A multi-source data fusion rail three-dimensional profile measurement method, device and system

Abstract

The present application belongs to the field of rail transit detection technology, and particularly relates to a multi-source data fusion rail three-dimensional profile measurement method, device and system, comprising the following steps: acquiring rail section profile data, high-low irregularity data and track irregularity data, and recording mileage information of each measurement position; selecting a local highest point as a vertical splicing reference point; selecting a sampling point horizontal coordinate as a horizontal splicing reference point; constructing a profile calibration model with a smoothing constraint, taking the high-low irregularity data and the track irregularity data as a reference, and calculating a vertical offset and a horizontal offset of each section; and performing global coordinate conversion on the original rail section profile data at each mileage position according to the calculated vertical offset and horizontal offset, to generate a continuous rail three-dimensional point cloud model with a unified spatial reference. The present application aims to solve the three-dimensional reconstruction distortion problem caused by the lack of global height and horizontal reference in the traditional laser scanning technology.

Description

Technical Field

[0001] This invention belongs to the field of rail transit inspection technology, specifically relating to a method, device, and system for measuring the three-dimensional profile of rails using multi-source data fusion. Background Technology

[0002] The three-dimensional profile of rails is one of the core parameters for assessing track geometry and guiding maintenance. Its accuracy directly affects the safety and stability of train operation. In recent years, with the development of high-speed railways and heavy-haul transportation, higher requirements have been placed on the spatial continuity, absolute height consistency, and inspection efficiency of rail profile measurement. Currently, the mainstream rail profile inspection methods mostly adopt line laser scanning technology: through line laser sensors installed on the track inspection vehicle, the rail cross-sectional profile is collected point by point during movement, and the position information is recorded by a mileage encoder. Finally, multiple two-dimensional cross-sections are spliced ​​along the longitudinal direction of the track to form an approximate three-dimensional surface model. This method has the advantages of being non-contact, high-resolution, and highly automated, and has been widely used in track inspection equipment.

[0003] Problems with existing technology:

[0004] However, this method has a fundamental flaw: line laser sensors can only acquire the relative height and lateral relationships between points within a cross-section, but cannot determine the absolute height and orbital position of the cross-section in the global vertical coordinate system. Therefore, when multiple cross-sections are stacked along the mileage direction, the lack of a unified vertical and lateral reference leads to problems such as height drift, step-like distortion, or smoothness distortion in the reconstructed 3D profile in the longitudinal direction. Especially in long-distance measurements, the accumulation of small errors can significantly reduce overall geometric accuracy, making it difficult to meet the demands of modern, refined track condition management.

[0005] Traditional three-dimensional profile measurement of rails typically relies on a single laser scanner for cross-sectional acquisition, but this method has the following problems:

[0006] The lack of longitudinal height and lateral displacement references between the sections leads to relative height and track drift;

[0007] It is difficult to accurately reflect the spatial continuity and smoothness of the track;

[0008] It is impossible to effectively integrate elevation / track alignment irregularities to improve the overall geometric modeling accuracy. Summary of the Invention

[0009] The purpose of this invention is to provide a method, device, and system for measuring the three-dimensional profile of rails by multi-source data fusion. It aims to solve the problem of three-dimensional reconstruction distortion caused by the lack of global height and lateral reference in traditional laser scanning technology. By integrating a line laser sensor, a high-precision height and unevenness measurement device, and a mileage encoder, it achieves high-precision and high-continuity three-dimensional modeling of the rail geometry.

[0010] The specific technical solution adopted by this invention is as follows:

[0011] A method for measuring the three-dimensional profile of rails through multi-source data fusion includes the following steps:

[0012] Rail cross-sectional profile data are acquired at multiple discrete mileage locations using a line laser scanning module;

[0013] The unevenness measurement module acquires rail elevation and alignment unevenness data at each mileage location, and the mileage positioning module records the mileage information at each measurement location.

[0014] Select the highest local point in the profile of each rail section as the vertical splicing reference point;

[0015] Select the x-coordinate of the sampling point at the working edge feature position in the rail section profile as the reference point for horizontal splicing.

[0016] Construct a profile calibration model with smoothing constraints, and calculate the vertical and lateral offsets of each section based on the elevation and track irregularity data.

