A rail profile registration method, device, equipment and storage medium

Abstract

The application discloses a rail profile registration method, device, equipment and storage medium, and relates to the technical field of track dynamic measurement. The method comprises the following steps: acquiring a standard rail profile; acquiring a dynamic measurement result output by a laser sensor, and obtaining a measured rail profile through preprocessing based on the dynamic measurement result; performing coarse registration of a rail waist profile based on a rail waist area of the standard rail profile and a rail waist area of the measured rail profile by using a geometric constraint and a feature point registration mode to obtain a coarse registration result; and performing fine registration based on the coarse registration result according to a preset rotation angle correction value range and a translation amount correction value range to obtain a final rail profile matching result. The method realizes fast and accurate registration of the measured rail profile and the standard rail profile, and greatly improves the efficiency of rail profile registration in track dynamic detection.

Description

Technical Field

[0001] This invention relates to the field of track dynamic measurement technology, and in particular to a method, apparatus, equipment and storage medium for rail profile registration. Background Technology

[0002] Track is a crucial infrastructure supporting train operation, and efficient track condition detection is essential for ensuring safe railway operation. Currently, non-contact detection technology based on laser displacement sensors is widely used in track inspection systems, enabling continuous dynamic measurement of the track. By comparing track profile data with standard profile data, some characteristic parameters of the actual track are obtained, and geometric irregularities of the actual track can be obtained through numerical calculations. Therefore, how to achieve high-precision automatic matching between the measured track profile and the standard track profile is a key issue in the high-precision dynamic detection of track geometric parameters.

[0003] In existing technologies, methods based on direct matching of datasets use the ICP (Iterative Closest Point) algorithm to calculate the optimal rotation and translation method. However, this method involves a large amount of computation, slow processing speed, and is greatly affected by noise and oil stains generated during measurement, easily leading to local optima. It is not easy to implement in measurements with high real-time requirements. Existing technologies also use static templates for feature point matching, and the accuracy of the localization depends on the accuracy of the fitted curve, which is easily affected by outliers and measurement error points. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, apparatus, equipment, and medium for registering rail profiles, which can achieve rapid and accurate registration between measured rail profiles and standard rail profiles, greatly improving the efficiency of rail profile registration in dynamic track inspection. The specific solution is as follows:

[0005] Firstly, this application discloses a method for registering rail profiles, including:

[0006] Obtain standard rail profiles;

[0007] The dynamic measurement results output by the laser sensor are obtained, and the measured rail profile is obtained through preprocessing based on the dynamic measurement results;

[0008] Based on the web region of the standard rail profile and the web region of the measured rail profile, coarse registration of the web profile is performed using geometric constraints and feature point registration methods to obtain the coarse registration result.

[0009] Based on the coarse registration result, fine registration is performed according to the preset rotation angle correction value range and translation amount correction value range to obtain the final rail profile matching result.

[0010] Optionally, the step of acquiring the dynamic measurement results output by the laser sensor, and obtaining the measured rail profile based on the dynamic measurement results through preprocessing, includes:

[0011] Based on the dynamic measurement results, an initial measured rail profile is obtained. The initial measured rail profile is then deduplicated to obtain the deduplicated measured rail profile.

[0012] Based on the relationship between the distance between two adjacent points in the measured rail profile after deduplication and the preset distance threshold, the starting point and ending point of the profile fracture are determined, and the target fracture interval is determined based on the density of the starting point and ending point.

[0013] The rail head region and the initial rail web region are obtained by segmenting the target fracture region. The rail web region is then extracted from the initial rail web region to obtain the measured rail profile.

[0014] Optionally, extracting the rail waist region from the initial rail waist region includes:

[0015] The target straight line is obtained by fitting the profile with a negative slope in the railhead area;

[0016] Based on the division of the initial track waist region according to the target straight line, the track waist region and the noise region are obtained.

[0017] Optionally, obtaining the standard rail profile includes:

[0018] Based on the curvature of the standard rail profile, the standard rail profile is divided into a large curvature range and a small curvature range;

[0019] Discrete points are extracted from the profile of the large curvature interval using a small step size, and discrete points are extracted from the profile of the small curvature interval using a large step size.

[0020] Optionally, before performing coarse registration of the rail web region based on the standard rail profile and the measured rail profile using geometric constraints and feature point registration, the method further includes:

[0021] The rail web region of the standard rail profile is divided into multiple target region profiles;

[0022] A coordinate system is established with the center of the rail base as the origin, and the function expression corresponding to the profile of each target area is determined based on the coordinate system.

[0023] Optionally, before performing coarse registration of the track waist profile using geometric constraints and feature point registration, the method further includes:

[0024] Establish endpoint lines based on the two endpoints of the web area of ​​the measured rail profile, and select the farthest point based on the distance from each point on the web area to the endpoint lines.

[0025] The target distance is obtained by multiplying the farthest distance from the farthest point to the endpoint by a scaling factor.

