Railway rail wear detection method, device, server and storage medium

Abstract

The application provides a rail wear detection method and device, a server and a storage medium, and relates to the technical field of rails. The rail wear detection method comprises the following steps: acquiring profile point cloud data of a to-be-detected rail in a first coordinate system; segmenting the profile point cloud data to obtain rail waist point cloud data and rail bottom point cloud data of the to-be-detected rail; determining a first center coordinate corresponding to the rail waist point cloud data and a second center coordinate of the rail bottom point cloud data; correcting the first center coordinate and the second center coordinate according to an LM algorithm to obtain a corrected second center coordinate and a corrected second center coordinate; adjusting the acquired profile point cloud data; and determining the total wear of the to-be-detected rail according to the distance between the rail head point cloud data of the adjusted profile point cloud data and the rail head point cloud data of standard profile point cloud data, so that the rail wear is accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rail, and particularly relates to a rail wear detection method and device, a server and a storage medium. BACKGROUND

[0002] Rail wear refers to the phenomenon that the rail head is gradually worn out due to the mutual friction between the train wheel and the rail during the operation of the train. In order to facilitate the maintenance of the rail by the staff, the wear of the rail needs to be detected.

[0003] At present, one way of detecting the rail wear is as follows: profile point cloud data of the rail is collected based on a laser sensor, the rail waist and the rail bottom arc profile in the profile point cloud data are segmented out through a vector data compression RDP algorithm, nonlinear quadratic fitting based on radius constraint is performed on the segmented rail waist and rail bottom to obtain the coordinates of the circle center. According to the obtained coordinates of the circle center, the rail measurement profile and the standard rail profile data are matched. Then, the rail wear is calculated according to the matching result.

[0004] However, the RDP algorithm has a large amount of calculation, and the accuracy of the obtained coordinates of the circle center is low. In this way, the calculation resources are wasted, and the reliability of the calculated rail wear is also low. SUMMARY

[0005] The present application provides a rail wear detection method and device, a server and a storage medium, which are used to solve the problem that the amount of calculation for determining the rail wear is large and the reliability is low in the prior art.

[0006] In a first aspect, the present application provides a method for detecting rail wear, comprising: obtaining profile point cloud data of a to-be-detected rail in a first coordinate system; segmenting the profile point cloud data to obtain rail waist point cloud data and rail bottom point cloud data of the to-be-detected rail; determining a first center coordinate corresponding to the rail waist point cloud data and a second center coordinate of the rail bottom point cloud data; correcting the first center coordinate and the second center coordinate according to an LM algorithm to obtain a corrected second center coordinate and a corrected second center coordinate, wherein a target function of the LM algorithm is: the first center coordinate is corrected so that the distances of each point in the rail waist point cloud data to the first center coordinate are respectively less than a first difference between a first standard radius and a set threshold value; and the second center coordinate is corrected so that the distances of each point in the rail bottom point cloud data to the second center coordinate are respectively less than a second difference between a second standard radius and the set threshold value; adjusting the obtained profile point cloud data so that the corrected first center coordinate overlaps with a third center coordinate corresponding to the rail waist point cloud data in the standard profile point cloud data in a second coordinate system, and so that the corrected second center coordinate overlaps with a fourth center coordinate corresponding to the rail bottom point cloud data in the standard profile point cloud data in the second coordinate system; and determining the total wear of the to-be-detected rail according to the distances between the rail head point cloud data of the adjusted profile point cloud data and the rail head point cloud data of the standard profile point cloud data.

[0007] In a possible implementation, the correcting the first center coordinate and the second center coordinate according to the LM algorithm to obtain the corrected second center coordinate and the corrected second center coordinate comprises: determining the distances of each point in the rail waist point cloud data to the first center coordinate and the distances of each point in the rail bottom point cloud data to the second center coordinate according to the LM algorithm; respectively judging whether the sum of the first differences is less than a set threshold value and whether the sum of the second differences is less than the set threshold value; when the sum of the first differences is greater than the set threshold value, correcting the first center coordinate and returning to determine the sum of the first differences until the sum of the first differences is less than the set threshold value; and when the sum of the second differences is greater than the set threshold value, correcting the second center coordinate and returning to determine the sum of the second differences until the sum of the second differences is less than the set threshold value.

[0008] In this way, the first center coordinate and the second center coordinate can be accurately and quickly corrected.

[0009] In a possible implementation, determining the first circle center coordinate corresponding to the rail waist point cloud data and the second circle center coordinate of the rail bottom point cloud data comprises: determining the first circle center coordinate corresponding to the rail waist point cloud data and the second circle center coordinate of the rail bottom point cloud data according to the least square method.

[0010] The first circle center coordinate corresponding to the rail waist point cloud data and the second circle center coordinate of the rail bottom point cloud data determined by the least square method have high accuracy, so that the iteration number of the LM algorithm can be reduced, the calculation resource is saved, and the efficiency is improved.

