A 3D point cloud visualization system for aircraft ducts
The 3D point cloud visualization system solves the problems of data separation and complex operation in aviation duct teaching, which are caused by the separation of 2D and 3D data. It realizes the synchronous display and parameter linkage of 2D and 3D, improves teaching efficiency and reduces costs.
Patent Information
- Application Number
- CN202611263755.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-19
- Publication Date
- 2026-11-17
AI Technical Summary
In existing technologies, aviation duct teaching software is complex to operate, has fragmented two-dimensional and three-dimensional data, lacks a professional parameter library, has uncontrollable point cloud simulation, and has high deployment costs, resulting in low teaching efficiency.
A 3D point cloud visualization system is adopted, including a human-computer interaction module, a 2D pipeline hand-drawing module, a coordinate mapping synchronization module, a 3D surface reconstruction module, a point cloud generation and display module, and a parameterized data management module. This system realizes the linear mapping from 2D maps to a 3D Cartesian coordinate system, generates smooth pipe curves and discrete point cloud data, and stores conduit engineering parameters.
It enables the synchronous display of two-dimensional and three-dimensional data, distinguishes between dimensional errors and geometric deviations, lowers the operational threshold, realizes the linkage between parameter modification and geometric model, reduces deployment costs, and improves teaching efficiency.
Smart Images

Figure CN123118962A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of aviation duct processing technology, and more specifically to a three-dimensional point cloud visualization system for aircraft aviation ducts. Background Technology
[0002] In aero-engine systems, fuel, lubricating oil, hydraulic, and bleed air pipelines form the backbone of the entire transmission medium. The routing, bends, pressure resistance, and temperature tolerance of these pipelines directly determine the engine's reliability and maintainability. Among all aerospace-related vocational colleges and technical schools in China, pipeline blueprint reading, assembly and debugging, and 3D reverse engineering are the core practical training components of the aviation maintenance major, and also among the most challenging modules in terms of teaching difficulty.
[0003] In traditional teaching models, students generally face the following three types of cognitive barriers in learning: (i) High threshold for spatial imagination and software learning: The learning and drawing of two-dimensional multi-view engineering drawings requires the use of orthogonal projection rules to express the standardized spatial pipeline layout. Lower grade students often need several weeks of training to transform the two-dimensional drawings into three-dimensional pipeline forms. A large amount of class time is consumed in drawing recognition training and software learning.
[0004] (ii) Serious disconnect between theory and practice: Commercial CAD software is complex to operate, and students’ learning focus has become the operation of the software rather than professional knowledge. Their understanding and learning of engineering constraints such as pipe materials, pressure bearing, and temperature resistance are completely disconnected from the three-dimensional model.
[0005] (III) Abstraction of reverse training: The point cloud data of 3D laser scanning is a collection of discrete points. Students find it difficult to correspond discrete points with real curved surfaces. However, abstract concepts such as sampling density and scanning noise can only be explained verbally by teachers, lacking intuitive demonstration methods, which leads to teaching difficulties.
[0006] Currently, mainstream 3D modeling software such as CATIA, SolidWorks, and AutoCAD Mechanical were originally designed for engineers working in enterprise design. While they offer extensive functionality, they are also complex to operate and have a high learning curve. Taking the construction of a complete aerospace pipeline model as an example, traditional CAD software requires the following steps: creating a new part file → selecting a sketch plane → drawing a 2D path sketch → exiting the sketch → selecting the sweep command → selecting the path and section → setting parameters such as wall thickness → confirming generation. The entire process requires at least 7-10 clicks and modifications to menus and parameter settings, making it too difficult and complex for teaching purposes. More importantly, these general-purpose modeling software programs lack parameter encapsulation for aerospace ducts. Aerospace standard parameters such as pipe grade, pressure limit, temperature range, minimum bending radius, and fatigue cycle count cannot be linked to the model. Students' learning focus becomes software operation and practice, completely disconnected from professional theory.