[0026] By translating the line connecting the endpoints, a target line is obtained whose perpendicular distance to the farthest point is the target distance;

[0027] The two endpoints where the target line intersects with the rail waist area, as well as the farthest point, are used as the registration feature points.

[0028] Optionally, the coarse registration of the track waist profile using geometric constraints and feature point registration methods to obtain the coarse registration result includes:

[0029] Based on the geometric features of the rail web, a registration feature point closer to the rail bottom is selected from the registration feature points as the first registration feature point, and the other is selected as the second registration feature point.

[0030] Align and coincide the first registration feature point with the discrete points of the rail web area of ​​the measured rail profile;

[0031] After overlapping, a first circle equation is constructed with the first registration feature point as the center and the straight-line distance from the first registration feature point to the second registration feature point as the radius. The intersection of the first circle equation and the function expression corresponding to the target region profile is taken as the third target point.

[0032] A second circle equation is constructed with the first registration feature point as the center and the straight-line distance from the first registration feature point to the farthest point as the radius. The intersection of the second circle equation and the function expression corresponding to the target region profile is taken as the fourth target point.

[0033] The first registration feature point, the second registration feature point, and the farthest point are used as the registration feature points of the measured rail profile, and the first registration feature point, the third target point, and the fourth target point are used as the registration feature points of the standard rail profile.

[0034] Determine the actual vertical distance from the fourth target point to the straight line formed by the first registration feature point and the third target point;

[0035] Compare the difference between the target distance and the actual vertical distance;

[0036] If the difference is less than a preset threshold, then the optimal rotation matrix and translation matrix from the registration feature points of the measured rail profile to the registration feature points of the standard rail profile are determined, and the registration error after rotation and translation is calculated.

[0037] By aligning and overlapping the first registration feature point with each discrete point in the web area of ​​the measured rail profile, a registration error set is obtained.

[0038] The minimum error is selected from the set of registration errors, and the measured rail profile is transformed into a coordinate system based on the position of each registration feature point corresponding to the minimum error, as well as the rotation and translation matrices, to obtain the coarse registration result.

[0039] Secondly, this application discloses a rail profile registration device, comprising:

[0040] The standard rail profile acquisition module is used to acquire the standard rail profile.

[0041] The measured rail profile acquisition module is used to acquire the dynamic measurement results output by the laser sensor, and obtain the measured rail profile based on the dynamic measurement results through preprocessing.

[0042] The coarse registration module is used to perform coarse registration of the rail web region based on the standard rail profile and the measured rail profile, using geometric constraints and feature point registration methods, to obtain the coarse registration result.

[0043] The fine matching module is used to perform fine registration based on the coarse registration result and according to the preset range of rotation angle correction value and translation amount correction value to obtain the final rail profile matching result.

[0044] Thirdly, this application discloses an electronic device, including:

[0045] Memory, used to store computer programs;

[0046] A processor is used to execute the computer program to implement the aforementioned rail profile registration method.

[0047] Fourthly, this application discloses a computer-readable storage medium for storing a computer program; wherein the computer program, when executed by a processor, implements the aforementioned rail profile registration method.

[0048] In this application, a standard rail profile is obtained; dynamic measurement results output by a laser sensor are acquired, and the measured rail profile is obtained through preprocessing based on the dynamic measurement results; based on the rail web area of ​​the standard rail profile and the rail web area of ​​the measured rail profile, coarse registration of the rail web profile is performed using geometric constraints and feature point registration methods to obtain a coarse registration result; based on the coarse registration result, fine registration is performed according to a preset range of rotation angle correction values ​​and translation correction values ​​to obtain the final rail profile matching result. It can be seen that by selecting the rail web portion that does not directly contact the wheel as the registration area, using geometric constraints to perform coarse registration of the measured rail profile and the standard rail profile globally, and then using the coarse registration result as a basis, fine-tuning is performed within a preset range of rotation angle correction values ​​and translation correction values ​​to obtain the optimal rigid body transformation relationship, thereby achieving rapid and accurate registration between the measured rail profile and the standard rail profile, greatly improving the efficiency of rail profile registration in track dynamic detection. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present 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 only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0050] Figure 1 A flowchart of a rail profile registration method provided in this application;

[0051] Figure 2 This application provides a schematic diagram of a specific standard rail cross-sectional profile.

[0052] Figure 3 This application provides a schematic diagram of the discrete points of a specific standard rail profile.

[0053] Figure 4 This application provides a specific measured schematic diagram of the rail profile.

[0054] Figure 5 This application provides a specific schematic diagram of noise removal from a measured rail profile.

[0055] Figure 6 This application provides a specific measured schematic diagram of the clean rail web and rail head of the rail profile.

[0056] Figure 7 This application provides a schematic diagram illustrating the process of determining the registration feature points for a specific measured rail profile.