[0011] In a possible implementation, the first circle center coordinate corresponding to the rail waist point cloud data and the second circle center coordinate of the rail bottom point cloud data are determined according to the least square method, comprising: determining the first circle center coordinate corresponding to the rail waist point cloud data and the second circle center coordinate of the rail bottom point cloud data according to the coordinates (x BCi , y BCi ) of the points in the i th rail waist point cloud data and the formula: The first minimum value of F(A BC , B BC , R BC ) is obtained, and the value of the circle center coordinate (A BC , B BC ) under the first minimum value and the value of the radius R BC corresponding to the rail waist point cloud data are obtained; wherein F(A BC , B BC , R BC ) is the sum of the distance of each point of the rail waist point cloud data to the circle center corresponding to the rail waist point cloud data and the square difference of the radius corresponding to the rail waist point cloud data, and n is the number of points in the rail waist point cloud data; the coordinates (x DEi , y DEi ) of the points in the i th rail bottom point cloud data are obtained, and the formula: The second minimum value of F(A DE , B DE , R DE ) is obtained, and the value of the circle center coordinate (A DE , B DE ) under the second minimum value and the value of the radius R D corresponding to the rail bottom point cloud data are obtained; F(A DE , B DE , R DE ) is the sum of the distance of each point of the rail bottom point cloud data to the circle center corresponding to the rail bottom point cloud data and the square difference of the radius corresponding to the rail bottom point cloud data, and m is the number of points in the rail bottom point cloud data; the value of the circle center coordinate (A BC , B BC ) corresponding to the first minimum value is taken as the first circle center coordinate, and the value of the circle center coordinate (ADE , B DE ) as the second center coordinates.

[0012] In this way, the first center coordinates and the second center coordinates can be obtained accurately and quickly.

[0013] In a possible implementation, the acquired profile point cloud data is adjusted, including: determining a first slope and a first position of a first line connecting the corrected first center coordinates and the corrected second center coordinates; and determining a second slope and a second position of a second line connecting third center coordinates corresponding to the rail waist point cloud data in the standard profile point cloud data and fourth center coordinates corresponding to the rail bottom point cloud data; rotating the acquired profile point cloud data so that the first slope is consistent with the second slope; and translating the acquired profile point cloud data so that the first position is consistent with the second position.

[0014] The method described above can quickly match the acquired profile point cloud data and has high reliability.

[0015] In a possible implementation, after the acquired profile point cloud data is adjusted, the method provided in the application further includes: determining a third slope and a third position of a non-working edge of a rail head of the acquired profile point cloud data, and determining a fourth slope and a fourth position of a non-working edge of a rail head of the standard profile point cloud data; when a difference between the third slope and the fourth slope is greater than a set slope threshold, rotating the profile point cloud data so that the difference between the third slope and the fourth slope is less than the slope threshold; and when a distance between the third position and the fourth position is greater than a first distance threshold, translating and rotating the profile point cloud data so that the distance between the third position and the fourth position is less than the first distance threshold.

[0016] When the difference between the third slope and the fourth slope is greater than the set slope threshold, it indicates that there is still an error in the adjusted profile point cloud data, and thus the profile point cloud data is translated and rotated so that the distance between the third position and the fourth position is less than the first distance threshold, which can improve the accuracy of the acquired profile point cloud data.

[0017] In a possible implementation, after the profile point cloud data of the steel rail to be detected in the first coordinate system is acquired, the method provided in the application further includes: removing noise in the acquired profile point cloud data according to a two-dimensional amplitude limiting filtering algorithm.

[0018] In this way, the accuracy of the acquired profile point cloud data can be improved.

[0019] In a possible implementation, the profile point cloud data of the steel rail to be detected in the first coordinate system is acquired, including: receiving the profile point cloud data in the first coordinate system from a laser sensing module.

[0020] In this way, the profile point cloud data in the first coordinate system can be quickly obtained.

[0021] In a possible implementation, the laser sensing module includes a first laser camera located above the left side of the rail to be detected and a second laser camera located above the right side of the rail to be detected; the first sub-profile point cloud data in the first coordinate system from the first laser camera is received, wherein the first sub-profile point cloud data includes the rail head point cloud data, the rail waist point cloud data, and the rail bottom point cloud data on the left side of the rail to be detected; the second sub-profile point cloud data in the first coordinate system from the second laser camera is received, wherein the second sub-profile point cloud data includes the rail head point cloud data, the rail waist point cloud data, and the rail bottom point cloud data on the right side of the rail to be detected; and the first sub-profile point cloud data and the second sub-profile point cloud data are fused to obtain the profile point cloud data in the first coordinate system.

[0022] In a possible implementation, the rail waist point cloud data and the rail bottom point cloud data of the rail to be detected are obtained by segmenting the profile point cloud data, including: determining that the point located above in the two adjacent points in the profile point cloud data with a distance greater than a second distance threshold is the starting point of the rail waist of the rail to be detected, and the point located above is the ending point of the rail waist of the rail to be detected; and segmenting the profile point cloud data of the rail to be detected in the first coordinate system according to a preset distance relationship between the starting point of the rail waist and each segment of the rail to be detected, or a preset distance relationship between the ending point of the rail waist and each segment of the rail to be detected, to obtain the rail waist point cloud data and the rail bottom point cloud data of the rail to be detected.

