Steel rail abrasion detection method and device and computer program

Abstract

The invention relates to a steel rail abrasion detection method and device and a computer program. The method comprises the following steps: acquiring actually measured point cloud data of an actual section contour of a steel rail; screening out a first arc and a second arc from the actually measured point cloud data, and respectively determining the coordinates of the center point of each arc; performing optimization processing on the two arcs according to the center point coordinates of the two arcs and the preset central angle of each arc so as to update the center point coordinates; determining a tangent point position between the two arcs according to the updated coordinates of the two circle center points; obtaining a first vector according to the new first circle center point coordinate and the tangent point position; according to a comparison result between the first vector and a preset second vector, determining a rotation angle and a translation vector, and generating a spatial transformation matrix; and rotating and translating the actually measured point cloud data according to the spatial transformation matrix to obtain a registration point cloud, and comparing the registration point cloud with a preset standard point cloud to determine the abrasion of the steel rail. The method has the advantages of small size, low cost, high calculation efficiency, high registration precision and the like.

Description

Technical Field

[0001] The present disclosure relates to the technical field of track maintenance, and in particular to a rail wear detection method, device and computer program. Background Art

[0002] Rail wear is widely present in my country's heavy-haul railways. Wear causes changes in the rail head profile, resulting in changes in the wheel-rail contact state and the dynamic response of the vehicle-track system, which has a great impact on the driving stability and safety of heavy-haul trains and the service life of various components of the vehicle-track system.

[0003] Some of the existing registration algorithms are based on the maximum curvature entropy of the dynamic window to perform small arc segmentation on the rail waist, and use the double center of the arc fitting at the bottom of the rail waist on both sides to obtain the reference point, so as to achieve contour alignment; others use the unworn triangular area formed by the arc curve at the bottom of the rail waist and the lower end point of the rail head to perform affine transformation, and use the ICP (Iterative Closest Point, ICP for short) algorithm to match the standard contour with the measured contour; others use the slope characteristics of the rail head measuring surface and the geometric relationship of the rail waist arc reference point to achieve two-segment arc segmentation and positioning, and then use the double center method to perform precise alignment with the standard contour.

[0004] The existing mainstream method is to first use the reference point to obtain the initial position, and then perform precise alignment according to the ICP algorithm. This method divides the rail waist area into large and small arc point sets when obtaining the reference point. However, since this method uses independent point set division methods in the process of obtaining the arc, it is difficult to form an overall arc system with two arcs, which can easily affect the process of obtaining the center of the circle and easily fall into their own local optimal solutions, thus affecting the initial position of the reference point.

[0005] In order to achieve rail profile registration in the existing registration methods for wear calculation, reference point-based and ICP-based methods can usually be used. Although the reference point-based method is fast, it may be interfered by some factors due to its low accuracy and poor robustness. Although the ICP algorithm can achieve higher accuracy, it is affected by data noise and initial values ​​and is prone to fall into local optimal solutions.

[0006] On this basis, some people have proposed a method that combines reference points with the ICP algorithm. In this method, the dual center method is generally used. This method divides the rail waist area into large and small arc point sets to achieve the two reference points. However, it is easy to fall into the local optimal solution in the process of obtaining it. This is because the division of the two arc point sets is relatively independent. Therefore, the reference points obtained by this method are prone to lead to their own local optimal solutions, thus affecting the subsequent precise alignment. At present, arc segmentation is generally based on the curvature segmentation method. This method requires second-order derivative operations when calculating derivatives, which has obvious fluctuations, and the computing power cost is high, making it difficult to ensure a suitable segmentation threshold.

[0007] In addition, most of the existing detection equipment adopts the method of single-sided dual sensors or triple sensors, which makes the overall device generally larger in size. Compared with single-sided sensors, the installation space and location requirements of sensors are more restricted for existing dynamic detection needs, reducing the installation adaptability of the equipment on different detection vehicles. The increase in the number of single-sided sensors will also increase the cost of equipment and subsequent maintenance costs; further, the multi-sensor solution will not only increase the cost of equipment and subsequent maintenance costs, but also require higher power consumption and higher computing power for equipment operation, which greatly restricts the actual application scenarios of sensors.

[0008] With the development of heavy-haul railway operation and maintenance, higher requirements are placed on rail wear detection, especially rail profile detection accuracy. Therefore, a registration algorithm with higher accuracy is urgently needed to adapt to higher requirements for wear detection and solve the above-mentioned defects in the existing technology. Summary of the invention

[0009] In order to solve the defects and shortcomings of the above-mentioned prior art, the present disclosure provides a new rail wear detection method, device and computer program, which realizes high-precision detection of rail wear through optimized data processing and registration technology, while reducing computing costs and improving detection efficiency. Among them, the present disclosure adopts a laser profilometer and a structured light sensor to measure the cross section of the rail, and designs a single-measurement single-sensor installation scheme based on the wear principle of the rail, which not only meets the wear measurement requirements, but also reduces costs, and reduces the power consumption and computing power requirements of the overall equipment.