[0057] Figure 8A specific coarse registration diagram is provided for this application;

[0058] Figure 9 A schematic diagram of a specific fine registration result provided in this application;

[0059] Figure 10 A flowchart of a specific rail profile registration method provided in this application;

[0060] Figure 11 A schematic diagram of a rail profile registration device provided in this application;

[0061] Figure 12 This application provides a structural diagram of an electronic device. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] In existing technologies, methods based on direct matching of datasets utilize the ICP algorithm to calculate the optimal rotation and translation method. However, this method involves a large computational load, slow processing speed, and is significantly affected by noise and oil stains generated during measurement, easily leading to local optima, making it difficult to implement in measurements with high real-time requirements. Existing technologies also rely on static templates for feature point matching, where the positioning accuracy depends on the accuracy of the fitted curve and is susceptible to outliers and measurement error points. To overcome these technical problems, this application proposes a rail profile registration method that enables rapid and accurate registration of measured rail profiles with standard rail profiles, greatly improving the efficiency of rail profile registration in dynamic track inspection.

[0064] This application discloses a method for registering rail profiles. See also... Figure 1 As shown, the method may include the following steps:

[0065] Step S11: Obtain the standard rail profile.

[0066] In this embodiment, the standard rail profile is obtained, that is, the outline shape of the standard rail, for example... Figure 2 The cross-sectional profile of a standard 60kg / m rail is shown.

[0067] In this embodiment, before performing coarse registration of the rail web region based on the standard rail profile and the measured rail profile using geometric constraints and feature point registration, the process may further include: segmenting the rail web region of the standard rail profile to obtain multiple target region profiles; establishing a coordinate system with the center of the rail base as the origin, and determining the function expression corresponding to each target region profile based on the coordinate system. That is, establishing a function for the rail web profile. Since the rail web profile is irregular, a corresponding function expression is established for each region profile after region segmentation. For example... Figure 2 As shown, a track reference coordinate system is defined within the rail cross-section plane. It has the center of the rail base as the origin, the Y-axis perpendicular to the track direction, and the X-axis vertical. Segment AB is a circular arc with a radius of 400mm, segment BC is a circular arc with a radius of 20mm, segments CD and EF are line segments with slopes of 1:3 and 1:9 respectively, and segment DE is a circular arc with a radius of 40mm connecting segments CD and EF. The standard profile is divided into the above five intervals, and the function expression for each curve segment is obtained based on the coordinate information in the design drawing, facilitating the subsequent determination of the intersection points, i.e., the determination of the third and fourth target points. Each of the above intervals satisfies the following expression:

[0068]

[0069] Among them, k1, D2, E2, k3, D4, E4, D5, and E5 are coefficients, and b1, F2, b3, F4, and F5 are constants.

[0070] In this embodiment, obtaining the standard rail profile may include: dividing the standard rail profile into a large curvature range and a small curvature range based on its curvature; extracting discrete points from the profile in the large curvature range using a small step size, and extracting discrete points from the profile in the small curvature range using a large step size. That is, extracting discrete points from the standard profile at a certain step size, capturing as many geometric features as possible, and using these extracted discrete points for subsequent feature point approval. Specifically, for example... Figure 3 As shown, intervals with greater curvature are represented by denser points, while intervals with less curvature are represented by sparser points. The discrete point set is denoted as X = {x}. i ,i=0,1,2……n}.

[0071] Step S12: Obtain the dynamic measurement results output by the laser sensor, and obtain the measured rail profile based on the dynamic measurement results through preprocessing.

[0072] This step aims to obtain the dynamic measurement results output in real time by the laser sensor, and to obtain the measured rail profile through preprocessing, which may include data cleaning and segmentation.

[0073] The above-mentioned acquisition of dynamic measurement results from the laser sensor output, and the obtaining of the measured rail profile through preprocessing based on the dynamic measurement results, may include: obtaining an initial measured rail profile based on the dynamic measurement results; deduplicating the initial measured rail profile to obtain a deduplicated measured rail profile; determining the start and end points of the profile fracture based on the relationship between the distance between two adjacent points in the deduplicated measured rail profile and a preset distance threshold; determining the target fracture interval based on the density of the start and end points; dividing the target fracture interval to obtain the rail head area and the initial rail web area; and extracting the rail web area from the initial rail web area to obtain the measured rail profile.

[0074] It is understandable that measured profile data is affected by the testing environment, resulting in measurement outliers. Therefore, it is necessary to perform deduplication, zero-point removal, and outlier elimination on the measured profile data to obtain the actual measured profile. After deduplication, the distance between two adjacent points is calculated and compared with a set distance threshold to find the break start and end points. However, in actual operation, there is often more than one break interval, requiring further processing. Only the two break intervals with the most points need to be retained, thus dividing the measured data into the part below the break interval and the part above the break interval. For example... Figure 4 As shown, it is divided into the rail head area and the initial rail web area. At this time, the initial rail web area contains noise.