[0023] It can be understood that the profile point cloud data of the rail to be detected in the first coordinate system is segmented according to the distance relationship between the starting point of the rail waist or the preset ending point of the rail waist and each segment of the rail to be detected, and the algorithm is simple and efficient.

[0024] In a second aspect, the present application provides a rail wear detection device, comprising: a data acquisition unit configured to acquire profile point cloud data of a to-be-detected rail in a first coordinate system; a profile division unit configured to segment the profile point cloud data to obtain rail waist point cloud data and rail bottom point cloud data of the to-be-detected rail; a center determination unit configured to determine a first center coordinate corresponding to the rail waist point cloud data and a second center coordinate of the rail bottom point cloud data; a center correction unit configured to correct the first center coordinate and the second center coordinate according to an LM algorithm to obtain a corrected second center coordinate and a corrected second center coordinate, wherein a target function of the LM algorithm is: the first center coordinate is corrected such that distances of each point in the rail waist point cloud data to the first center coordinate are respectively lower than a set threshold value by a first difference value of a first standard radius; and the second center coordinate is corrected such that distances of each point in the rail bottom point cloud data to the second center coordinate are respectively lower than the set threshold value by a second difference value of a second standard radius; a profile matching unit configured to adjust the acquired profile point cloud data such that the corrected first center coordinate overlaps with a third center coordinate corresponding to the rail waist point cloud data in standard profile point cloud data in a second coordinate system, and the corrected second center coordinate overlaps with a fourth center coordinate corresponding to the rail bottom point cloud data in the standard profile point cloud data in the second coordinate system; and a wear determination unit configured to determine total wear of the to-be-detected rail according to distances between rail head point cloud data of the adjusted profile point cloud data and rail head point cloud data of the standard profile point cloud data.

[0025] In a third aspect, the present application provides a computer readable storage medium storing a computer program, which, when executed by a processor, causes a computer to perform the method provided by the above-described embodiments.

[0026] In a fourth aspect, the present application provides a server, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, causes the server to perform the method provided by the above-described embodiments.

[0027] In a fifth aspect, the present application provides a computer program product, comprising a computer program, which, when executed, causes a computer to perform the method provided by the above-described embodiments.

[0028] The application provides a rail wear detection method and device, a server and a storage medium, which can determine a first circle center coordinate corresponding to rail waist point cloud data of a to-be-detected rail and a second circle center coordinate of rail bottom point cloud data; and correct the first circle center coordinate and the second circle center coordinate according to an LM algorithm to obtain a corrected second circle center coordinate and a corrected second circle center coordinate. Since the LM algorithm introduces a trust region and a damping coefficient, the speed of correcting the first circle center coordinate and the second circle center coordinate is improved, and the reliability of the corrected second circle center coordinate and the corrected second circle center coordinate obtained by using the LM algorithm is higher, and thus the accuracy of the obtained rail wear is also high. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0030] Figure 1 One of the flowcharts of the rail wear detection method provided by the embodiments of the present application;

[0031] Figure 2 The schematic diagram of the laser sensing module collecting profile point cloud data of the to-be-detected rail provided by the embodiments of the present application;

[0032] Figure 3 The second flowchart of the rail wear detection method provided by the embodiments of the present application;

[0033] Figure 4 The specific flowchart of S104 in Figure 1

[0034] The specific flowchart of S104 in Figure 5

[0035] The specific flowchart of S104 in Figure 6

[0036] The third flowchart of the rail wear detection method provided by the embodiments of the present application; Figure 7

[0037] The fourth flowchart of the rail wear detection method provided by the embodiments of the present application; Figure 8

[0038] The fourth flowchart of the rail wear detection method provided by the embodiments of the present application; Figure 9 The function module block diagram of the rail wear detection device provided by the embodiments of the present application.DETAILED DESCRIPTION

[0039] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments made by those of ordinary skill in the art under the inspiration of the embodiments of the present application belong to the scope of protection of the present application.

[0040] The terms "first", "second", "third", "fourth" and the like (if any) in the description, claims and drawings of the present application are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0041] Rail wear refers to the phenomenon that the rail head is gradually worn out due to the mutual friction between the wheel and the rail during the operation of the train. In order to facilitate the maintenance of the rail by the staff, the wear of the rail needs to be detected. At present, the detection of the rail wear has a large amount of calculation, and the accuracy of the obtained center coordinates is low. As a result, the calculation resources are wasted, and the reliability of the calculated rail wear is also low.

[0042] Based on the above technical problem, the present application provides a rail wear detection method and device, and a rail wear detection system and device.

[0043] The following will describe the technical solutions of the present application and how the technical solutions of the present application solve the above technical problems with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described again in some embodiments. The embodiments of the present application will be described with reference to the drawings.