[0010] In the embodiment disclosed in the present application, for the 20mm arc area of ​​60 rails, a tangent vector projection method based on the first-order derivative characteristics is used for rough screening. A sliding window is taken to smooth the data, and the tangent vector calculation of the linear least squares method is performed using this window, thereby improving the efficiency of the initial screening calculation. A rough registration algorithm for ICP fine registration is designed. The segmented arc point set obtained by the tangent vector projection method is subjected to the geometric constraints of the double centers of the rails using the Powell multiplier penalty function method and the LM algorithm, thereby achieving joint optimization of the double arc point set, and then the constraint is performed using the relationship between the double centers and the tangent points, thereby reducing the situation where the LM algorithm is difficult to consider the local optimal solution from the geometric characteristics of the double arc angles, improving the accuracy of the registration, reducing the dependence on the number of iterations of the subsequent ICP algorithm fine registration, reducing the calculation time, and improving the efficiency of real-time processing.

[0011] In a first aspect, the present disclosure provides a rail wear detection method, comprising:

[0012] Acquire measured point cloud data for identifying the actual cross-sectional profile of the rail;

[0013] Based on a preset screening strategy, the measured point cloud data is screened and processed by a tangent vector projection method to obtain a first circular arc for identifying a first preset position of the rail and a second circular arc for identifying a second preset position of the rail, and the coordinates of a first center point of the first circular arc and the coordinates of a second center point of the second circular arc are determined respectively;

[0014] According to the first center point coordinates, the second center point coordinates, the preset first center angle corresponding to the first arc, and the preset second center angle corresponding to the second arc, the first arc and the second arc are optimized by a preset optimization method, and the first center point coordinates and the second center point coordinates are updated according to the optimized first arc and the second arc;

[0015] Determine the tangent point position between the optimized first arc and the optimized second arc according to the updated coordinates of the first center point and the updated coordinates of the second center point;

[0016] Determine whether the tangent point position satisfies a preset condition, and if the preset condition is satisfied, obtain a first vector according to the updated first circle center point coordinates and the tangent point position;

[0017] Determine a rotation angle and a translation vector according to a comparison result between the first vector and a preset second vector of a standard rail profile, and generate a space transformation matrix based on the rotation angle and the translation vector;

[0018] The measured point cloud data is rotated and translated according to the spatial transformation matrix to obtain a registration point cloud, and the rail wear is determined according to a comparison result between the registration point cloud and a preset standard point cloud.

[0019] In some embodiments, the method of filtering and processing the measured point cloud data based on a preset filtering strategy by a tangent vector projection method to obtain a first arc for identifying a first preset position of the rail and a second arc for identifying a second preset position of the rail includes:

[0020] Determining a first preset position and a second preset position of the rail according to the preset screening strategy;

[0021] Through the tangent vector projection method, a first point cloud data set for identifying the first preset part and a second point cloud data set for identifying the second preset part are respectively screened from the measured point cloud data, and the first point cloud data set is used as the first arc corresponding to the first preset part, and the second point cloud data set is used as the second arc corresponding to the second preset part.

[0022] In some embodiments, the first preset position is a first preset circular arc area located at the waist of the rail; the second preset position is a second preset circular arc area located at the waist of the rail.

[0023] In some embodiments, the obtaining of measured point cloud data for identifying the actual cross-sectional profile of the rail includes:

[0024] Acquire measured point cloud data collected by a structured light sensor for identifying the actual cross-sectional profile of the rail.

[0025] In some embodiments, the method for determining the cut point position comprises:

[0026] The relative position relationship between the first center point coordinates and the second center point coordinates is used as a constraint condition, and the tangent point position is determined according to a nonlinear least squares method under the constraint condition.

[0027] In some embodiments, it also includes:

[0028] When the preset condition is not met, determining the endpoint position of the arc from the first arc according to the preset first center angle and the tangent point position;

[0029] Selecting data points between the endpoint position and the tangent point position from the first circular arc, using all the selected data points as the updated first circular arc, and determining the coordinates of the first center point of the updated first circular arc;

[0030] According to the first center point coordinates, the second center point coordinates, the preset first center angle, and the preset second center angle, the first arc and the second arc are optimized by the preset optimization method, and the first center point coordinates and the second center point coordinates are updated according to the optimized first arc and the second arc;

[0031] Determine the tangent point position between the optimized first arc and the optimized second arc according to the updated coordinates of the first center point and the updated coordinates of the second center point;

[0032] The first vector is regenerated based on the updated coordinates of the first circle center point and the tangent point position.