[0017] The original rail cross-sectional profile data at each mileage location are transformed globally according to the calculated vertical and lateral offsets to generate a continuous three-dimensional point cloud model of the rail with a unified spatial reference.

[0018] According to another aspect of the present invention, the profile calibration model with smoothing constraints includes a data fidelity term and a smoothing regularization term, and the reconstructed three-dimensional profile is optimized to both conform to the non-smoothness measurement benchmark and maintain longitudinal smoothness.

[0019] According to another aspect of the present invention, the data fidelity term makes the vertical and lateral offsets of each section as close as possible to the theoretical calculated values, and the smoothing regularization term penalizes abrupt changes in offset between adjacent sections.

[0020] According to another aspect of the present invention, the line laser scanning results are used as supplementary sensor data, and together with the displacement sensor data in the one-chord N-point chord measurement method, the elevation irregularity data and the track alignment irregularity data are calculated.

[0021] According to another aspect of the present invention, the global coordinate transformation includes: performing coordinate transformation on each contour point in the original rail cross-section contour data according to the calculated vertical and lateral offsets to obtain the three-dimensional coordinates of the point in the global coordinate system, and splicing and integrating the three-dimensional point sets corresponding to all mileage positions in mileage order to form a continuous rail three-dimensional point cloud model.

[0022] According to another aspect of the present invention, a multi-source data fusion three-dimensional profile measurement device for rails is also provided, comprising:

[0023] Line laser scanning module is used to acquire rail cross-sectional profile data at multiple discrete mileage locations;

[0024] The irregularity measurement module is used to obtain rail elevation irregularity data and rail alignment irregularity data at each mileage location using the one-chord N-point chord measurement method.

[0025] The mileage positioning module is used to record mileage information at each measurement location;

[0026] The data fusion processing module, connected to the line laser scanning module, the irregularity measurement module, and the mileage positioning module, is used to perform spatiotemporal alignment of the rail cross-section contour data with the elevation irregularity data and the track alignment irregularity data, and to calculate the global offset of each cross-section based on the vertical reference point and the lateral reference point, thereby realizing the reconstruction of the three-dimensional profile of the rail.

[0027] According to another aspect of the present invention, the vertical reference point is a local highest point in the rail profile, and the lateral reference point is the abscissa of a sampling point at a working edge feature position in the rail profile.

[0028] According to another aspect of the present invention, the data fusion processing module constructs a profile calibration model with smoothing constraints. The profile calibration model includes a data fidelity term and a smoothing regularization term, which are used to optimize the solution of the vertical and lateral offsets of each section.

[0029] According to another aspect of the present invention, a rail three-dimensional profile measurement system based on multi-source data fusion is also provided, comprising the rail three-dimensional profile measurement device described in any one of the preceding embodiments, and:

[0030] A transport platform is used to carry the measuring device and move it along the track;

[0031] The data storage module, connected to the data fusion processing module, is used to store the reconstructed three-dimensional point cloud model of the rail.

[0032] The display terminal is connected to the data storage module and is used to visualize the three-dimensional profile of the rail.

[0033] According to another aspect of the present invention, the scanning data of the line laser scanning module and the displacement sensor data of the irregularity measurement module are fused together to jointly calculate the track elevation irregularity and track alignment irregularity data, thereby improving the accuracy of irregularity measurement.

[0034] According to another aspect of the present invention, an electronic device is also provided, the electronic device including a memory and a processor; the memory is used to store a program; the processor executes the program to implement the method described in any one of the foregoing.

[0035] According to another aspect of the present invention, a computer-readable storage medium is also provided, the storage medium storing a computer program that, when executed by a processor, implements the method described in any one of the preceding embodiments.

[0036] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the method described in any one of the preceding embodiments.

[0037] The technical effects achieved by this invention are as follows:

[0038] This invention utilizes high-precision data on elevation and track irregularities as vertical / lateral references to achieve efficient data fusion and a three-dimensional rail reconstruction algorithm.