[0075] In this embodiment, extracting the rail waist region from the initial rail waist region may include: fitting a target straight line based on the negative slope profile of the rail head region; and dividing the initial rail waist region according to the target straight line to obtain the rail waist region and the noise region. Even after deduplication, the data still contains some noise points outside the rail waist and rail head regions, mainly noise points and sleeper points present in the initial rail waist region, which need to be removed. Specifically, this is done by extracting a set of points with a negative slope from the rail head, i.e., the straight line portion of the rail head, and fitting this segment of data using the least squares method to obtain the optimal fitted straight line L1, for example... Figure 5 As shown, the distance from the rail web data points to L1 is calculated, distinguishing between points above and below L1. The points below L1 are the noise points that are removed, ultimately yielding clean rail web and rail head data. Figure 6 As shown.

[0076] Step S13: Based on the web area of ​​the standard rail profile and the web area of ​​the measured rail profile, perform coarse registration of the web profile using geometric constraints and feature point registration methods to obtain the coarse registration result.

[0077] After obtaining clean rail web data of the measured rail profile, the rail web area of ​​the standard rail profile and the rail web area of ​​the measured rail profile are matched. It can be understood that since the rail head part is in direct contact with the wheel, it is prone to surface defects and oil stains, and is not suitable as a registration area. Therefore, the rail web part that is not in direct contact with the wheel is selected as the registration area. Specifically, the rail web profile can be coarsely registered by using feature point registration through geometric constraints. That is, the initial coarse matching is performed by using feature points on the rail web area of ​​the standard rail profile and the rail web area of ​​the measured rail profile.

[0078] In this embodiment, before performing coarse registration of the rail waist profile using geometric constraints and feature point registration, the method may further include: establishing endpoint lines connecting the two endpoints of the rail waist region of the measured rail profile; selecting the farthest point based on the distance from each point on the rail waist region to the endpoint lines; obtaining the target distance by multiplying the farthest distance from the farthest point to the endpoint lines by a scaling factor; obtaining a target line whose perpendicular distance to the farthest point is the target distance by translating the endpoint lines; and using the two endpoints where the target line intersects with the rail waist region and the farthest point as the registration feature points.

[0079] First, registration feature points are selected in the rail web area of ​​the measured rail profile. Essentially, the matching algorithm calculates the coordinates of the areas or feature points without surface defects in the measured rail profile with their corresponding parts in the standard rail profile to obtain the overall rotation and translation parameters. The measured rail profile parts or feature points involved in the parameter calculation are called registration areas or registration feature points. The rail web data of the measured rail profile obtained through the above processing is prone to noise and measurement errors at breakpoints. To avoid affecting the matching accuracy, such as... Figure 7 As shown, feature points can be filtered as follows:

[0080] Let the endpoints of the rail web data be P(x1,y1) and Q(x2,y2), respectively. Calculate the perpendicular distance from each point on the rail web to the line connecting the endpoints (the line connecting P and Q). Find the point with the largest distance from the curve secant and denote it as the farthest point M1(x3,y3). Define the farthest distance as:

[0081]

[0082] Find the secant line whose distance to the farthest point is equal to the product of the farthest distance and the scaling factor. For example, find the secant line whose distance to the farthest point is equal to 70%max_d. The two endpoints where the secant line intersects with the rail waist area are P1(x4,y4) and Q1(x5,y5), which are the registration feature points.

[0083] In this embodiment, the coarse registration of the track waist profile using geometric constraints and feature point registration methods to obtain the coarse registration result may include:

[0084] 1. Based on the geometric features of the rail web, select the registration feature point closest to the rail bottom from the registration feature points as the first registration feature point, and the other as the second registration feature point; the obtained registration feature points are the first registration feature point P1(x4,y4), the second registration feature point Q1(x5,y5), and M1(x3,y3). Combining with the prior geometric information, it can be determined that point P1 is a point on the rail web closer to the middle section of the rail web, and point Q1 is a point on the rail web closer to the rail bottom.

[0085] 2. Align and coincide the first registration feature point with the discrete points of the rail web region of the measured rail profile; if point Q1 is aligned and coincident with any discrete point in the discrete point set of the standard rail profile, such as Figure 8 As shown.

[0086] 3. After overlap, construct a first circle equation with the first registration feature point as the center and the straight-line distance from the first registration feature point to the second registration feature point as the radius. The intersection of the first circle equation and the function expression corresponding to the target region profile is taken as the third target point; for example... Figure 8 As shown, given the spatial relationship and geometric distance constraints between points P1 and Q1, a possible point, denoted as P2, is searched on the standard profile curve. Specifically, the equation of the first circle is obtained with P1 and Q1 as the radius, and the Euclidean distance between points P1 and Q1 is... The intersection of the first circle equation and the standard rail profile is the third target point P2. That is, the coordinates of the intersection point are obtained by solving the equations of the first circle equation and the corresponding function expression equations of the target area profile.