[0033] In a second aspect, the present disclosure provides a rail wear detection device, comprising:

[0034] Structured light sensor, used to collect measured point cloud data to identify the actual cross-sectional profile of the rail;

[0035] The processor is in communication with the structured light sensor and is used to detect rail wear according to the measured point cloud data by the method described above.

[0036] In some embodiments, the structured light sensor comprises:

[0037] Laser profile sensor. In some embodiments, the number of the structured light sensors is multiple, wherein:

[0038] The measured point cloud data of a single-side rail is collected by one or more structured light sensors.

[0039] In a third aspect, the present disclosure provides a computer program product, including a computer program, which implements the steps of the method described in the above aspects when executed by a processor.

[0040] The rail wear detection method provided by the present disclosure belongs to a rail registration method based on Powell multiplier penalty function and dynamic region segmentation and combined with ICP algorithm, which can achieve registration that meets the railway-related accuracy requirements and reliability requirements in real-time detection of rail wear in heavy-duty railways. In the case of using a single sensor to measure rail wear, it has the characteristics of small size and low cost; the rough screening of the area is achieved by using the tangent vector projection method, and the calculation efficiency is high; in the registration algorithm, the geometric relationship such as the center position and tangent point of the arc is used to realize the dynamic segmentation of the double arc area, and the information association of the two arcs is realized in the registration algorithm, which can provide better initial value information for subsequent fine registration, improve the overall accuracy, and reduce the dependence on the subsequent ICP algorithm; the rail registration method based on multiplier penalty function and dynamic region segmentation is adopted, and the registration accuracy is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The present disclosure will be described in more detail below based on embodiments and with reference to the accompanying drawings:

[0042] Figure 1 A schematic diagram of a flow chart of a rail wear detection method provided in an embodiment of the present disclosure;

[0043] Figure 2 A schematic diagram of a rail waist arc provided by an embodiment of the present disclosure;

[0044] Figure 3 A schematic diagram of a 20 mm arc of a rail provided in an embodiment of the present disclosure;

[0045] Figure 4 A schematic diagram of a tangential projection provided by an embodiment of the present disclosure;

[0046] Figure 5 A schematic flow chart of another rail wear detection method provided in an embodiment of the present disclosure;

[0047] Figure 6 A schematic structural diagram of a rail wear detection device provided in an embodiment of the present disclosure;

[0048] Figure 7 A schematic diagram of an application scenario of a rail wear detection device provided in an embodiment of the present disclosure;

[0049] Figure 8 A schematic diagram of rail wear provided in an embodiment of the present disclosure.

[0050] In the drawings, the same reference numerals are used for the same components, and the drawings are not drawn to scale. DETAILED DESCRIPTION

[0051] In order to enable those skilled in the art to better understand the technical solution of the present disclosure, and understand how the present disclosure applies technical means to solve technical problems, and to fully understand and implement the implementation process of the corresponding technical effects, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only embodiments of a part of the present disclosure, not all of the embodiments. The embodiments of the present disclosure and the various features in the embodiments can be combined with each other without conflict, and the technical solutions formed are all within the scope of protection of the present disclosure. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present disclosure.

[0052] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0053] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0054] In recent years, with the rapid development of laser technology, rail wear detection methods based on structured light have been widely used. Rail wear detection equipment using structured light detection generally obtains the cross-sectional profile data of the rail through devices such as laser displacement devices, and aligns the profile data with the standard profile to calculate the wear value of the rail.

[0055] In the technical solution of the present application, a corresponding data screening algorithm is disclosed. A tangent vector projection method based on the first-order derivative characteristics is used to perform a rough screening of a 20mm arc area (taking 60 rails as an example). A sliding window is taken to smooth the data, and the tangent vector calculation of the linear least squares method is performed with this window, thereby improving the efficiency of the initial screening calculation. A rough registration algorithm for ICP fine registration is designed. The segmented arc point set obtained by the tangent vector projection method uses the Powell multiplier penalty function method and the LM algorithm to perform geometric constraints on the double center of the rail, thereby realizing the joint optimization of the double arc point set, and then using the relationship between the double center and the tangent point to constrain, thereby reducing the situation where the LM algorithm is difficult to consider the local optimal solution from the double arc angle geometric characteristics, improving the registration accuracy, reducing the dependence on the number of iterations of the subsequent ICP algorithm fine registration, reducing the calculation time, and improving the efficiency of real-time processing.