[0044] As shown in Figure 1 The rail wear detection method provided by the embodiments of the present application is applied to a server. The rail wear detection method provided by the embodiments of the present application comprises the following steps.

[0045] S101: The server acquires profile point cloud data of the to-be-detected steel rail 100 in a first coordinate system.

[0046] Specifically, the specific implementation of S101 can be that the server receives profile point cloud data in the first coordinate system from the laser sensing module. In this way, the profile point cloud data in the first coordinate system can be quickly acquired. The first coordinate system is a coordinate system under the shooting angle of the laser sensing module.

[0047] As shown in the figure, Figure 2 The to-be-detected steel rail 100 includes a rail head A-A2 segment, a rail jaw A-B segment and A2-B2 segment, a rail waist B-D segment and B2-D2 segment, and a rail bottom segment D-D2 segment. The rail waist B-D segment and B2-D2 segment are respectively in a circular arc shape, and the joint part D-E segment and D2-E2 segment of the rail bottom and the rail waist are also respectively in a circular arc shape.

[0048] As shown in the figure, Figure 2 The laser sensing module includes a first laser camera 200 located at the upper left of the to-be-detected steel rail 100 and a second laser camera 300 located at the upper right of the to-be-detected steel rail 100. Then, the server can receive first sub-profile point cloud data in the first coordinate system from the first laser camera 200, wherein the first sub-profile point cloud data includes rail head point cloud data T1-A segment, rail waist point cloud data B-D segment, and rail bottom point cloud data D-E segment on the left side of the to-be-detected steel rail 100. It can be understood that since the first laser camera 200 is located at the upper left of the to-be-detected steel rail 100, under this detection angle, the rail jaw of the to-be-detected steel rail is blocked by the rail head. Therefore, the first sub-profile point cloud data does not include rail jaw point cloud data A-B segment on the left side of the to-be-detected steel rail 100.

[0049] In addition, the server can also receive second sub-profile point cloud data in the first coordinate system from the second laser camera 300, wherein the second sub-profile point cloud data includes rail head point cloud data T1-A2 segment, rail waist point cloud data B2-D2 segment, and rail bottom point cloud data B2-D2 segment on the right side of the to-be-detected steel rail 100. It can be understood that since the second laser camera 300 is located at the upper right of the to-be-detected steel rail 100, under this detection angle, the rail jaw of the to-be-detected steel rail is blocked by the rail head. Therefore, the second sub-profile point cloud data does not include rail jaw point cloud data A2-B2 segment on the right side of the to-be-detected steel rail 100.

[0050] In this way, the server can fuse the first sub-profile point cloud data and the second sub-profile point cloud data to obtain the profile point cloud data in the first coordinate system.

[0051] Further, the server can also remove the noise in the obtained profile point cloud data 400 according to a two-dimensional amplitude limiting filtering algorithm. Specifically, the two-dimensional amplitude limiting filtering algorithm is to set a maximum allowable deviation threshold, compare the distance D between two adjacent points in the obtained rail profile point cloud data, and if D>T, consider one of the two adjacent points as a random interference point and remove the random interference point. In this way, the accuracy of the obtained profile point cloud data 400 can be improved.

[0052] S102: The server segments the profile point cloud data to obtain rail waist point cloud data and rail bottom point cloud data of the to-be-detected rail 100.

[0053] In Figure 2 Based on the corresponding embodiments, as Figure 3 shown, the specific implementation of S102 can include:

[0054] S301: The server determines that, among two adjacent points in the profile point cloud data with a distance greater than the second distance threshold, the point located on the upper side is the starting point of the rail waist of the to-be-detected rail 100, and the point located on the upper side is the ending point of the rail waist of the to-be-detected rail 100.

[0055] It can be understood that, since in Figure 2 the corresponding embodiments, the first sub-profile point cloud data does not include the rail jaw point cloud data A-B segment on the left side of the to-be-detected rail 100; the second sub-profile point cloud data does not include the rail jaw point cloud data A2-B2 segment on the right side of the to-be-detected rail 100. Therefore, the server can determine that, among two adjacent points in the profile point cloud data with a distance greater than the second distance threshold, the point located on the upper side is the starting point of the rail waist of the to-be-detected rail 100, and the point located on the upper side is the ending point of the rail waist of the to-be-detected rail 100.

[0056] S302: The server segments the profile point cloud data of the to-be-detected rail 100 under the first coordinate system according to a preset distance relationship between the starting point of the rail waist and each segment of the to-be-detected rail 100, or a preset distance relationship between the ending point of the rail waist and each segment of the to-be-detected rail 100, to obtain rail waist point cloud data and rail bottom point cloud data of the to-be-detected rail 100.