[0087] 4. Construct a second circle equation with the first registration feature point as the center and the straight-line distance from the first registration feature point to the farthest point as the radius. The intersection of the second circle equation and the function expression corresponding to the target region profile is taken as the fourth target point; similarly, as... Figure 8 As shown, the possible point on the standard rail profile that satisfies the geometric distance Q1M1 is denoted as the fourth target point M2. The Euclidean distance between points Q1 and M1 is...

[0088] 5. The first registration feature point, the second registration feature point, and the farthest point are used as the registration feature points of the measured rail profile, and the first registration feature point, the third target point, and the fourth target point are used as the registration feature points of the standard rail profile; that is, Q1, M1, and P1 are the registration feature points of the measured rail profile, and Q1, M2, and P3 are the registration feature points of the standard rail profile.

[0089] 6. Determine the actual vertical distance from the fourth target point to the straight line formed by the first registration feature point and the third target point; that is, calculate the distance from point M2 to the secant line Q1P2, and denot it as the actual vertical distance d1.

[0090] 7. Compare the difference between the target distance and the actual vertical distance; that is, calculate the distance difference between d1 and 70%max_d, which is to compare the distance from point M2 to secant Q1P2 and the distance from point M1 to secant Q1P1.

[0091] 8. If the difference is less than a preset threshold, the optimal rotation and translation matrix from the registration feature point of the measured rail profile to the registration feature point of the standard rail profile is determined, and the registration error after rotation and translation is calculated. The preset threshold can be 2mm. It can be understood that when the distance difference is greater than the set threshold, it can be considered that the discrete point selected at this time that coincides with Q1 is definitely not the point in the standard rail profile corresponding to Q1. Therefore, no error calculation is performed, which can avoid a large number of error calculations.

[0092] If the difference is less than or equal to a preset threshold, the rotation matrix R1 and translation matrix T1 are calculated using the registered points Q1 and P2 as references. It can be understood that for the same rail profile, the center of the rail web light strip is extracted, and a set of registered points U = {u} in the measurement coordinate system is calculated using a visual mathematical calculation model. i Let i = 0, 1, 2, ..., n} be the set of registration points X = {x_i, i = 0, 1, 2, ..., n} in the standard coordinate system. i , i = 0, 1, 2, ..., n}, for each point u i =(u i1 u i2 ,......u in ) and x i =(x i1 x i2 , ...x in According to the spatial coordinate transformation relationship, we have:

[0093] X = R·U + T;

[0094] R is the rotation matrix, and T is the translation matrix. To minimize the registration error, the minimum error satisfies the following formula:

[0095]

[0096] R is a rotation matrix that satisfies T is a translation matrix that satisfies

[0097] 9. By aligning and overlapping the first registration feature point with each discrete point of the rail web region of the measured rail profile, a registration error set is obtained; by aligning and overlapping Q1 with each discrete point of the standard rail profile, the error corresponding to each discrete point whose difference satisfies the condition is obtained, that is, the nth registration criterion assumes that point Q1 is aligned with the corresponding nth rail web discrete point, and then repeats the above subsequent operations to obtain the registration error set error={errori,i=1,2,……n}.

[0098] 10. Select the minimum error from the set of registration errors, and based on the positions of the registration feature points corresponding to the minimum error, as well as the rotation and translation matrices, perform coordinate system transformation on the measured rail profile to obtain the coarse registration result. Find the registration method with the minimum error; the rotation angle is denoted as θ, and the translation is denoted as t. x and t y By substituting the transformation parameters into X = R·U + T, the measured rail profile is transformed into the plane coordinate system of the standard rail profile, thus completing the coarse matching of the profile.

[0099] Step S14: Based on the coarse registration result, perform fine registration according to the preset rotation angle correction value range and translation amount correction value range to obtain the final rail profile matching result.

[0100] To eliminate the spatial positional error between the selected standard rail profile discrete points and the measured registration points, the profile is adjusted based on the coarse matching reference. Specifically, fine adjustments are made within a given range of rotation angle correction values ​​Δθ and translation correction values ​​Δt to find the optimal combination of correction values ​​that minimizes the registration error, thus completing the fine registration. The final matching result is as follows: Figure 9 As shown.

[0101] As can be seen, this embodiment proposes an automatic rail profile matching method based on laser sensors, for example... Figure 10As shown, using the rail web segment data of the standard profile as the matching benchmark, geometric constraints are used to globally search for possible alignment points. The point with the smallest alignment error is selected to perform coarse registration between the measured rail profile and the standard rail profile. Based on the coarse registration result, rotation and translation thresholds are set, and through iterative calculations, the optimal rigid body transformation relationship is obtained, thereby achieving fast and accurate registration between the measured profile and the standard profile. Fully utilizing prior knowledge for profile registration improves the data utilization rate of previous static measurements and experiments, avoiding the waste of historical data. The use of simple and easy-to-implement geometric constraints accelerates the search speed while ensuring registration accuracy, exhibiting high real-time performance. Furthermore, in calculating the registration error, not all registration cases are included in the calculation; geometric constraints are further considered to exclude points that cannot be the optimal registration method, greatly reducing the computational load, accelerating the registration time, and to some extent avoiding data matching getting trapped in local optima.