[0007] Industrial-grade point cloud processing software such as MeshLab, ParaView, and CloudCompare are designed specifically for quality inspection and data analysis professionals. The software interface is filled with professional modules and technical terms such as error quantification, mesh reconstruction, and normal estimation. The operation logic is very complex and has not been simplified or adapted for teaching scenarios at all.
[0008] More importantly, existing software for point cloud noise simulation generally uses a single overall noise parameter, making it impossible to distinguish the impact of "range measurement error" and "positioning deviation" on point cloud morphology during operation. When explaining the difference between "dimensional deviation" and "geometric deviation," teachers can only describe it verbally, unable to provide a visual demonstration through software. This results in students' understanding of these two core concepts remaining at an abstract level, increasing the difficulty of teaching. A comprehensive analysis of existing technologies reveals the following five major technical shortcomings: (1) Separation of two-dimensional and three-dimensional data: Two-dimensional sketches and three-dimensional models are independent of each other. After modifying the planar layout, it is necessary to remodel. The comparison efficiency in the teaching plan is extremely low.
[0009] (2) The operating threshold is too high: commercial software has a lot of functions but is difficult to learn and is not suitable for quick learning in teaching scenarios.
[0010] (3) Lack of professional parameter library: General software does not encapsulate aviation standard parameters, and engineering constraints are disconnected from geometric models.
[0011] (4) Uncontrollable point cloud simulation: A single noise parameter cannot distinguish the different sources of error, resulting in poor reverse training teaching effect.
[0012] (5) High deployment costs: Commercial software has expensive annual license fees, high costs for batch deployment in data centers, and cannot be used offline. Summary of the Invention
[0013] To address the above problems, this invention proposes a three-dimensional point cloud visualization system for aircraft ductwork.
[0014] The technical solution of the present invention is: a three-dimensional point cloud visualization system for aircraft aviation ducts, comprising a human-computer interaction module, a two-dimensional pipeline hand-drawing module, a coordinate mapping synchronization module, a three-dimensional surface reconstruction module, a point cloud generation and display module, and a parameterized data management module; The human-computer interaction module is used to receive parameters and feed them back to the two-dimensional pipeline hand-drawing module, the coordinate mapping synchronization module, the three-dimensional surface reconstruction module, and the point cloud generation and display module; The 2D pipeline sketching module is used to draw pipeline topology control points on a 2D map and generate smooth pipe bend curves; The coordinate mapping synchronization module is used to perform linear mapping from a two-dimensional atlas to a three-dimensional Cartesian coordinate system based on a smooth pipe curve. The 3D surface reconstruction module is used to generate cylindrical mesh surfaces in a 3D Cartesian coordinate system; The point cloud generation and display module is used to superimpose noise onto the standard cylindrical sampling point set corresponding to the cylindrical mesh surface to generate discrete point cloud data and perform visualization display. The parameterized data management module is used to store conduit engineering parameters.
[0015] Furthermore, in the two-dimensional pipeline hand-drawing module, the Catmull-Rom cubic spline interpolation method is used to generate smooth pipe bend curves.
[0016] Furthermore, in the coordinate mapping synchronization module, the expression for linear mapping is: ; ; ; in, The x-coordinate of a pixel in a three-dimensional Cartesian coordinate system. The ordinate of the pixel in a three-dimensional Cartesian coordinate system. The vertical coordinate of a pixel in a three-dimensional Cartesian coordinate system. The x-coordinate of a pixel on a two-dimensional map. The vertical coordinate of a pixel on a two-dimensional map. The x-coordinate of the reference origin of the graph. The ordinate of the origin of the graph is the reference point. This is the scaling factor. This is the value of the depth slider.
[0017] Furthermore, the 3D surface reconstruction module generates a cylindrical mesh surface by: calculating the tangent vector, principal normal vector, and subnormal vector of each segment along the central spline of the pipeline; constructing a local coordinate system using the tangent vector, principal normal vector, and subnormal vector; performing uniform sampling in each local coordinate system; and stitching them together to obtain the cylindrical mesh surface.