[0056] The method disclosed in this application focuses on real-time wear detection in heavy-duty railways and can be applied to embedded devices on board vehicles. It proposes a wear detection method with higher computational efficiency without reducing the registration results. It not only improves the accuracy and stability of the registration, but also reduces the dependence on the ICP algorithm, reduces the number of iterations, and enables the registration accuracy of dynamic detection to reach 0.08mm and the standard deviation to 0.02mm.

[0057] Embodiment 1

[0058] The embodiment of the present disclosure provides a rail wear detection method. Figure 1 A schematic diagram of a rail wear detection method provided in an embodiment of the present disclosure is shown in FIG. Figure 1 As shown, the following steps are included:

[0059] Step 110: Acquire measured point cloud data for identifying the actual cross-sectional profile of the rail.

[0060] In some embodiments, the obtaining of measured point cloud data for identifying the actual cross-sectional profile of the rail includes:

[0061] Acquire measured point cloud data collected by a structured light sensor for identifying the actual cross-sectional profile of the rail.

[0062] Optionally, measured point cloud data collected by a structured light sensor may be obtained; wherein the measured point cloud data is used to identify the actual cross-sectional profile of the rail; wherein the structured light sensor includes a laser profile sensor.

[0063] Among them, the structured light sensor can accurately capture the three-dimensional information of the rail surface and provide basic data for subsequent wear detection.

[0064] The laser profile sensor is a device that uses the principle of laser triangulation to perform non-contact high-precision measurement. The built-in laser diode emits a linear laser beam to the object being measured, and the contour line formed is reflected by the imaging lens group to the photoelectric receiver for imaging. Subsequently, the image data is analyzed using image processing technology to accurately calculate the three-dimensional contour information of the object being measured.

[0065] Optionally, the rail on one side is measured by a laser profile sensor to collect the measured point cloud data. For example, a set of horizontal and vertical coordinate values ​​of 1200 points is collected; wherein the coordinate form of a single point is, for example, (78.09, 100.67) mm, representing the coordinates of a point. The original profile of the detected rail can be obtained through these point cloud data. The left and right laser profile sensors collect the left and right rail profiles respectively, and their data structure can be represented as a 2*1200 matrix, in which each element represents a point in the point cloud.

[0066] Step 120: Based on a preset screening strategy, the measured point cloud data is screened and processed by a tangent vector projection method to obtain a first arc for identifying a first preset position of the rail and a second arc for identifying a second preset position of the rail, and the coordinates of a first center point of the first arc and the coordinates of a second center point of the second arc are determined respectively.

[0067] In step 120, the process utilizes the above-mentioned tangent vector projection method to efficiently screen the point cloud data to accurately identify key parts of the rail, such as the 20 mm arc area and the 400 mm arc area of ​​the rail waist, thereby providing accurate geometric information for subsequent optimization processing and wear detection.

[0068] It is understandable that the preset filtering strategy is used to set which part of the data needs to be filtered, and can be set according to actual needs.

[0069] In some embodiments, the method of filtering and processing the measured point cloud data based on a preset filtering strategy by a tangent vector projection method to obtain a first arc for identifying a first preset position of the rail and a second arc for identifying a second preset position of the rail includes:

[0070] Determining a first preset position and a second preset position of the rail according to the preset screening strategy;

[0071] Through the tangent vector projection method, a first point cloud data set for identifying the first preset part and a second point cloud data set for identifying the second preset part are respectively screened from the measured point cloud data, and the first point cloud data set is used as the first arc corresponding to the first preset part, and the second point cloud data set is used as the second arc corresponding to the second preset part.

[0072] In some embodiments, the first preset position is a first preset circular arc area located at the waist of the rail; the second preset position is a second preset circular arc area located at the waist of the rail.

[0073] It is understandable that for different types of rails, the radius of the arc area at the same position on the rail waist is different, for example:

[0074] For 60 rails, the first preset position is a 20 mm arc area at the rail waist; the second preset position is a 400 mm arc area at the rail waist;

[0075] For the 75 rail, the first preset position is the 25 mm arc area at the waist of the rail; the second preset position is the 450 mm arc area at the waist of the rail.

[0076] In addition, in addition to the above two types of rails, there are other types of rails that also have arc areas of different sizes, which are not listed here one by one. Specifically, in this embodiment, a 60-rail is taken as an example.

[0077] In this step, according to the collected original data (i.e., measured point cloud data), a method based on tangent vector projection is proposed to perform coarse screening on the 20 mm arc area of ​​the rail profile, thereby obtaining the initial rail waist coarse segmentation point set, including the R20 arc point set and the R400 arc point set.