[0102] As can be seen from the above, this embodiment acquires the standard rail profile; acquires the dynamic measurement results output by the laser sensor, and obtains the measured rail profile through preprocessing based on the dynamic measurement results; based on the rail web area of ​​the standard rail profile and the rail web area of ​​the measured rail profile, coarse registration of the rail web profile is performed using geometric constraints and feature point registration methods to obtain the coarse registration result; based on the coarse registration result, fine registration is performed according to the preset rotation angle correction value range and translation correction value range to obtain the final rail profile matching result. It can be seen that by selecting the rail web portion that does not directly contact the wheel as the registration area, the accuracy of the registration is improved; geometric relationship constraints are used globally to perform coarse registration of the measured rail profile and the standard rail profile; and then, based on the coarse registration result, fine adjustments are made within the preset rotation angle correction value range and translation correction value range to obtain the optimal rigid body transformation relationship. This achieves rapid and accurate registration of the measured rail profile and the standard rail profile, greatly improving the efficiency of rail profile registration in track dynamic detection.

[0103] Accordingly, this application also discloses a rail profile registration device, see [link to relevant documentation]. Figure 11 As shown, the device includes:

[0104] Standard rail profile acquisition module 11 is used to acquire standard rail profiles;

[0105] The measured rail profile acquisition module 12 is used to acquire the dynamic measurement results output by the laser sensor, and obtain the measured rail profile based on the dynamic measurement results through preprocessing.

[0106] The coarse registration module 13 is used to perform coarse registration of the rail web region based on the standard rail profile and the rail web region of the measured rail profile, using geometric constraints and feature point registration methods, to obtain the coarse registration result.

[0107] The fine matching module 14 is used to perform fine matching based on the coarse registration result and according to the preset rotation angle correction value range and translation amount correction value range to obtain the final rail profile matching result.

[0108] As can be seen from the above, this embodiment acquires the standard rail profile; acquires the dynamic measurement results output by the laser sensor, and obtains the measured rail profile through preprocessing based on the dynamic measurement results; based on the rail web area of ​​the standard rail profile and the rail web area of ​​the measured rail profile, coarse registration of the rail web profile is performed using geometric constraints and feature point registration methods to obtain the coarse registration result; based on the coarse registration result, fine registration is performed according to the preset rotation angle correction value range and translation correction value range to obtain the final rail profile matching result. It can be seen that by selecting the rail web portion that does not directly contact the wheel as the registration area, the accuracy of the registration is improved; geometric relationship constraints are used globally to perform coarse registration of the measured rail profile and the standard rail profile; and then, based on the coarse registration result, fine adjustments are made within the preset rotation angle correction value range and translation correction value range to obtain the optimal rigid body transformation relationship. This achieves rapid and accurate registration of the measured rail profile and the standard rail profile, greatly improving the efficiency of rail profile registration in track dynamic detection.

[0109] In some specific embodiments, the measured rail profile acquisition module 12 may specifically include:

[0110] The deduplication unit is used to obtain an initial measured rail profile based on the dynamic measurement results, and to deduplicate the initial measured rail profile to obtain a deduplicated measured rail profile.

[0111] The target fracture interval determination unit is used to determine the starting point and ending point of the profile fracture based on the relationship between the distance between two adjacent points in the measured rail profile after deduplication and a preset distance threshold, and to determine the target fracture interval based on the density of the starting point and the ending point.

[0112] The rail web area determination unit is used to divide the target fracture section to obtain the rail head area and the initial rail web area, and extract the rail web area from the initial rail web area to obtain the measured rail profile.

[0113] In some specific embodiments, the rail waist area determination unit may specifically include:

[0114] The line fitting unit is used to fit the target line based on the profile with a negative slope in the railhead area;

[0115] The partitioning unit is used to divide the initial track waist region according to the target straight line, thereby obtaining the track waist region and the noise region.

[0116] In some specific embodiments, the standard rail profile acquisition module 11 may specifically include:

[0117] Curvature division unit, used to divide the standard rail profile into large curvature range and small curvature range according to the curvature of the standard rail profile;

[0118] The discrete point extraction unit is used to extract discrete points from the profile of the large curvature interval using a small step size, and to extract discrete points from the profile of the small curvature interval using a large step size.

[0119] In some specific embodiments, the rail profile matching device may specifically include:

[0120] The rail web segmentation unit is used to segment the rail web area of ​​the standard rail profile to obtain multiple target area profiles;

[0121] The function expression generation unit is used to establish a coordinate system with the center of the rail base as the origin, and to determine the function expression corresponding to the profile of each target area based on the coordinate system.