[0018] Furthermore, the tangent vector The expression is: ; ; ; ; in, The original tangent vector, The original tangent vector of the first point, The original tangent vector at the endpoint. The original tangent vector at the midpoint. The coordinates of the end point of the first segment. The coordinates of the starting point of the first segment. This is the last control point of the pipeline. The second to last control point of the pipeline. The next control point after the current intermediate point. The control point preceding the current intermediate point.
[0019] Furthermore, the principal normal vector The expression is: ; ; in, The initial principal normal vector before orthogonal normalization. The normalized principal normal vector from the previous sampling point. This is the normalized unit principal normal vector; binormal vector The expression is: .
[0020] Furthermore, the point cloud generation and display module generates point clouds by superimposing a first layer of radial Gaussian noise onto the standard cylindrical sampling point set corresponding to the cylindrical mesh surface using a radial noise mathematical model, and then superimposing a second layer of triaxial spatial Gaussian noise using a spatial noise mathematical model to generate discrete point cloud data.
[0021] Furthermore, the mathematical model for radial noise is expressed as follows: ; in, , The actual radius of the duct cross-section after adding ranging noise. The theoretical standard catheter radius, This is radial Gaussian ranging noise. The standard deviation of radial noise; The mathematical model for spatial noise is expressed as follows: ; in, , The spatial three-dimensional coordinates of the point after superimposing positioning noise. These are standard sampling coordinate points with no error. This is triaxial synthesized spatial positioning noise. This represents the spatial noise component along the X-axis. This represents the spatial noise component along the Y-axis. This represents the spatial noise component along the Z-axis. The standard deviation of uniaxial spatial noise.
[0022] The beneficial effects of this invention are as follows: This invention adopts a hierarchical overall system architecture that synchronizes two-dimensional hand-drawn topology to three-dimensional entities at millisecond levels, fundamentally solving the problem of data fragmentation between two and three dimensions; it employs an autonomous point cloud simulation core algorithm with radial and spatial dual-dimensional Gaussian noise superposition to achieve separation of dimensional errors and geometric deviations in teaching; it introduces the Frenet pipeline surface reconstruction algorithm with Gram-Schmidt orthogonal correction to solve the problems of large-curvature pipeline frame flipping and surface distortion; it adopts an integrated structured data model of process parameters, geometric topology, and point cloud acquisition to achieve global parameter modification; this invention also integrates a standardized duct parameter library with five built-in types of HB and GJB navigation beacons to achieve a closed-loop teaching system integrating theory and practice. Attached Figure Description
[0023] Figure 1 This is a flowchart of a 3D point cloud visualization system for aircraft ductwork. Detailed Implementation
[0024] The embodiments of the present invention will be further described below with reference to the accompanying drawings.
[0025] like Figure 1 As shown, the present invention provides a three-dimensional point cloud visualization system for aircraft air ducts, including a human-computer interaction module, a two-dimensional pipeline hand-drawing module, a coordinate mapping synchronization module, a three-dimensional surface reconstruction module, a point cloud generation and display module, and a parameterized data management module. The human-computer interaction module is used to receive parameters and feed them back to the two-dimensional pipeline hand-drawing module, the coordinate mapping synchronization module, the three-dimensional surface reconstruction module, and the point cloud generation and display module; The 2D pipeline sketching module is used to draw pipeline topology control points on a 2D map and generate smooth pipe bend curves; The coordinate mapping synchronization module is used to perform linear mapping from a two-dimensional atlas to a three-dimensional Cartesian coordinate system based on a smooth pipe curve. The 3D surface reconstruction module is used to generate cylindrical mesh surfaces in a 3D Cartesian coordinate system; The point cloud generation and display module is used to superimpose noise onto the standard cylindrical sampling point set corresponding to the cylindrical mesh surface to generate discrete point cloud data and perform visualization display. The parameterized data management module is used to store conduit engineering parameters.