[0078] like Figure 2 As shown, O1 is the center of the R20 arc (i.e., the first arc, also known as the small arc), O2 is the center of the R400 arc (i.e., the second arc, also known as the large arc), and α is the center angle corresponding to the small arc, which is 66.54° in the standard rail.

[0079] The preprocessing algorithm for the original point cloud disclosed in the present invention adopts a method of arc area screening based on tangent vector projection, which has high calculation efficiency.

[0080] The arc segmentation under the existing technology generally adopts the curvature segmentation method, which has high computing cost and large volatility. The present invention adopts the tangent vector projection method based on the first-order derivative characteristics to perform a rough screening of the 20mm arc area. The key to the segmentation of the rail waist arc lies in the BC segment area with a larger central angle, that is, the 20mm area. The central angle of its arc is larger than that of the AB segment arc and contains more radius information. In the process of calculating the tangent vector, the downsampling method is adopted, and the data is filtered with a sliding window, and the linear least squares calculation is performed on this basis, which significantly improves the calculation efficiency of the tangent vector. Due to the installation position, as the angle of the sensor changes, the data density collected at the 400mm arc and the 20mm arc is different. For example, Figure 2 As shown in the figure. Since the arc at 20mm is parallel to the lens of the sensor, the acquisition density is the highest, which results in the tangent vector projection variance at the 20mm arc being smaller than that at the 400mm arc. Finally, through the dual screening of threshold and variance, the regional segmentation of the tangent vector projection method is achieved.

[0081] Step 130: optimize the first arc and the second arc by a preset optimization method according to the first center point coordinates, the second center point coordinates, the preset first center angle corresponding to the first arc, and the preset second center angle corresponding to the second arc, and update the first center point coordinates and the second center point coordinates according to the optimized first arc and the second arc.

[0082] In step 130, the optimization process aims to further improve the accuracy of arc recognition. Through preset optimization methods, such as the Powell multiplier penalty function method and the LM algorithm, the arc is finely adjusted to ensure the accuracy of the center point coordinates, thereby providing reliable geometric parameters for subsequent wear detection.

[0083] In some embodiments, the preset optimization method includes:

[0084] Powell multiplier penalty function method and LM algorithm.

[0085] Optionally, the multiplier coefficients and penalty function coefficients are initialized, and the center points O1 and O2 of the two arcs and the arc division point c are initialized according to the arc point set obtained by preprocessing, so as to realize the division of the rail waist arc.

[0086] Step 140: Determine the tangent point position between the optimized first arc and the optimized second arc according to the updated coordinates of the first center point and the second center point.

[0087] In step 140, the position where the first arc and the second arc are tangent to each other in the rail waist profile is determined, which ensures a smooth transition between the two arcs.

[0088] In some embodiments, the method for determining the cut point position comprises:

[0089] The relative position relationship between the first center point coordinates and the second center point coordinates is used as a constraint condition, and the tangent point position is determined according to a nonlinear least squares method under the constraint condition.

[0090] Step 150: determine whether the tangent point position satisfies a preset condition. If the preset condition is met, obtain a first vector according to the updated first circle center point coordinates and the tangent point position.

[0091] In step 150, the accuracy of the tangent point position is specifically ensured. By determining the preset conditions, inaccurate results caused by data noise or calculation errors can be effectively avoided, thereby ensuring the reliability of subsequent vector calculations.

[0092] In some embodiments, it also includes:

[0093] When the preset condition is not met, determining the endpoint position of the arc from the first arc according to the preset first center angle and the tangent point position;

[0094] Selecting data points between the endpoint position and the tangent point position from the first circular arc, using all the selected data points as the updated first circular arc, and determining the coordinates of the first center point of the updated first circular arc;

[0095] According to the first center point coordinates, the second center point coordinates, the preset first center angle, and the preset second center angle, the first arc and the second arc are optimized by the preset optimization method, and the first center point coordinates and the second center point coordinates are updated according to the optimized first arc and the second arc;

[0096] Determine the tangent point position between the optimized first arc and the optimized second arc according to the updated coordinates of the first center point and the updated coordinates of the second center point;

[0097] The first vector is regenerated based on the updated coordinates of the first circle center point and the tangent point position.