[0039] This invention solves the problem of lacking longitudinal height and lateral displacement benchmarks between sections by integrating high-precision elevation / track alignment irregularity data as a global benchmark, thus eliminating relative height and track alignment drift. Through an optimized calibration model with smoothing constraints, the invention suppresses the effects of measurement noise and identification errors, ensuring that the reconstructed 3D profile conforms to the irregularity measurement benchmark while maintaining longitudinal smoothness, truly reflecting the spatial continuity of the track. By using line laser scanning results as a supplementary sensor, combined with displacement sensor data from the one-chord N-point chord measurement method, the invention improves overall measurement accuracy, achieving data cross-calibration and accuracy mutual protection. The optimized algorithm has high computational efficiency, is suitable for real-time processing in embedded systems, meets on-site inspection requirements, and provides a reliable data foundation for rail wear analysis, profile comparison, and maintenance decisions.

[0040] This invention, through the reconstruction of the three-dimensional profile of the rail, achieves global consistency and longitudinal smoothness, effectively eliminating height drift and step-like distortion problems, and can accurately reflect the actual geometric state of the rail. Attached Figure Description

[0041] Figure 1 This is a flowchart of the method of the present invention;

[0042] Figure 2 This is the process for measuring the three-dimensional profile of the rail in this invention;

[0043] Figure 3 This is a schematic diagram of the structure of the three-dimensional profile measuring device for rails in this invention. Figure 1 ;

[0044] Figure 4 This is a schematic diagram illustrating how laser scanning results are used to supplement the accuracy of the N-point string measurement method in this invention.

[0045] Figure 5 This is a schematic diagram of the rail profile at the mileage location in this invention;

[0046] Figure 6 This is a schematic diagram of the three-dimensional profile reconstruction of the rail in this invention;

[0047] Figure 7 This is a rendering of the three-dimensional profile of the rail in this invention;

[0048] Figure 8 This is a schematic diagram of the structure of the three-dimensional profile measuring device for rails in this invention. Figure 2 ;

[0049] Figure 9 This is a schematic diagram of the three-dimensional profile measurement system for rails in this invention. Detailed Implementation

[0050] To make the objectives and advantages of this invention clearer, the invention will be specifically described below with reference to embodiments. It should be understood that the following text is merely used to describe one or more specific embodiments of the invention and does not strictly limit the scope of protection specifically claimed by the invention.

[0051] It should be noted that the terms "first," "second," etc., used in this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0052] According to an embodiment of the present invention, a method embodiment for measuring the three-dimensional profile of rails by multi-source data fusion is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0053] like Figure 1 and Figure 2 As shown, a method for measuring the three-dimensional profile of rails through multi-source data fusion includes the following steps:

[0054] S1. Obtain rail cross-sectional profile data at multiple discrete mileage locations using a line laser scanning module;

[0055] S2. Obtain rail elevation and alignment irregularity data at each mileage location through the irregularity measurement module, and record the mileage information at each measurement location through the mileage positioning module.

[0056] S3. Select the highest local point in the profile of each rail section as the vertical splicing reference point;

[0057] S4. Select the abscissa of the sampling point at the working edge feature position in the rail section profile as the reference point for horizontal splicing.

[0058] S5. Construct a profile calibration model with smoothing constraints, and calculate the vertical and lateral offsets of each section based on the elevation and track irregularity data and the track irregularity data.

[0059] S6. Perform global coordinate transformation on the original rail cross-section contour data at each mileage location according to the calculated vertical and lateral offsets to generate a continuous three-dimensional point cloud model of the rail with a unified spatial reference.

[0060] Data collection:

[0061] like Figures 1 to 3 As shown, the rail 3D profile measurement system includes a line laser scanning module, an irregularity measurement module, and a mileage positioning module. In steps S1 and S2, the line laser scanning module acquires rail cross-sectional profile data at multiple discrete mileage locations with fixed sampling intervals during the movement of the rail inspection vehicle. The rail cross-sectional profile data at each mileage location can be represented as a point set. ,in The x-axis represents the profile. The vertical axis represents the profile. Indicates the mileage location index. Indicates the index of sampling points on the cross-section (see reference). Figure 5 ).

[0062] Meanwhile, the irregularity measurement module uses the one-chord N-point chord measurement method (based on Chinese invention patent: Track Waveform Determination Method, Device and Electronic Equipment, Storage Medium, Publication No. CN112033316B) to obtain rail elevation irregularity data at each mileage location. and track irregularity data In the one-chord N-point chord measurement method, the distance from the rail surface to the reference chord is measured by a displacement sensor to obtain high-precision data on elevation irregularities and track alignment irregularities.