[0026] The human-computer interaction module is used to receive user hand-drawn sketches, parameter adjustments, view switching, noise parameter configuration, and point cloud export operation commands, which are then uniformly distributed to various functional modules and provide real-time feedback on the visualization of 2D canvas, 3D surface, and discrete point cloud.
[0027] The system's underlying layer employs a three-in-one structured dictionary data model of "process parameters, geometric topology, and point cloud sampling," encapsulating and storing a complete set of information, including the duct's engineering parameters, 2D coordinates, 3D coordinates, and point cloud configuration. All modules share the same data source, ensuring that the global view is updated synchronously when parameters are modified or the topology changes, fundamentally guaranteeing the speed of 2D and 3D synchronization and greatly reducing latency.
[0028] In this embodiment of the invention, the two-dimensional pipeline hand-drawing module uses the Catmull-Rom cubic spline interpolation method to generate smooth pipe bend curves.
[0029] The 2D pipeline sketching module is implemented using Qt's Graphics View component, and the curve smoothing kernel employs the Catmull-Rom cubic spline interpolation algorithm. Students can generate topology control points by clicking on the 2D canvas. These control points can also be used as monitoring points; their coordinates can be adjusted by long-pressing the mouse, and points can be deleted by double-clicking.
[0030] Catmull-Rom spline interpolation algorithm principle: For four consecutive control points ,exist and The parametric equations for the curve segments between them are: ; in, It is a parameter.
[0031] M is the Catmull-Rom basis matrix: M=[[0,2,0,0],[-1,0,1,0],[2,-5,4,-1],[-1,3,-3,1]]; The unfolded form is as follows: ; Convert to Bessel control point formula: ; ; ; ; in, and The endpoints of the curve (passing through control points). and As control points, they are drawn from the previous and next control points respectively. The denominator 6 is an empirical value, ensuring curve C... 1 Continuous and with moderate tension.
[0032] For the first and last points of the curve, an endpoint replication strategy is used: , This ensures that the entire pipeline is free of broken line distortion. The algorithm has a small floating-point operation requirement, allowing it to run smoothly even on low-spec computers.
[0033] In this embodiment of the invention, the expression for linear mapping in the coordinate mapping synchronization module is: ; ; ; in, The x-coordinate of a pixel in a three-dimensional Cartesian coordinate system. The ordinate of the pixel in a three-dimensional Cartesian coordinate system. The vertical coordinate of a pixel in a three-dimensional Cartesian coordinate system. The x-coordinate of a pixel on a two-dimensional map. The vertical coordinate of a pixel on a two-dimensional map. The x-coordinate of the reference origin of the graph. The ordinate of the origin of the graph is the reference point. This is the scaling factor. This is the value of the depth slider.
[0034] The coordinate mapping synchronization module enables linear mapping from a 2D side view to a 3D Cartesian coordinate system, specifically adapted for 2D hand-drawn engine tutorials. The mapping rules are: 2D X coordinate → 3D X-axis (horizontal), 2D Y coordinate → 3D Z-axis (vertical, Y-axis reversed), depth → 3D Y-axis (front and back direction).
[0035] When a user adds, deletes, or moves any path control point, the module can complete all coordinate calculations within milliseconds and send the complete 3D path array to the 3D reconstruction module. The latency for synchronizing the two views can be controlled to within 10 milliseconds, significantly reducing system resource consumption.
[0036] In this embodiment of the invention, the three-dimensional surface reconstruction module generates a cylindrical mesh surface by: calculating the tangent vector, principal normal vector, and subnormal vector of each segment along the central spline of the pipeline; constructing a local coordinate system using the tangent vector, principal normal vector, and subnormal vector; performing uniform sampling in each local coordinate system; and stitching together to obtain a cylindrical mesh surface.
[0037] The 3D surface reconstruction module is implemented based on the matplotlib 3D open-source rendering engine. The surface generation kernel adopts the classic Frenet-Serret differential geometry algorithm and adds a Gram-Schmidt orthogonalization recursive strategy, which specifically solves the problems of frame flipping and surface distortion that are prone to occur in pipelines with large curvature.