[0057] It can be understood that the starting point of the preset rail and each segment of the steel rail 100 to be detected have a preset distance relationship. For example, the server can determine the starting point of the rail waist according to the distance between the starting point of the preset rail and the starting point of the rail waist; the server determines the ending point of the rail waist according to the distance between the starting point of the preset rail and the ending point of the rail waist, so that the server can obtain the rail waist point cloud data. For another example, the server can determine the starting point of the rail bottom according to the distance between the starting point of the preset rail and the starting point of the rail bottom; the server determines the ending point of the rail bottom according to the distance between the starting point of the preset rail and the ending point of the rail bottom, so that the rail bottom point cloud data can be obtained. For another example, the server can determine the starting point of the rail head according to the distance between the starting point of the preset rail and the starting point of the rail head; the server determines the ending point of the rail head according to the distance between the starting point of the preset rail and the ending point of the rail head, so that the rail head point cloud data can be obtained.

[0058] It can be understood that the rail waist point cloud data corresponding to the first center coordinates and the rail bottom point cloud data corresponding to the second center coordinates are segmented from the profile point cloud data of the steel rail 100 in the first coordinate system according to the distance relationship between the starting point of the rail waist or the preset rail and each segment of the steel rail 100 to be detected, and the algorithm is simple and efficient.

[0059] S103: The server determines the first center coordinates corresponding to the rail waist point cloud data and the second center coordinates of the rail bottom point cloud data.

[0060] Specifically, the server can determine the first center coordinates corresponding to the rail waist point cloud data and the second center coordinates of the rail bottom point cloud data according to the least square method. The accuracy of the first center coordinates corresponding to the rail waist point cloud data and the second center coordinates of the rail bottom point cloud data determined by the least square method is high, so that the iteration number of the subsequent LM algorithm can be reduced, the calculation resources are saved, and the efficiency is improved.

[0061] Exemplarily, S103 can be specifically implemented as: the server obtains the first minimum value of F(A BCi , B BCi , R BC ) according to the coordinates (x BC , y BC ) of the points in the i-th rail waist point cloud data and the formula: , and obtains the value of the center coordinates (A BC , B BC ) and the value of the radius R BC corresponding to the rail waist point cloud data under the first minimum value.

[0062] F(A BC , B BC , R BC) is the sum of the distance of each point of the rail waist point cloud data to the center of the circle corresponding to the rail waist point cloud data and the square difference of the radius corresponding to the rail waist point cloud data, and n is the number of points in the rail waist point cloud data. It can be understood that BCi = (x BCi -A BC ) 2 + (y BCi -B BC ) 3 -R BC 2 = x BC 2 + y BC 2 + C BC x BC + D BC y BC + E BC , and δ BC is the distance of the point of the rail waist point cloud data to the center of the corresponding circle and the square difference of the radius corresponding to the rail waist point cloud data. It can be understood that F(A BC , B BC , R BC ) = ∑δ BCi 2 = ∑[x BC 2 + y BC 2 + C BC x BC + D BC y BC + E BC ] 2 . Further, the server can obtain the minimum value of F(A BC , B BC , R BC ) by taking the partial derivative of the algorithm F(A BCi 2 = ∑[x BC 2 + y BC 2 + C BC x BC + D BC y BC + E BC ] 2 , and the minimum value corresponding to BC , B BC , R BC ). The minimum value of F(A

[0063] Further, the server can obtain the minimum value of F(A DEi , y DEi), and the formula: The second minimum value of F(A DE , B DE , R DE ) is obtained, and the values of the center coordinates (A DE , B DE ) and the value of the radius R DE corresponding to the sagittal point cloud data at the second minimum value are obtained.

[0064] wherein F(A DE , B DE , R DE ) is the sum of the distances of each point in the sagittal point cloud data to the center of the circle corresponding to the sagittal point cloud data and the square difference of the radius corresponding to the sagittal point cloud data, and m is the number of points in the sagittal point cloud data.

[0065] It can be understood that δ DEi = (x DEi -A DE ) 2 +(y DEi -B DE ) 2 -R DE 2 =x DE 2 +y DE 2 +C DE x DE +D DE y DE +E DE , and δ DE is the distance of the point in the sagittal point cloud data to the corresponding center and the square difference of the radius corresponding to the sagittal point cloud data. It can be understood that F(A DE , B DE , R DE )=∑δ DEi 2 =∑[x DE 2 +y DE 2 +C DE x DE +D DE y DE +E DE ] 2 . Further, the server can calculate the formula F(A DE , B DE , R De )=∑δ DEi 2 =∑[x DE 2 +y DE2 +C DE x DE +D DE y DE +E DE ] 2 Taking the partial derivative, we can obtain F(A) DE B DE R DE The minimum value of ), and the corresponding minimum value

[0066]

[0067] Finally, the server will assign the center coordinates (A) of the circle corresponding to the first minimum value. BC B BC The value of ) is used as the coordinate of the first center of the circle, and the coordinate of the center of the circle corresponding to the second minimum value (A) is used as the coordinate of the second minimum value. DE B DE The value of ) is used as the coordinate of the second center.

[0068] Understandably, based on the above S401-S403, the coordinates of the first and second center circles can be obtained accurately and quickly.