[0098] Optionally, the Powell multiplier penalty function method and the LM algorithm are used to perform geometric constraints on the dual centers of the rails, thereby realizing joint optimization of the dual arc point set. By iteratively updating the coordinates of the center of the circle and the coefficients of the multiplier penalty function, the tangent point position of the arc is searched and confirmed until the tangent point position converges, thereby realizing accurate fitting and optimization of the arc, and realizing the update of the coordinates of the center of the circle and the update of the coefficients of the multiplier penalty function; using the nonlinear least squares method under constraints, the relative position relationship of the centers of the circles is used as a constraint condition, the tangent point position c of the two arcs is searched, and it is determined whether the tangent point position coincides with the previous one. If it coincides, it proves that the tangent point coordinates are correct, otherwise the endpoint b of the R20 arc is searched according to the center angle α of the R20 arc and the existing tangent point coordinates, and the nonlinear least squares algorithm based on the Powell multiplier penalty function and the LM algorithm are jointly optimized according to the newly divided R20 arc bc, and then a new tangent point is obtained, and this cycle is repeated until the tangent point position converges.

[0099] This paper adopts a rail registration method based on multiplier penalty function and dynamic region segmentation, which has high registration accuracy. Figure 3 and Figure 4 ,in, Figure 3 This is a schematic diagram of a 20 mm arc of a rail provided in an embodiment of the present disclosure. Figure 4 A schematic diagram of a tangential projection provided in an embodiment of the present disclosure.

[0100] The present invention adopts the constrained least square method to realize the joint optimization of two arcs and dynamically segment the curve area according to the geometric relationship between the position of the center of the circle, the size of the center angle of the circle and the position of the tangent point, and the relative position relationship between the tangent point and the center of the circle and the size of the center angle of the circle in the iterative process. The geometric relationship diagram is as follows: Figure 4As shown. First, in order to ensure that the relative distance between the two arcs remains unchanged, the square of the distance between the centers of the two arcs minus the square of the theoretical Euclidean distance between the centers of the two arcs is used as a constraint condition to obtain the constraint equation; then, the multiplier penalty function is used to avoid the problem of computational difficulties caused by the penalty coefficient tending to infinity. The iteration of the multiplier penalty function can complete the update of the weights of the relevant constraints. In addition to constraining the center of the circle, the tangent area of ​​the two arcs also needs to be constrained; finally, the two arc areas are relocated according to the center angle to realize the dynamic segmentation of the least squares method under the constraint equation.

[0101] The present invention utilizes geometric relationships such as the center position and tangent point of the arc in the registration algorithm to realize dynamic segmentation of the double arc area, and realizes information association of the two arcs in the registration algorithm, thereby avoiding the lack of association between the two arcs in the ICP algorithm and each falling into a local optimal solution. The algorithm provides better initial value information for subsequent precise registration, improves the overall accuracy, and reduces the dependence on the subsequent ICP algorithm.

[0102] Furthermore, when the position of the tangent point converges, the vector O1c (ie, the first vector) can be obtained according to the center and the tangent point of the arc R20 that are finally determined.

[0103] Step 160: Compare the first vector with the preset second vector of the standard rail profile to obtain a rotation angle and a translation vector, and generate a space transformation matrix based on the rotation angle and the translation vector. Further, compare the vector O1c with O1'c' (i.e., the preset second vector) in the standard rail profile, and obtain its rotation angle and translation vector based on the comparison result, and then determine its space transformation matrix.

[0104] Step 170: performing rotation and translation operations on the measured point cloud data according to the space transformation matrix to obtain a registration point cloud, and determining the rail wear according to a comparison result between the registration point cloud and a preset standard point cloud.

[0105] Among them, the registration point cloud is obtained by accurately rotating and translating the measured point cloud data through the spatial transformation matrix. The comparison results between the registered point cloud and the standard point cloud can directly reflect the wear of the rail, thereby realizing high-precision detection of rail wear.

[0106] Furthermore, after the spatial transformation matrix is ​​determined, the original point cloud is rotated and translated using the spatial transformation matrix to obtain the final registration point cloud, and finally the rail wear is determined based on the comparison result between the registration point cloud and the preset standard point cloud.

[0107] In addition, in order to facilitate understanding of the technical solution of the present disclosure, reference may also be made to Figure 5 , Figure 5A schematic flow chart of another rail wear detection method provided in an embodiment of the present disclosure.

[0108] Embodiment 2

[0109] Based on the above embodiments, the present disclosure provides a rail wear detection device, such as Figure 6 As shown, the rail wear detection device comprises:

[0110] Structured light sensor, used to collect measured point cloud data to identify the actual cross-sectional profile of the rail;

[0111] A processor is connected in communication with the structured light sensor, and is used to detect rail wear according to the measured point cloud data by using the rail wear detection method as described above.

[0112] In some embodiments, the structured light sensor comprises a laser profile sensor.

[0113] In some embodiments, the number of the structured light sensors is multiple, wherein:

[0114] The measured point cloud data of a single-side rail is collected by one or more structured light sensors.