[0038] The algorithm calculates the tangent vector, principal normal vector, and binormal vector segment by segment along the central spline of the pipeline. These three orthogonal vectors are used to construct a local coordinate system that moves along the path. In each local coordinate system, samples are uniformly taken along the circumference, and finally, a complete and smooth cylindrical mesh surface is stitched together.
[0039] The parallel transmission method ensures that the normal vector rotates smoothly along the curve; otherwise, a sudden flip could cause errors in the piping. In contrast, the traditional Frenet frame directly uses the second derivative to calculate the normal vector, which can easily lead to a sudden 180-degree flip at large curvatures or inflection points, causing surface distortion and errors.
[0040] The module includes a complete 3D benchmark model of an aero-engine, encompassing five core structures: the air intake, compressor, combustion chamber, turbine, and exhaust nozzle, which can serve as a spatial reference for piping assembly. Interactively, it supports full-degree-of-freedom 3D observation, allowing users to rotate the view with the left mouse button, pan with the right mouse button, and zoom using the scroll wheel.
[0041] In this embodiment of the invention, the tangent vector The expression is: ; ; ; ; in, The original tangent vector, The original tangent vector of the first point, The original tangent vector at the endpoint. The original tangent vector at the midpoint. The coordinates of the end point of the first segment. The coordinates of the starting point of the first segment. This is the last control point of the pipeline. The second to last control point of the pipeline. The next control point after the current intermediate point. The control point preceding the current intermediate point.
[0042] In this embodiment of the invention, the principal normal vector The expression is: ; ; in, The initial principal normal vector before orthogonal normalization. The normalized principal normal vector from the previous sampling point. This is the normalized unit principal normal vector; binormal vector The expression is: .
[0043] In this embodiment of the invention, the point cloud generation and display module generates point cloud data by superimposing a first layer of radial Gaussian noise onto the standard cylindrical sampling point set corresponding to the cylindrical mesh surface using a radial noise mathematical model, and then superimposing a second layer of triaxial spatial Gaussian noise using a spatial noise mathematical model to generate discrete point cloud data.
[0044] The point cloud generation and display module is the core innovative module of this system. It adopts a simulation algorithm that superimposes radial and spatial Gaussian noise, which is fundamentally different from the traditional single noise simulation.
[0045] When the algorithm runs, it first generates a standard cylindrical sampling point set without errors, and then superimposes the first layer of radial Gaussian noise to simulate the size deviation of the laser ranging itself; then it superimposes the second layer of triaxial spatial Gaussian noise to simulate hand-held shaking and positioning offset of the scanning device.
[0046] Point cloud generation algorithm flow: Traverse each path segment (two adjacent control points); Calculate the segment length: ; Calculate the number of axial sampling points based on density: ; Calculate the number of sampling points in the circumferential direction: ; For each axial position and each circumferential angle, calculate the coordinates of a point on the cylindrical surface; Add radial noise and spatial position noise.
[0047] In this embodiment of the invention, the expression for the radial noise mathematical model is: ; in, , The actual radius of the duct cross-section after adding ranging noise. The theoretical standard catheter radius, This is radial Gaussian ranging noise. The standard deviation of radial noise; Radial noise simulates the measurement error of the laser rangefinder itself. The radius direction of each point can be disturbed independently, which will directly affect the "thickness" and dimensional accuracy of the point cloud.
[0048] The mathematical model for spatial noise is expressed as follows: ; in, , The spatial three-dimensional coordinates of the point after superimposing positioning noise. These are standard sampling coordinate points with no error. This is triaxial synthesized spatial positioning noise. This represents the spatial noise component along the X-axis. This represents the spatial noise component along the Y-axis. This represents the spatial noise component along the Z-axis. The standard deviation of uniaxial spatial noise.