[0063] By using line laser scanning results as supplementary sensor data, and fusing them with displacement sensor data from the one-chord N-point chord measurement method, such as... Figure 4 As shown, the line laser scanning results can directly obtain the distance from the rail surface to the reference chord, which can be regarded as the n+1th displacement sensor. Furthermore, the height / rail direction irregularity data can be calculated by using the one-chord (N+1) point chord measurement method, which significantly improves the measurement accuracy of track irregularity.

[0064] Reference point selection:

[0065] In steps S3 and S4, to achieve precise alignment between the sections, suitable reference points need to be selected. The highest local point in the profile of each rail section is chosen as the vertical splicing reference point, i.e. For example, the x-coordinate of a sampling point 16mm below the top surface of the rail is selected as the reference point for lateral splicing, i.e.: The highest point in this area corresponds to the top of the rail. Under normal operating conditions, it has geometric characteristics that make it easy to identify and has good stability.

[0066] For the lateral splicing reference point, the abscissa of the sampling point at the working edge feature position in the rail cross-section profile is selected. The working edge is the main contact area between the rail and the wheel, and its geometric features have high stability across different cross-sections. Selecting this feature position as the lateral reference point can effectively avoid identification errors caused by wear or contamination on the rail head side, and improve the reliability of lateral alignment.

[0067] Profile calibration:

[0068] like Figure 6 As shown, if noise and measurement errors are disregarded, the mileage position... Rail profile data obtained from the location The reconstruction result can be expressed as: However, considering the noise and detection errors in the measurement process, such as the reference point in the line laser cross-section... (Such as the top of the rail and the working edge) may be inaccurately identified due to dirt, obstruction or noise; Although accurate, it is slightly out of sync with laser measurements in time and space;

[0069] The aim is to further suppress local anomalies and make the three-dimensional profile more consistent with the physical continuity of the orbit;

[0070] Therefore: In step S5, the mathematical expression for constructing the profile calibration model with smoothing constraints is: in, and These are the vertical and lateral global offsets to be solved, respectively, used to determine the displacement mileage of the i-th cross section (i.e., the displacement mileage). (Measurement results of rail profile at the location) The entire structure is lifted / moved laterally to the correct absolute position, where M is the total number of cross-sections (total number of profiles). This is a parameter for controlling the smoothness intensity.

[0071] The profile calibration model with smoothing constraints consists of two parts:

[0072] Data fidelity items: Make the vertical and lateral offsets of each section as close as possible to the theoretical calculation values. Ideally, if the measurement data is completely accurate, these theoretical values ​​can be used directly for calibration. However, in actual measurements, reference points may be inaccurately identified due to rail surface oil contamination, abnormal laser reflection, or noise interference, and direct use may lead to "jumping" phenomena in individual cross-sections.

[0073] Smoothing regularization terms: The constraint penalizes abrupt shifts in offset between adjacent sections. This constraint is based on the physical characteristics of the track—the actual rail should have good smoothness in the longitudinal direction, and the changes in height and lateral position between adjacent sections should be gradual and continuous. If the difference in offset between adjacent sections is too large, it indicates that the reconstructed 3D profile has "steps" or "jitter," which does not conform to the actual geometry of the track.

[0074] The choice of the smoothing parameter λ is crucial: when λ→0, the model completely trusts the original measurement data and does not perform smoothing; when λ→∞, all cross-sections are forced to have the same offset, resulting in rigid translation. In practical applications, a moderate value for λ is chosen, achieving a balance between fidelity and smoothing through cross-validation or empirical values.

[0075] The above optimization problem is an unconstrained quadratic programming problem, and the objective function is about and All are convex functions, and there exists a unique global optimum. The problem can be transformed into matrix form: in, D is Given the first-order difference matrix, setting the gradient to 0, we obtain the system of linear equations: The coefficient matrix of the linear equation system is a symmetric positive definite sparse banded matrix, which can be solved quickly using the conjugate gradient method or the Thomas algorithm. It has high computational efficiency and is suitable for real-time processing in embedded systems.