[0069] S104: The server corrects the coordinates of the first and second center circles according to the LM algorithm to obtain the corrected coordinates of the second center circle.

[0070] The objective function of the LM algorithm is to correct the coordinates of the first center of the circle so that the sum of the first difference between the distance of each point in the track waist point cloud data to the coordinates of the first center of the circle and the first standard radius is lower than a set threshold; and to correct the coordinates of the second center of the circle so that the sum of the second difference between the distance of each point in the track bottom point cloud data to the coordinates of the second center of the circle and the second standard radius is lower than a set threshold.

[0071] Specifically, such as Figure 4 As shown, S104 can be specifically implemented as follows:

[0072] S401: The server determines the distance from each point in the orbital waist point cloud data to the coordinates of the first circle center based on the LM algorithm, and sums the first difference between each distance and the first standard radius.

[0073] S402: The server determines whether the sum of the first differences is less than the set threshold. If not, it executes S403; if yes, it ends.

[0074] S403: The server corrects the coordinates of the first center and returns to execute S401.

[0075] S501: The server determines distances of each point in the rail bottom point cloud data to the second center coordinate according to the LM algorithm, and a sum of second differences between the distances and the second standard radius respectively.

[0076] S502: The server determines whether the sum of the second differences is less than a set threshold value. If not, S403 is executed; if yes, the process ends.

[0077] S503: The second center coordinate is corrected, and S501 is executed.

[0078] It should be noted that the steps of S501-S503 and S401-S403 have no sequence. In this way, the center coordinate can be corrected accurately and quickly.

[0079] S105: The server adjusts the obtained profile point cloud data 400 so that the corrected first center coordinate overlaps with a third center coordinate corresponding to the rail waist point cloud data in the standard profile point cloud data 500 in the second coordinate system, and so that the corrected second center coordinate overlaps with a fourth center coordinate corresponding to the rail bottom point cloud data in the standard profile point cloud data 500 in the second coordinate system.

[0080] Taking the profile point cloud data of the left side of the obtained steel rail to be detected as an example, as shown in (a) of Figure 5 , the first center coordinate of the corrected profile point cloud data is O1, and the second center coordinate of the corrected profile point cloud data is O2. As shown in (b) of Figure 5 , the third center coordinate corresponding to the rail waist point cloud data in the standard profile point cloud data 500 in the second coordinate system is O3, and the fourth center coordinate corresponding to the rail bottom point cloud data in the standard profile point cloud data 500 in the second coordinate system is O4. In this way, as shown in Figure 6 , the corrected profile point cloud data is translated and rotated so that the first center coordinate O1 overlaps with the third center coordinate O3, and the second center coordinate O2 overlaps with the fourth center coordinate O4.

[0081] Exemplarily, as shown in Figure 7 , the specific implementation of S105 can include:

[0082] S701: The server determines a first slope of a first line connecting the corrected first center coordinate and the corrected second center coordinate and a first position, and determines a second slope of a second line connecting the third center coordinate corresponding to the rail waist point cloud data in the standard profile point cloud data 500 and the fourth center coordinate corresponding to the rail bottom point cloud data and a second position.

[0083] S702: The server rotates the acquired profile point cloud data 400 so that the first slope and the second slope are consistent; the server translates the acquired profile point cloud data 400 so that the first position and the second position are consistent.

[0084] Understandably, when the first slope is the same as the second slope, and the first position is the same as the second position, the coordinates of the first center of the circle overlap with the coordinates of the second center of the circle, and the coordinates of the third center of the circle overlap with the coordinates of the fourth center of the circle.

[0085] Understandably, based on the above S701-S702, the acquired profile point cloud data 400 can be matched quickly and reliably.

[0086] S106: The server determines the total wear of the rail 100 to be inspected based on the distance between the rail head point cloud data of the adjusted profile point cloud data and the rail head point cloud data of the standard profile point cloud data 500.

[0087] like Figure 6 As shown, after adjusting the acquired profile point cloud data 400, the coordinates (A1, the first vertical wear calculation point) of the first vertical wear calculation point are found in the profile point cloud data. aest B aest The coordinates of the first level wear calculation point b1 (A) best B best ), and find the coordinates of the second vertical wear calculation point a2 in the standard profile point cloud data 500 (A aint B aint The coordinates of the second level wear calculation point b2 (A) bint B bint Furthermore, the server can be based on the formula. Determine the vertical wear ΔH of the rail 100 to be tested, and based on the formula ΔM=A best -A bint The horizontal wear ΔM of the rail to be inspected (100 mm) is determined. In this way, the server can use a formula... Determine the total wear ΔW of the rail to be tested, 100.

[0088] In summary, the rail wear detection method provided in this application can determine the first center coordinates of the rail web point cloud data and the second center coordinates of the rail base point cloud data corresponding to the rail web point cloud data of the rail to be tested. The first and second center coordinates are then corrected using the LM algorithm to obtain the corrected second center coordinates. Because the LM algorithm introduces a trust region and a damping coefficient, it improves the speed of correcting the first and second center coordinates. Furthermore, the reliability of the corrected second center coordinates obtained using the LM algorithm is higher, resulting in higher accuracy of the obtained rail wear data.