[0049] Spatial noise simulates the positioning jitter and attitude error of the scanning equipment. The entire point cloud is offset in space, affecting the "shape and position accuracy" of the point cloud rather than its dimensional accuracy.
[0050] The magnitudes of the three independent normal variables follow a Maxwell-Boltzmann distribution: ; ; The final coordinate formula for each point cloud point synthesized from the two-dimensional noise is: ; The first part is radial noise, acting on the cylindrical surface, and the second part is the overall spatial offset. The two noise sources have different physical meanings and different effects on the point cloud morphology: radial noise mainly affects the accuracy of the cross-sectional profile of the point cloud, while spatial noise mainly affects the accuracy of the overall position of the point cloud.
[0051] The total number of points along the entire catheter is: ; in, This represents the total length of the conduit. There is no upper limit to the point cloud sampling density, allowing the generation of multi-level discrete point cloud data ranging from sparse to millions of levels, suitable for layered practical training.
[0052] The standard deviations of the two types of noise can be adjusted separately. By simply modifying the parameters, the three different point cloud effects of high-precision industrial scanning, ordinary teaching scanning, and simple low-precision scanning can be reproduced.
[0053] In this embodiment of the invention, the parameterized data management module adopts a parameter-driven topology linkage architecture at its underlying level, using a structured dictionary to uniformly store complete engineering data for five types of aviation ducts. It includes five standardized parameter templates for main fuel, return fuel, afterburner fuel, starting fuel, and drain fuel. Each set of data contains fields specified by aviation standards, such as duct number, total length, number of bends, inner and outer diameters, wall thickness, material, pressure range, high and low temperature operating range, self-weight, and medium function description.
[0054] Table 1 is a comparison table of five types of standard catheter parameters.
[0055] Table 1
[0056] When a user modifies any process parameter, the system will automatically and synchronously modify the pipeline geometry and point cloud sampling range, enabling the linked updates of geometry and sampling data when parameters change.
[0057] The formula for calculating the pressure bearing capacity of a thin-walled cylinder is: ; in, For circumferential stress, For internal pressure, For pipe diameter, This refers to the wall thickness. When the user modifies the pressure range parameters, the system will automatically calculate the required minimum wall thickness based on the material's allowable stress, and then update the pipe diameter and wall thickness parameters accordingly.
[0058] To further improve the smoothness of interaction, the system adopts an incremental local redrawing mechanism. Traditional 2D / 3D synchronous solutions use a global recalculation strategy: every time a control point is modified, all 3D coordinates of the entire pipeline need to be recalculated and the entire surface redrawn. This solution results in a linear increase in computational load when the pipeline is long and there are many control points, which can easily cause stuttering on devices with integrated graphics cards.
[0059] The incremental local redrawing mechanism of this system includes: (1) Dirty control point marking: When a control point is modified, the point and the two control points before and after it are marked as "dirty" areas, and only the three-dimensional coordinates within the influence range of these 5 control points are recalculated. Since the Catmull-Rom spline interpolation is only affected by the two control points before and after it, the curve segments outside this range are completely unaffected and do not need to be recalculated.
[0060] (2) Dirty rectangle marker: At the 3D rendering level, calculate the bounding box of the affected surface area, and only redraw the primitives within the bounding box, rather than the entire viewport.
[0061] (3) Double buffer pre-rendering: Maintain two rendering buffers before and after the current frame is rendered. The display buffer is switched immediately after the current frame is rendered and the next frame is rendered in the background buffer to avoid screen tearing and stuttering.
[0062] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.