[0076] 3D reconstruction:

[0077] In step S6, the original rail cross-sectional profile data at each mileage location are adjusted according to the calculated vertical offset. and lateral offset Perform a global coordinate transformation for the first... Mileage location Each contour point Its three-dimensional coordinates in the global coordinate system are: in, To obtain the mileage coordinates of this section, a complete 3D point cloud data is generated. All 3D point sets corresponding to each mileage location are then stitched together and integrated in mileage order to form a continuous 3D point cloud model of the rail with a unified spatial reference, covering the entire track section. and These are the optimal vertical and lateral offset adjustments obtained by solving the profile calibration model with smoothing constraints.

[0078] Based on the above steps, by fusing high-precision elevation / track alignment irregularity data as a global benchmark, the problem of lacking longitudinal height and lateral displacement benchmarks between sections is solved, eliminating relative height and track alignment drift phenomena. Through an optimized calibration model with smoothing constraints, the effects of measurement noise and identification errors are suppressed, ensuring that the reconstructed three-dimensional profile conforms to the irregularity measurement benchmark while maintaining longitudinal smoothness, truly reflecting the spatial continuity of the track. By using line laser scanning results as a supplementary sensor, and combining them with displacement sensor data from the one-chord N-point chord measurement method to calculate irregularity data, the overall measurement accuracy is improved, achieving data cross-calibration and accuracy mutual protection. The optimized algorithm has high computational efficiency, is suitable for real-time processing in embedded systems, meets on-site inspection needs, and provides a reliable data foundation for rail wear analysis, profile comparison, and maintenance decisions.

[0079] Furthermore, the reconstructed three-dimensional profile of the rail possesses global consistency and longitudinal smoothness, effectively eliminating height drift and step-like distortion problems. It accurately reflects the actual geometric state of the rail and can be used as a reference. Figure 7 The restored result image.

[0080] As an optional embodiment, refer to Figure 3 and Figure 8 It also includes a multi-source data fusion rail three-dimensional profile measurement device, comprising:

[0081] Line laser scanning module is used to acquire rail cross-sectional profile data at multiple discrete mileage locations;

[0082] The irregularity measurement module is used to obtain rail elevation irregularity data and rail alignment irregularity data at each mileage location using the one-chord N-point chord measurement method.

[0083] The mileage positioning module is used to record mileage information at each measurement location;

[0084] The data fusion processing module, connecting the laser scanning module, the irregularity measurement module, and the mileage positioning module are used to perform spatiotemporal alignment of rail cross-section contour data with elevation and lateral irregularity data and track alignment irregularity data, and calculate the global offset of each cross-section based on vertical and lateral reference points to realize the reconstruction of the three-dimensional profile of the rail.

[0085] Reference Figure 8 In one alternative embodiment, the rail three-dimensional profile measuring device adopts a modular design, with each functional module working together to achieve high-precision three-dimensional profile measurement.

[0086] The line laser scanning module includes, but is not limited to, employing a high-precision line laser sensor, mounted at the bottom of the measurement system at an appropriate distance from the rail surface. During the measurement system's movement, the line laser sensor emits a laser line at a fixed frequency, illuminating the rail surface to form a light stripe. The deformed image of the light stripe is captured by a CCD or CMOS imaging element, and the point cloud data of the rail cross-section contour is calculated using triangulation principles. The line laser scanning module features non-contact measurement characteristics and can adapt to different rail types and surface conditions.

[0087] The irregularity measurement module includes a reference chord, multiple displacement sensors, and a data acquisition unit. The displacement sensors are arranged at equal intervals along the reference chord to measure the distance from the rail surface to the reference chord in real time. The line laser scanning result is regarded as the data of the N+1th displacement sensor and fused with the data of the traditional displacement sensor to further improve the accuracy of irregularity measurement.

[0088] The mileage positioning module combines a high-precision encoder with an inertial navigation unit. The encoder is installed on the detection wheel axle to record the number of wheel rotations in real time and convert them into accurate mileage. The inertial navigation unit provides auxiliary positioning information to compensate for mileage errors caused by wheel diameter changes and slippage, ensuring that each measurement section can be accurately assigned spatial position information, providing a reliable spatiotemporal reference for subsequent data fusion.