[0089] In addition, in Figure 1 Based on the corresponding embodiments, as Figure 8 As shown in S105, the method provided by the embodiments of the present application further includes:

[0090] S801: The server determines the third slope and the third position of the non-working edge of the rail head of the obtained profile point cloud data 400, and determines the fourth slope and the fourth position of the non-working edge of the rail head of the standard profile point cloud data 500.

[0091] S802: The server determines whether the difference between the third slope and the fourth slope is greater than a set slope threshold value, and whether the distance between the third position and the fourth position is greater than a first distance threshold value, and if so, performs S803.

[0092] S803: The server translates and rotates the profile point cloud data so that the difference between the third slope and the fourth slope is less than the slope threshold value, and the distance between the third position and the fourth position is less than the first distance threshold value.

[0093] It can be understood that based on the above S801-S803, when the difference between the third slope and the fourth slope is greater than the set slope threshold value, it indicates that the adjusted profile point cloud data still has errors, so that the profile point cloud data is translated and rotated to make the distance between the third position and the fourth position less than the first distance threshold value, which can improve the accuracy of the obtained profile point cloud data 400.

[0094] Please refer to Figure 9 The embodiments of the present application also provide a rail wear detection device 900. It should be noted that the basic principles and technical effects of the rail wear detection device 900 provided by the embodiments of the present application are the same as those of the above-mentioned embodiments. For brief description, the part not mentioned in the embodiments of the present application can refer to the corresponding content in the above-mentioned embodiments. As shown in Figure 9 The rail wear detection device 900 includes a data acquisition unit 901, a profile division unit 902, a center determination unit 903, a center correction unit 904, a profile matching unit 903, and a wear determination unit 904, wherein

[0095] The data acquisition unit 901 is configured to acquire profile point cloud data of a to-be-detected rail in a first coordinate system.

[0096] In one possible implementation, the data acquisition unit 901 is specifically configured to receive profile point cloud data in the first coordinate system from a laser sensing module.

[0097] In a possible implementation, the laser sensing module includes a first laser camera located above the left upper side of the steel rail to be detected and a second laser camera located above the right upper side of the steel rail to be detected. The data acquisition unit 901 is further configured to receive first sub-profile point cloud data in the first coordinate system from the first laser camera, where the first sub-profile point cloud data includes rail head point cloud data, rail waist point cloud data, and rail bottom point cloud data on the left side of the steel rail to be detected; receive second sub-profile point cloud data in the first coordinate system from the second laser camera, where the second sub-profile point cloud data includes rail head point cloud data, rail waist point cloud data, and rail bottom point cloud data on the right side of the steel rail to be detected; and fuse the first sub-profile point cloud data and the second sub-profile point cloud data to obtain profile point cloud data in the first coordinate system.

[0098] In a possible implementation, the data acquisition unit 901 is further configured to remove noise in the obtained profile point cloud data according to a two-dimensional amplitude limiting filtering algorithm.

[0099] The profile division unit 902 is configured to segment the profile point cloud data to obtain rail waist point cloud data and rail bottom point cloud data of the steel rail to be detected.

[0100] In a possible implementation, the profile division unit 902 is specifically configured to determine, from two adjacent points in the profile point cloud data with a distance greater than a second distance threshold, a point located above as a starting point of a rail waist of the steel rail to be detected and a point located above as an ending point of the rail waist of the steel rail to be detected; and segment the profile point cloud data of the steel rail to be detected in the first coordinate system according to a preset distance relationship between the starting point of the rail waist and each segment of the steel rail to be detected or a preset distance relationship between the ending point of the rail waist and each segment of the steel rail to be detected, to obtain the rail waist point cloud data and the rail bottom point cloud data of the steel rail to be detected.

[0101] The center determination unit 903 is configured to determine a first center coordinate corresponding to the rail waist point cloud data and a second center coordinate corresponding to the rail bottom point cloud data.

[0102] In a possible implementation, the center determination unit 903 is specifically configured to determine the first center coordinate corresponding to the rail waist point cloud data and the second center coordinate corresponding to the rail bottom point cloud data according to a least square method.