Claims
1. A three-dimensional point cloud visualization system for aircraft ducts, characterized in that, It includes a human-computer interaction module, a 2D pipeline hand-drawing module, a coordinate mapping synchronization module, a 3D surface reconstruction module, a point cloud generation and display module, and a parametric data management module; The human-computer interaction module is used to receive parameters and feed them back to the two-dimensional pipeline hand-drawing module, the coordinate mapping synchronization module, the three-dimensional surface reconstruction module, and the point cloud generation and display module. The two-dimensional pipeline hand-drawing module is used to draw pipeline topology control points on a two-dimensional map and generate smooth pipe curves; The coordinate mapping synchronization module is used to perform linear mapping from a two-dimensional map to a three-dimensional Cartesian coordinate system based on a smooth pipe curve. The three-dimensional surface reconstruction module is used to generate cylindrical mesh surfaces in a three-dimensional Cartesian coordinate system; The point cloud generation and display module is used to superimpose noise on the standard cylindrical sampling point set corresponding to the cylindrical mesh surface to generate discrete point cloud data and perform visualization display. The parameterized data management module is used to store conduit engineering parameters.
2. The three-dimensional point cloud visualization system for aircraft ducts according to claim 1, characterized in that, In the two-dimensional pipeline hand-drawing module, the Catmull-Rom cubic spline interpolation method is used to generate smooth pipe bend curves.
3. The three-dimensional point cloud visualization system for aircraft aeronautical ducts according to claim 1, characterized in that, In the coordinate mapping synchronization module, the expression for linear mapping is: ; ; ; in, The x-coordinate of a pixel in a three-dimensional Cartesian coordinate system. The ordinate of the pixel in a three-dimensional Cartesian coordinate system. The vertical coordinate of a pixel in a three-dimensional Cartesian coordinate system. The x-coordinate of a pixel on a two-dimensional map. The vertical coordinate of a pixel on a two-dimensional map. The x-coordinate of the reference origin of the graph. The ordinate of the origin of the graph is the reference point. This is the scaling factor. This is the value of the depth slider.
4. The three-dimensional point cloud visualization system for aircraft ducts according to claim 1, characterized in that, The three-dimensional surface reconstruction module generates a cylindrical mesh surface by: calculating the tangent vector, principal normal vector, and subnormal vector of each segment along the central spline of the pipeline; constructing a local coordinate system using the tangent vector, principal normal vector, and subnormal vector; performing uniform sampling in each local coordinate system; and stitching them together to obtain the cylindrical mesh surface.
5. The three-dimensional point cloud visualization system for aircraft aeronautical ducts according to claim 4, characterized in that, The tangent vector The expression is: ; ; ; ; in, The original tangent vector, The original tangent vector of the first point, The original tangent vector at the endpoint. The original tangent vector at the midpoint. The coordinates of the end point of the first segment. The coordinates of the starting point of the first segment. This is the last control point of the pipeline. The second to last control point of the pipeline. The next control point after the current intermediate point. The control point preceding the current intermediate point.
6. The three-dimensional point cloud visualization system for aircraft ducts according to claim 4, characterized in that, The principal normal vector The expression is: ; ; in, The initial principal normal vector before orthogonal normalization. The normalized principal normal vector from the previous sampling point. This is the normalized unit principal normal vector; The secondary normal vector The expression is: 。 7. The three-dimensional point cloud visualization system for aircraft ductwork according to claim 1, characterized in that, The point cloud generation and display module generates point clouds by superimposing a first layer of radial Gaussian noise onto the standard cylindrical sampling point set corresponding to the cylindrical mesh surface using a radial noise mathematical model, and then superimposing a second layer of triaxial spatial Gaussian noise using a spatial noise mathematical model to generate discrete point cloud data.
8. The three-dimensional point cloud visualization system for aircraft ducts according to claim 7, characterized in that, The mathematical model for radial noise is expressed as follows: ; in, , The actual radius of the duct cross-section after adding ranging noise. The theoretical standard catheter radius, This is radial Gaussian ranging noise. The standard deviation of radial noise; The expression for the spatial noise mathematical model is as follows: ; in, , The spatial three-dimensional coordinates of the point after superimposing positioning noise. These are standard sampling coordinate points with no error. This is triaxial synthesized spatial positioning noise. This represents the spatial noise component along the X-axis. This represents the spatial noise component along the Y-axis. This represents the spatial noise component along the Z-axis. The standard deviation of uniaxial spatial noise.
