment may further include a data preprocessing module 2011, which is used to perform denoising, outlier removal, stitching and registration, sparsification, resampling, coordinate transformation, boundary enhancement and region segmentation on the collected environmental data.

[0159] The Model Import and Coordinate Unification Module 202 is used to import the standard workpiece 3D model and establish a unified coordinate relationship between the standard workpiece 3D model and the 3D point cloud data of the installation environment, that is, to convert the standard workpiece 3D model to the same reference coordinate system.

[0160] The virtual assembly module 203 is used to solve the target pose of a standard workpiece based on assembly constraints such as installation datum, stop relationship, surface continuity and assembly direction.

[0161] The geometric difference analysis module 204 is used to extract the true boundary of the installation cavity from the three-dimensional point cloud data, generate the target contour based on the true boundary, calculate the geometric difference and identify the area to be removed, and identify the area with material shortage risk.

[0162] The process model generation module 205 is used to reconstruct the process model based on the recognition results output by the geometric difference analysis module. The process model includes the retained area, the removed area, the target contour, the hole position, and process attribute information.

[0163] The CAM interface module 2051 is used to convert process models into data structures, intermediate models, or machining instructions that can be recognized by CNC machining software.

[0164] The machining execution module 206 is used to control CNC trimming equipment, five-axis machining center, drilling equipment or composite machining equipment to perform partial repairs on standard workpieces.

[0165] The quality verification module 207 is used to inspect the workpiece dimensions, contour deviations, gap uniformity, and surface difference consistency after processing or assembly, and to feed back the inspection results to the aforementioned module. Specifically, the aforementioned module refers to at least one of the following: virtual assembly module 203, geometric difference analysis module 204, process model generation module 205, CAM interface module 2051, and processing execution module 206.

[0166] In summary, the embodiments of the present invention provide a constrained virtual assembly scheme for the complex and confined space of the EMU driver's cab. This scheme does not simply compare the standard workpiece model with the on-site scanning data in a static manner, but solves the optimal target pose of the standard workpiece under constraints such as installation datum fit, continuous peripheral surface difference, and assembly direction reachability. This ensures that the subsequent difference analysis results are more consistent with the actual assembly state, and improves the repair accuracy and the first-time assembly success rate.

[0167] The above-mentioned solution provided by the embodiments of the present invention extracts the real cavity boundary from the point cloud of the installation environment and automatically generates the target processing contour by combining the design gap, adhesive layer compensation, sealing requirements and process offset rules, so that the processing basis is transformed from theoretical numerical model to on-site measured boundary, effectively solving the problem of mismatch between standard parts and actual installation cavity caused by error accumulation.

[0168] The above-described solution provided in this invention also proposes specific technical means for automatic identification of the area to be removed and automatic generation of the process model. By calculating the geometric difference between the standard workpiece model and the target contour, the material removal area exceeding the target boundary is automatically identified, and a process model with color annotations, target contour, cutting amount, and hole position information is generated, achieving automatic conversion from scanning analysis results to a machining semantic model.

[0169] The above-described solution provided in this invention constructs a closed-loop digital process chain from on-site scanning and difference analysis to local CNC repair and assembly verification. This solution can directly drive CAM software to generate local machining paths only for the area to be processed, avoiding traditional manual scribing, repeated trial assembly, and manual grinding, significantly improving assembly efficiency, machining consistency, and standard parts utilization, while reducing dust pollution and the risk of parts scrapping.

[0170] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440. The processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus. The processor 410 can call logical instructions stored in the memory 430 to execute the following methods: The process involves: acquiring 3D point cloud data of the installation environment corresponding to the interior trim component to be assembled; the installation environment including the mounting cavity for mounting the interior trim component; importing the 3D model of the standard workpiece corresponding to the interior trim component and transforming the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system; solving the target pose of the standard workpiece 3D model in the installation environment based on the standard workpiece 3D model and the 3D point cloud data in the same reference coordinate system; extracting the true boundary of the mounting cavity from the 3D point cloud data and generating the target contour based on the true boundary; detecting the area to be removed relative to the target contour in the target pose state of the standard workpiece 3D model; determining the area to be retained for the interior trim component and acquiring the process attribute information for processing the interior trim component; generating a process model based on the area to be removed, the area to be retained, the target contour, and the process attribute information; and automatically fitting the standard workpiece entity of the interior trim component according to the process model to obtain a target interior trim component that fits the mounting cavity.

[0171] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to related technologies, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0172] This invention discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can perform the methods provided in the above-described method embodiments, such as including: The process involves: acquiring 3D point cloud data of the installation environment corresponding to the interior trim component to be assembled; the installation environment including the mounting cavity for mounting the interior trim component; importing the 3D model of the standard workpiece corresponding to the interior trim component and transforming the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system; solving the target pose of the standard workpiece 3D model in the installation environment based on the standard workpiece 3D model and the 3D point cloud data in the same reference coordinate system; extracting the true boundary of the mounting cavity from the 3D point cloud data and generating the target contour based on the true boundary; detecting the area to be removed relative to the target contour in the target pose state of the standard workpiece 3D model; determining the area to be retained for the interior trim component and acquiring the process attribute information for processing the interior trim component; generating a process model based on the area to be removed, the area to be retained, the target contour, and the process attribute information; and automatically fitting the standard workpiece entity of the interior trim component according to the process model to obtain a target interior trim component that fits the mounting cavity.