[0089] The data fusion processing module consists of a high-performance embedded processor, a storage unit, and a data transmission interface. It receives cross-sectional contour data from the line laser scanning module, elevation / track-direction irregularity data from the irregularity measurement module, and position information from the odometer positioning module, and performs the following key processing steps:

[0090] Spatiotemporal registration: Based on mileage information, data collected by different modules are synchronized in time and aligned in space to ensure accurate matching of multi-source data at the same location;

[0091] Reference point extraction: Automatically identify the highest local point in the profile of each section as the vertical reference point, and identify the feature position of the working edge as the lateral reference point;

[0092] Offset calculation: Construct a profile calibration model with smooth constraints and solve for the vertical and lateral global offsets of each section;

[0093] Coordinate transformation: The original cross-sectional data is transformed to a unified global coordinate system based on the calculated offset to generate a 3D point cloud model.

[0094] The various modules of the rail 3D profile measurement device are tightly connected through a high-speed data bus. The raw data from the line laser scanning module and the irregularity measurement module are transmitted to the data fusion processing module in real time. The synchronization signal of the mileage positioning module ensures that all data have a unified timestamp. This hardware architecture design guarantees the real-time performance and integrity of data processing, enabling the device to generate a high-precision rail 3D profile model in real time during the movement of the inspection vehicle, providing a basis for track condition assessment.

[0095] As an optional embodiment, and referring to Figure 9 A multi-source data fusion-based three-dimensional profile measurement system for rails includes a three-dimensional profile measurement device for rails, and:

[0096] A transport platform used to carry the measuring device as it moves along a track;

[0097] The data storage module, connected to the data fusion processing module, is used to store the reconstructed three-dimensional point cloud model of the rail.

[0098] The display terminal is connected to the data storage module and is used to visualize the three-dimensional profile of the rail.

[0099] Based on the above, refer to Figure 9 In one optional embodiment, the rail three-dimensional profile measurement system is a complete detection platform built on the aforementioned measurement device, realizing an integrated operation process from data acquisition to result display.

[0100] The transport platform includes, but is not limited to: mobile platforms for short distances, standard track inspection vehicle platforms, and tracked / trackless flying platforms, providing a stable operating foundation and installation space for the entire measurement system. The platform is equipped with precision vibration isolation devices to effectively isolate vibration interference caused by track irregularities, ensuring that the line laser scanning module and the irregularity measurement module maintain a stable measurement posture during vehicle operation. The platform integrates a power supply system, environmental protection devices, and an operation control room, providing the necessary power support, temperature and humidity control, and operating environment for the measurement equipment.

[0101] The data storage module adopts a distributed storage architecture, comprising a high-speed cache unit and a large-capacity storage unit. The high-speed cache unit is used for temporary storage of raw data acquired in real time, ensuring that no measurement information is lost even at high detection speeds. The large-capacity storage unit is used for long-term storage of the processed 3D point cloud model and raw measurement data. The storage module employs data compression and redundant backup mechanisms, saving storage space while ensuring data security. All data is organized according to mileage location and indexed for easy subsequent rapid retrieval and analysis.

[0102] The display terminal consists of two parts: an operation control interface and a data visualization interface. The operation control interface is used for system parameter setting, measurement process monitoring, and anomaly handling, providing operators with intuitive system status feedback. The data visualization interface focuses on the display and analysis of the 3D point cloud model, supporting multiple display modes, including a global bird's-eye view, a magnified local view, a profile analysis view, and a multi-dimensional color mapping view. Through this interface, technicians can intuitively observe the 3D geometry of the rails, identify abnormal areas, and perform professional analyses such as wear calculation and profile comparison. The display terminal also supports exporting measurement results, allowing 3D point cloud data to be exported in a standard format for further analysis and processing by other professional software.

[0103] The rail 3D profile measurement system adopts a modular design, with each component connected through standardized interfaces. This ensures the overall coordination of the system while providing good scalability and maintainability. The system architecture allows the measuring device to fully leverage its technical advantages, transforming high-precision rail 3D profile measurement technology into a practical engineering tool and providing a basis for railway maintenance.

[0104] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0105] According to another aspect of the present invention, an electronic device is also provided, the electronic device including a memory and a processor; the memory is used to store a program; the processor executes the program to implement the method of any of the foregoing.

[0106] According to another aspect of the present invention, a computer-readable storage medium is also provided, the storage medium storing a computer program that, when executed by a processor, implements the method of any of the foregoing.