[0115] In this embodiment, the cross section of the rail is measured by a structured light sensor, and a single-measurement single-sensor solution is adopted based on the wear principle of the rail. This not only meets the wear measurement requirements, but also reduces the cost, and also reduces the power consumption and computing power requirements of the overall equipment.

[0116] For example, in this embodiment, the measured point cloud data of a single-side rail is collected by a structured light sensor as an example for explanation. Figure 7 The figure shows an application scenario of a rail wear detection device. Two laser profile sensors are included in this application scenario. The figure shows the relative positions of the left and right laser profile sensors and the rail. The laser emitted by the laser profile sensor can illuminate the tread and the rail waist. A single sensor is used to collect data on each side of the rail to realize the collection of profile data. The data is transmitted to the data processing unit for rail profile alignment and subsequent wear calculation.

[0117] The present invention uses a single sensor to measure rail wear, which has the characteristics of small size and low cost. The acquisition frame rate can be set to a maximum of 200fps (Frames Per Second, fps for short), and two laser profile sensors are used to detect the rails on both sides respectively. After the registration is achieved, the definition of rail wear is required to calculate the vertical wear, side wear and total wear of the rail, so that real-time detection of wear can be achieved.

[0118] refer to Figure 8 Rail wear refers to the phenomenon that the friction between the wheel and the rail causes wear on the rail head, which is divided into vertical wear, side wear and total wear. Vertical wear refers to the value read at 1 / 3 of the rail head width from the working edge, with the non-working edge and the non-working edge rail jaw as the measurement reference; side wear refers to the value read at 16mm below the rail surface, with the non-working edge and the non-working edge rail jaw as the measurement reference; based on vertical wear and side wear, the total wear is equal to the sum of vertical wear and 1 / 2 side wear. The wear of each part of the rail is shown in the figure, where W v Indicates vertical wear, W h Indicates the amount of side wear.

[0119] Since the rail waist has no direct contact with the wheel and no wear occurs, the above-mentioned profile registration algorithm is performed on the rail waist so that the point cloud of the measured rail waist coincides with the point cloud of the standard rail waist, and then the rail head wear is calculated.

[0120] In practical applications, the rail wear detection device disclosed in the present invention can be deployed on freight railway lines to monitor the wear of rails in real time. The device is equipped with two high-precision laser profile sensors, which are installed on both sides of the track to ensure full coverage of the tread and waist areas of the rails. The laser profile sensor scans the rail section at an acquisition frame rate of up to 200fps to generate a detailed point cloud data set. The data is then transmitted to a built-in high-performance processor, which uses the rail wear detection method disclosed in the present invention to analyze and process the data in real time.

[0121] Through the method in the embodiment, the device can accurately identify the vertical wear, lateral wear and total wear of the rails, and feed back this information to the railway maintenance department in real time. For example, in practical applications, the rail wear detection device disclosed in this application successfully identified that the vertical wear of a certain section of rails was 0.5mm, the lateral wear was 0.3mm, and the total wear was 0.65mm. These data are fed back to relevant maintenance personnel, helping them to promptly discover potential safety hazards and take corresponding maintenance measures, thereby effectively avoiding possible train derailment accidents.

[0122] Embodiment 3

[0123] On the basis of the above embodiments, this embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the above embodiments.

[0124] In some implementations of this embodiment, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the method described in the above embodiment are implemented.

[0125] In some implementations of this embodiment, a computer program product is provided, including a computer program, characterized in that when the computer program is executed by a processor, the steps of the method described in the above embodiment are implemented.

[0126] The processor may include, but is not limited to, one or more processors or microprocessors, etc. Each processor may be an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components to execute the methods in the above embodiments.

[0127] The computer-readable storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, and the computer-readable storage medium may include but is not limited to, for example, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (e.g., hard disk, floppy disk, solid-state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).

[0128] The computer-readable storage medium may also store at least one computer executable program / instruction, which may be, for example, a computer-readable instruction. The computer-readable storage medium includes, but is not limited to, for example, a volatile memory and / or a non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The computer-readable storage medium may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. For example, a non-transitory computer-readable storage medium may be connected to a computing device such as a computer, and then, when the computing device runs the computer-readable instructions stored on the computer-readable storage medium, the various methods described above may be performed.

[0129] In addition, the computer device may also include (but not limited to) a data bus, an input / output (I / O) bus, a display, and input / output devices (eg, keyboard, mouse, speaker, etc.), etc.

[0130] The processor may communicate with external devices via an I / O bus via a wired or wireless network.