[0173] On the other hand, embodiments of the present invention also provide a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the methods provided in the above embodiments, including, for example: The process involves: acquiring 3D point cloud data of the installation environment corresponding to the interior trim component to be assembled; the installation environment including the mounting cavity for mounting the interior trim component; importing the 3D model of the standard workpiece corresponding to the interior trim component and transforming the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system; solving the target pose of the standard workpiece 3D model in the installation environment based on the standard workpiece 3D model and the 3D point cloud data in the same reference coordinate system; extracting the true boundary of the mounting cavity from the 3D point cloud data and generating the target contour based on the true boundary; detecting the area to be removed relative to the target contour in the target pose state of the standard workpiece 3D model; determining the area to be retained for the interior trim component and acquiring the process attribute information for processing the interior trim component; generating a process model based on the area to be removed, the area to be retained, the target contour, and the process attribute information; and automatically fitting the standard workpiece entity of the interior trim component according to the process model to obtain a target interior trim component that fits the mounting cavity.

[0174] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0175] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of software products. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0176] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for assembling interior components of a rail transit vehicle, characterized in that, The method includes: Obtain 3D point cloud data of the installation environment corresponding to the interior trim component to be assembled; the installation environment includes an installation cavity for installing the interior trim component; Import the standard workpiece 3D model corresponding to the interior trim, and transform the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system; Under the same reference coordinate system, based on the standard workpiece 3D model and the 3D point cloud data, the target pose of the standard workpiece 3D model in the installation environment is solved; Extract the true boundary of the mounting cavity from the three-dimensional point cloud data, and generate the target contour based on the true boundary; Detect the region to be removed in the standard workpiece 3D model relative to the target contour in the target pose state; Determine the retention area corresponding to the interior trim part and obtain the process attribute information for processing the interior trim part; generate a process model based on the area to be removed, the retention area, the target contour, and the process attribute information; The standard workpiece entity of the interior trim is automatically fitted according to the process model to obtain the target interior trim that fits the installation cavity.

2. The method according to claim 1, characterized in that, Based on the standard workpiece 3D model and the 3D point cloud data, the target pose of the standard workpiece 3D model in the installation environment is solved, including: Determine assembly constraints; the assembly constraints include at least one of the following constraints: installation constraints, boundary fit relationships, surface difference control requirements, surface continuity requirements, and installation direction accessibility requirements; Based on the standard workpiece 3D model and the 3D point cloud data, the target pose of the standard workpiece 3D model in the installation environment that meets the assembly constraints is determined.

3. The method according to claim 1, characterized in that, Generating a target contour based on the real boundary information includes: Based on at least one of the following: preset assembly gap, sealing strip thickness, sealing compensation amount, adhesive layer thickness, thermal expansion compensation amount, adhesive layer compression amount, edge rounding transition requirements, or local process rules, the real boundary information is offset to generate the target contour.

4. The method according to claim 1, characterized in that, The installation environment includes the installation cavity and surrounding associated structures; the surrounding associated structures include common reference features for coordinate transformation; Transforming the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system includes: Based on the common reference features, the three-dimensional point cloud data and the three-dimensional model of the standard workpiece are transformed to the same reference coordinate system.

5. The method according to any one of claims 1-4, characterized in that, The standard workpiece entity of the interior trim is automatically repaired and fitted according to the process model, including: Based on the region to be removed, a local processing path is generated for the region to be removed; The standard workpiece entity is automatically repaired according to the local processing path; the repair includes one or more operations such as local cutting, trimming, drilling, grooving, chamfering or grinding.

6. The method according to claim 5, characterized in that, Generate a local processing path for the region to be removed, including: Based on at least one of the following factors—cutting depth, edge normal, material type, local thickness, rate of curvature change, and tool interference risk—the local machining path is dynamically adjusted to include at least one of the following machining parameters: machining strategy, feed rate, path step distance, or tool axis posture.

7. The method according to claim 5, characterized in that, After identifying the area to be removed based on the process model, the method further includes: The area to be removed is identified using at least one of the following methods: color labeling, layer labeling, attribute field labeling, manufacturing feature labeling, and independent modeling labeling of the cutting body.

8. The method according to claim 1, characterized in that, The process model is at least one of the following: solid model, surface model, mesh model, point cloud model with attributes, independent cut-off volume model, and three-dimensional model with product manufacturing information.

9. An assembly device for interior components of rail transit vehicles, characterized in that, The device includes: An environmental perception module is used to acquire three-dimensional point cloud data of the installation environment corresponding to the interior trim component to be assembled; the installation environment includes an installation cavity for installing the interior trim component. The model import and coordinate unification module is used to import the standard workpiece 3D model corresponding to the interior parts, and transform the 3D point cloud data and the standard workpiece 3D model to the same reference coordinate system. The virtual assembly module is used to solve the target pose of the standard workpiece 3D model in the installation environment based on the standard workpiece 3D model and the 3D point cloud data under the same reference coordinate system. The geometric difference analysis module is used to extract the true boundary of the mounting cavity from the three-dimensional point cloud data, generate a target contour based on the true boundary, and detect the area to be removed of the standard workpiece three-dimensional model relative to the target contour in the target pose state. The process model generation module is used to determine the retention area corresponding to the interior trim part and obtain the process attribute information for processing the interior trim part; based on the area to be removed, the retention area, the target contour and the process attribute information, a process model is generated; The processing execution module is used to automatically repair the standard workpiece entity of the interior trim according to the process model to obtain the target interior trim that fits the installation cavity.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for assembling interior components of rail transit vehicles as described in any one of claims 1 to 8.

Images