[0107] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the method described in any of the foregoing.

[0108] The above description is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention are implemented according to conventional methods in the art unless otherwise specified or limited.

Claims

1. A method for measuring the three-dimensional profile of rails through multi-source data fusion, characterized in that, Includes the following steps: Rail cross-sectional profile data are acquired at multiple discrete mileage locations using a line laser scanning module. The unevenness measurement module acquires rail elevation and alignment unevenness data at each mileage location, and the mileage positioning module records the mileage information at each measurement location. Select the highest local point in the profile of each rail section as the vertical splicing reference point; Select the x-coordinate of the sampling point at the working edge feature position in the rail section profile as the reference point for horizontal splicing. A profile calibration model with smoothing constraints is constructed. Based on elevation and track irregularity data, the vertical and lateral offsets of each section are calculated. The profile calibration model with smoothing constraints includes a data fidelity term and a smoothing regularization term. Through optimization, the reconstructed three-dimensional profile conforms to the irregularity measurement benchmark while maintaining longitudinal smoothness. The data fidelity term makes the vertical and lateral offsets of each section as close as possible to the theoretical calculation values, while the smoothing regularization term penalizes abrupt changes in offset between adjacent sections. The original rail cross-section contour data at each mileage location are transformed globally according to the calculated vertical and lateral offsets to generate a continuous three-dimensional point cloud model of the rail with a unified spatial reference. The line laser scanning results are used as supplementary sensor data, and together with the displacement sensor data in the one-chord N-point chord measurement method, the elevation irregularity data and the track alignment irregularity data are calculated.

2. The method for measuring the three-dimensional profile of a rail by multi-source data fusion according to claim 1, characterized in that: The global coordinate transformation includes: transforming the coordinates of each contour point in the original rail cross-section contour data according to the calculated vertical and lateral offsets to obtain the three-dimensional coordinates of the point in the global coordinate system; and stitching together the three-dimensional point sets corresponding to all mileage positions in mileage order to form a continuous rail three-dimensional point cloud model.

3. A rail three-dimensional profile measurement device based on multi-source data fusion, according to the rail three-dimensional profile measurement method based on multi-source data fusion as described in claim 1 or 2, characterized in that, include: Line laser scanning module is used to acquire rail cross-sectional profile data at multiple discrete mileage locations; The irregularity measurement module is used to obtain rail elevation irregularity data and rail alignment irregularity data at each mileage location using the one-chord N-point chord measurement method. The mileage positioning module is used to record mileage information at each measurement location; The data fusion processing module, connected to the line laser scanning module, the irregularity measurement module, and the mileage positioning module, is used to perform spatiotemporal alignment of the rail cross-section contour data with the elevation irregularity data and the track alignment irregularity data, and to calculate the global offset of each cross-section based on the vertical reference point and the lateral reference point, thereby realizing the reconstruction of the three-dimensional profile of the rail.

4. The rail three-dimensional profile measurement device based on multi-source data fusion according to claim 3, characterized in that: The vertical reference point is the highest local point in the rail profile, and the lateral reference point is the abscissa of the sampling point at the working edge feature position in the rail profile.

5. The rail three-dimensional profile measurement device based on multi-source data fusion according to claim 3, characterized in that: The data fusion processing module constructs a profile calibration model with smoothing constraints. The profile calibration model includes data fidelity terms and smoothing regularization terms, which are used to optimize the solution of the vertical and lateral offsets of each section.

6. A multi-source data fusion system for measuring the three-dimensional profile of rails, characterized in that, Includes the rail three-dimensional profile measuring device as described in any one of claims 3 to 5, and: A transport platform is used to carry the measuring device and move it along the track; The data storage module, connected to the data fusion processing module, is used to store the reconstructed three-dimensional point cloud model of the rail. The display terminal is connected to the data storage module and is used to visualize the three-dimensional profile of the rail.

7. The rail three-dimensional profile measurement system based on multi-source data fusion according to claim 6, characterized in that, The scanning data from the line laser scanning module is fused with the displacement sensor data from the irregularity measurement module to jointly calculate the track elevation irregularity and track alignment irregularity data, thereby improving the accuracy of irregularity measurement.

Images