[0131] In one embodiment, the at least one computer executable instruction may also be compiled into or constitute a software product / computer program product, wherein one or more computer executable instructions are executed by a processor to perform the various functions and / or method steps in the embodiments described in the present technology.

[0132] In the embodiments provided in the present disclosure, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0133] It should be noted that in the present disclosure, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element limited by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0134] Although the embodiments disclosed in the present disclosure are as above, the above contents are only embodiments adopted for facilitating the understanding of the present disclosure and are not intended to limit the present disclosure. Any technician in the technical field to which the present disclosure belongs can make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in the present disclosure, but the scope of patent protection of the present disclosure shall still be subject to the scope defined in the attached claims.

Claims

1. A rail wear detection method, characterized in that: The method comprises: Acquire measured point cloud data for identifying the actual cross-sectional profile of the rail; Based on a preset screening strategy, the measured point cloud data is screened and processed by a tangent vector projection method to obtain a first circular arc for identifying a first preset position of the rail and a second circular arc for identifying a second preset position of the rail, and the coordinates of a first center point of the first circular arc and the coordinates of a second center point of the second circular arc are determined respectively; Optimizing the first arc and the second arc by a preset optimization method according to the first center point coordinates, the second center point coordinates, a preset first center angle corresponding to the first arc, and a preset second center angle corresponding to the second arc, and updating the first center point coordinates and the second center point coordinates according to the optimized first arc and the second arc; Determine the tangent point position between the optimized first arc and the optimized second arc according to the updated coordinates of the first center point and the second center point; Determine whether the tangent point position satisfies a preset condition, and if the preset condition is satisfied, generate a first vector based on the updated first circle center point coordinates and the tangent point position; Performing a comparison process on the first vector and a preset second vector of a standard rail profile to obtain a rotation angle and a translation vector, and generating a space transformation matrix based on the rotation angle and the translation vector; The measured point cloud data is rotated and translated according to the spatial transformation matrix to obtain a registration point cloud, and the rail wear is determined according to a comparison result between the registration point cloud and a preset standard point cloud.

2. The method according to claim 1, characterized in that The method of screening and processing the measured point cloud data based on the preset screening strategy by a tangent vector projection method to obtain a first circular arc for identifying a first preset position of the rail and a second circular arc for identifying a second preset position of the rail includes: Determining a first preset position and a second preset position of the rail according to the preset screening strategy; Through the tangent vector projection method, a first point cloud data set for identifying the first preset part and a second point cloud data set for identifying the second preset part are respectively screened from the measured point cloud data, and the first point cloud data set is used as the first arc corresponding to the first preset part, and the second point cloud data set is used as the second arc corresponding to the second preset part.

3. The method according to claim 2, characterized in that The first preset position is a first preset arc area located at the rail waist of the rail; The second preset position is a second preset arc area located at the rail waist of the rail.

4. The method according to claim 1, characterized in that: The step of obtaining the measured point cloud data for identifying the actual cross-sectional profile of the rail includes: Acquire measured point cloud data collected by a structured light sensor for identifying the actual cross-sectional profile of the rail.

5. The method according to claim 1, characterized in that The method for determining the position of the cut point comprises: The relative position relationship between the first center point coordinates and the second center point coordinates is used as a constraint condition, and the tangent point position is determined according to a nonlinear least squares method under the constraint condition.

6. The method according to claim 1, characterized in that Also includes: When the preset condition is not met, determining the endpoint position of the arc from the first arc according to the preset first center angle and the tangent point position; Selecting data points between the endpoint position and the tangent point position from the first circular arc, using all the selected data points as the updated first circular arc, and determining the coordinates of the first center point of the updated first circular arc; According to the first center point coordinates, the second center point coordinates, the preset first center angle, and the preset second center angle, the first arc and the second arc are optimized by the preset optimization method, and the first center point coordinates and the second center point coordinates are updated according to the optimized first arc and the second arc; Determine the tangent point position between the optimized first arc and the optimized second arc according to the updated coordinates of the first center point and the updated coordinates of the second center point; The first vector is regenerated based on the updated coordinates of the first circle center point and the tangent point position.

7. A rail wear detection device, characterized in that: include: Structured light sensor, used to collect measured point cloud data to identify the actual cross-sectional profile of the rail; A processor is communicatively connected to the structured light sensor, and is used to detect rail wear according to the measured point cloud data by the method described in any one of claims 1 to 6.

8. The device according to claim 7, characterized in that The structured light sensor comprises: Laser profile sensor.

9. The device according to claim 7, characterized in that The number of the structured light sensors is multiple, wherein: The measured point cloud data of a single-side rail is collected by one or more structured light sensors.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images