[0122] In some specific embodiments, the rail profile matching device may specifically include:

[0123] The farthest point determination unit is used to establish an endpoint connection line based on the two endpoints of the rail web area of ​​the measured rail profile, and to filter out the farthest point based on the distance of each point on the rail web area to the endpoint connection line.

[0124] The target distance determination unit is used to obtain the target distance by multiplying the farthest distance of the line connecting the farthest point to the endpoint by a scaling factor;

[0125] The target line determination unit is used to obtain a target line whose perpendicular distance to the farthest point is the target distance by translating the endpoint line;

[0126] The registration point determination unit is used to take the two endpoints where the target line intersects with the rail waist area and the farthest point as the registration feature points.

[0127] In some specific embodiments, the coarse registration module 13 may specifically include:

[0128] The first registration feature point determination unit is used to select the registration feature point closest to the rail bottom from the registration feature points based on the geometric features of the rail web as the first registration feature point, and the other as the second registration feature point.

[0129] Alignment unit, used to align and coincide the first registration feature point with the discrete points of the rail web area of ​​the measured rail profile;

[0130] The third target point determination unit is used to construct a first circle equation after overlap, with the first registration feature point as the center and the straight-line distance from the first registration feature point to the second registration feature point as the radius, and the intersection of the first circle equation and the function expression corresponding to the target region profile is taken as the third target point.

[0131] The fourth target point determination unit is used to construct a second circle equation with the first registration feature point as the center and the straight-line distance from the first registration feature point to the farthest point as the radius, and to take the intersection of the second circle equation and the function expression corresponding to the target region profile as the fourth target point.

[0132] The registration feature point allocation unit is used to use the first registration feature point, the second registration feature point and the farthest point as the registration feature points of the measured rail profile, and to use the first registration feature point, the third target point and the fourth target point as the registration feature points of the standard rail profile.

[0133] The actual vertical distance determination unit is used to determine the actual vertical distance from the fourth target point to the straight line formed by the first registration feature point and the third target point;

[0134] A comparison unit is used to compare the difference between the target distance and the actual vertical distance;

[0135] The matrix determination unit is used to determine the optimal rotation matrix and translation matrix from the registration feature points of the measured rail profile to the registration feature points of the standard rail profile if the difference is less than a preset threshold, and to calculate the registration error after rotation and translation.

[0136] The registration error set determination unit is used to obtain the registration error set by aligning and overlapping the first registration feature point with each discrete point of the rail web area of ​​the measured rail profile.

[0137] The coordinate transformation unit is used to select the minimum error from the set of registration errors, and transform the measured rail profile into a coordinate system according to the position of each registration feature point corresponding to the minimum error and the rotation and translation matrices to obtain the coarse registration result.

[0138] Furthermore, this application also discloses an electronic device, see [link to relevant documentation]. Figure 12 As shown, the content in the figure should not be considered as any limitation on the scope of use of this application.

[0139] Figure 12This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the rail profile registration method disclosed in any of the foregoing embodiments.

[0140] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0141] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon include operating system 221, computer program 222 and data 223 including standard rail profile, etc. The storage method can be temporary storage or permanent storage.

[0142] The operating system 221 manages and controls the various hardware devices on the electronic device 20 and the computer program 222 to enable the processor 21 to perform calculations and processing on the massive data 223 in the memory 22. The operating system 221 can be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the rail profile registration method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0143] Furthermore, this application also discloses a computer storage medium storing computer-executable instructions. When the computer-executable instructions are loaded and executed by a processor, they implement the rail profile registration method steps disclosed in any of the foregoing embodiments.

[0144] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0145] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0146] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0147] The foregoing has provided a detailed description of the rail profile registration method, apparatus, equipment, and medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for matching rail profiles, characterized in that, include: Obtain standard rail profiles; The dynamic measurement results output by the laser sensor are obtained, and the measured rail profile is obtained through preprocessing based on the dynamic measurement results; Based on the web region of the standard rail profile and the web region of the measured rail profile, coarse registration of the web profile is performed using geometric constraints and feature point registration methods to obtain the coarse registration result. Based on the coarse registration result, fine registration is performed according to the preset rotation angle correction value range and translation correction value range to obtain the final rail profile matching result; The step of acquiring the dynamic measurement results output by the laser sensor, and obtaining the measured rail profile through preprocessing based on the dynamic measurement results, includes: Based on the dynamic measurement results, an initial measured rail profile is obtained. The initial measured rail profile is then deduplicated to obtain the deduplicated measured rail profile. Based on the relationship between the distance between two adjacent points in the measured rail profile after deduplication and the preset distance threshold, the starting point and ending point of the profile fracture are determined, and the target fracture interval is determined based on the density of the starting point and ending point. The rail head region and the initial rail web region are obtained by segmenting the target fracture region. The rail web region is then extracted from the initial rail web region to obtain the measured rail profile.

2. The rail profile registration method according to claim 1, characterized in that, Extracting the rail waist region from the initial rail waist region includes: The target straight line is obtained by fitting the profile with a negative slope in the railhead area; Based on the division of the initial track waist region according to the target straight line, the track waist region and the noise region are obtained.

3. The rail profile registration method according to claim 1, characterized in that, The process of obtaining the standard rail profile includes: Based on the curvature of the standard rail profile, the standard rail profile is divided into a large curvature range and a small curvature range; Discrete points are extracted from the profile of the large curvature interval using a small step size, and discrete points are extracted from the profile of the small curvature interval using a large step size.

4. The rail profile registration method according to claim 3, characterized in that, Before performing coarse registration of the rail web region based on the standard rail profile and the measured rail profile using geometric constraints and feature point registration methods, the process also includes: The rail web region of the standard rail profile is divided into multiple target region profiles; A coordinate system is established with the center of the rail base as the origin, and the function expression corresponding to the profile of each target area is determined based on the coordinate system.

5. The rail profile registration method according to claim 4, characterized in that, Before performing coarse registration of the track waist profile using geometric constraints and feature point registration methods, the following steps are also included: Establish endpoint lines based on the two endpoints of the web area of ​​the measured rail profile, and select the farthest point based on the distance from each point on the web area to the endpoint lines. The target distance is obtained by multiplying the farthest distance from the farthest point to the endpoint by a scaling factor. By translating the line connecting the endpoints, a target line is obtained whose perpendicular distance to the farthest point is the target distance; The two endpoints where the target line intersects with the rail waist area, as well as the farthest point, are used as the registration feature points.

6. The rail profile registration method according to claim 5, characterized in that, The method of using geometric constraints and feature point registration to perform coarse registration of the track waist profile yields the following coarse registration results: Based on the geometric features of the rail web, a registration feature point closer to the rail bottom is selected from the registration feature points as the first registration feature point, and the other is selected as the second registration feature point. Align and coincide the first registration feature point with the discrete points of the rail web area of ​​the measured rail profile; After overlapping, a first circle equation is constructed with the first registration feature point as the center and the straight-line distance from the first registration feature point to the second registration feature point as the radius. The intersection of the first circle equation and the function expression corresponding to the target region profile is taken as the third target point. A second circle equation is constructed with the first registration feature point as the center and the straight-line distance from the first registration feature point to the farthest point as the radius. The intersection of the second circle equation and the function expression corresponding to the target region profile is taken as the fourth target point. The first registration feature point, the second registration feature point, and the farthest point are used as the registration feature points of the measured rail profile, and the first registration feature point, the third target point, and the fourth target point are used as the registration feature points of the standard rail profile. Determine the actual vertical distance from the fourth target point to the straight line formed by the first registration feature point and the third target point; Compare the difference between the target distance and the actual vertical distance; If the difference is less than a preset threshold, then the optimal rotation matrix and translation matrix from the registration feature points of the measured rail profile to the registration feature points of the standard rail profile are determined, and the registration error after rotation and translation is calculated. By aligning and overlapping the first registration feature point with each discrete point in the web area of ​​the measured rail profile, a registration error set is obtained. The minimum error is selected from the set of registration errors, and the measured rail profile is transformed into a coordinate system based on the position of each registration feature point corresponding to the minimum error, as well as the rotation and translation matrices, to obtain the coarse registration result.

7. A rail profile registration device, characterized in that, include: The standard rail profile acquisition module is used to acquire the standard rail profile. The measured rail profile acquisition module is used to acquire the dynamic measurement results output by the laser sensor, and obtain the measured rail profile based on the dynamic measurement results through preprocessing. The coarse registration module is used to perform coarse registration of the rail web region based on the standard rail profile and the measured rail profile, using geometric constraints and feature point registration methods, to obtain the coarse registration result. The fine matching module is used to perform fine registration based on the coarse registration result and according to the preset rotation angle correction value range and translation amount correction value range to obtain the final rail profile matching result. The measured rail profile acquisition module includes: The deduplication unit is used to obtain an initial measured rail profile based on the dynamic measurement results, and to deduplicate the initial measured rail profile to obtain a deduplicated measured rail profile. The target fracture interval determination unit is used to determine the starting point and ending point of the profile fracture based on the relationship between the distance between two adjacent points in the measured rail profile after deduplication and a preset distance threshold, and to determine the target fracture interval based on the density of the starting point and the ending point. The rail web area determination unit is used to divide the target fracture section to obtain the rail head area and the initial rail web area, and extract the rail web area from the initial rail web area to obtain the measured rail profile.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the rail profile registration method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, Used to store computer programs; wherein the computer programs, when executed by a processor, implement the rail profile registration method as described in any one of claims 1 to 6.

Images