[0103] In a possible implementation, the center determination unit 903 is further specifically configured to determine the first center coordinate corresponding to the rail waist point cloud data and the second center coordinate corresponding to the rail bottom point cloud data according to coordinates (x BCi , y BCi ) of points in the i th rail waist point cloud data and a formula: F(A BC , B BC , R BCthe first minimum value, and obtaining the value of the center coordinates (A BC , B BC ) at the first minimum value and the value of the radius R BC corresponding to the track waist point cloud data; wherein F(A BC , B BC , R BC ) is the sum of the distance of each point in the track waist point cloud data to the center corresponding to the track waist point cloud data and the square difference of the radius corresponding to the track waist point cloud data, n is the number of points in the track waist point cloud data; according to the coordinates (x DEi , y DEi ) of the points in the i-th track bottom point cloud data, and the formula: obtaining the second minimum value of F(A DE , B DE , R DE ), and obtaining the value of the center coordinates (A DE , B DE ) at the second minimum value and the value of the radius R DE corresponding to the track bottom point cloud data; F(A DE , B DE , R DE ) is the sum of the distance of each point in the track bottom point cloud data to the center corresponding to the track bottom point cloud data and the square difference of the radius corresponding to the track bottom point cloud data, m is the number of points in the track bottom point cloud data; the value of the center coordinates (A BC , B BC ) at the first minimum value is taken as the first center coordinates, and the value of the center coordinates (A DE , B DE ) at the second minimum value is taken as the second center coordinates.

[0104] The center correction unit 904 is configured to correct the first center coordinates and the second center coordinates according to the LM algorithm to obtain the corrected second center coordinates and the corrected second center coordinates.

[0105] The objective function of the LM algorithm is to correct the first center coordinates so that the distance of each point in the track waist point cloud data to the first center coordinates is lower than a set threshold value, and to correct the second center coordinates so that the distance of each point in the track bottom point cloud data to the second center coordinates is lower than a set threshold value.

[0106] In a possible implementation, the center correction unit 904 is specifically configured to: determine distances of the points in the rail waist point cloud data to the first center coordinate according to the LM algorithm, and sums of first differences between the distances and the first standard radius respectively; determine distances of the points in the rail bottom point cloud data to the second center coordinate, and sums of second differences between the distances and the second standard radius respectively; respectively determine whether the sum of the first differences is less than a set threshold value and whether the sum of the second differences is less than the set threshold value; when the sum of the first differences is greater than the set threshold value, correct the first center coordinate, and return to determine the distances of the points in the rail waist point cloud data to the first center coordinate, and the sums of the first differences between the distances and the first standard radius, until the sum of the first differences is less than the set threshold value; when the sum of the second differences is greater than the set threshold value, correct the second center coordinate, and return to determine the distances of the points in the rail bottom point cloud data to the second center coordinate, and the sums of the second differences between the distances and the second standard radius, until the sum of the second differences is less than the set threshold value.

[0107] The profile matching unit 903 is configured to adjust the obtained profile point cloud data, so that the corrected first center coordinate overlaps with a third center coordinate corresponding to the rail waist point cloud data in the standard profile point cloud data in the second coordinate system, and so that the corrected second center coordinate overlaps with a fourth center coordinate corresponding to the rail bottom point cloud data in the standard profile point cloud data in the second coordinate system.

[0108] In a possible implementation, the profile matching unit 903 is specifically configured to: determine a first slope of a first line connecting the corrected first center coordinate and the corrected second center coordinate and a first position; determine a second slope of a second line connecting the third center coordinate corresponding to the rail waist point cloud data and the fourth center coordinate corresponding to the rail bottom point cloud data in the standard profile point cloud data and a second position; rotate the obtained profile point cloud data so that the first slope is consistent with the second slope; and translate the obtained profile point cloud data so that the first position is consistent with the second position.

[0109] In a possible implementation, the profile matching unit 903 is further configured to: determine a third slope of a non-working edge of a rail head of the obtained profile point cloud data and a third position, and determine a fourth slope of the non-working edge of the rail head of the standard profile point cloud data and a fourth position; when a difference between the third slope and the fourth slope is greater than a set slope threshold value, rotate the profile point cloud data so that the difference between the third slope and the fourth slope is less than the slope threshold value; and when a distance between the third position and the fourth position is greater than a first distance threshold value, translate and rotate the profile point cloud data so that the distance between the third position and the fourth position is less than the first distance threshold value.

[0110] The wear determination unit 904 is configured to determine total wear of the to-be-detected rail according to distances between rail head point cloud data of the adjusted profile point cloud data and rail head point cloud data of the standard profile point cloud data.

[0111] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to make the computer execute the method provided by the above embodiment.

[0112] The embodiment of the present application further provides a server, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to make the server execute the method provided by the above embodiment.

[0113] The embodiment of the present application further provides a computer program product, which comprises a computer program, and the computer program is executed to make the computer execute the method provided by the above embodiment.

[0114] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for detecting rail wear, characterized in that, The method includes: Acquire the profile point cloud data of the rail to be inspected in the first coordinate system; The profile point cloud data is segmented to obtain the rail web point cloud data and rail bottom point cloud data of the rail to be detected. Determine the coordinates of the first center of the orbital waist point cloud data and the coordinates of the second center of the orbital bottom point cloud data; The first and second center coordinates are corrected according to the LM algorithm to obtain the corrected second center coordinates. The objective function of the LM algorithm is to correct the first center coordinates so that the sum of the first difference between the distance from each point in the track waist point cloud data to the first center coordinate and the first standard radius is lower than a set threshold; and to correct the second center coordinates so that the sum of the second difference between the distance from each point in the track bottom point cloud data to the second center coordinate and the second standard radius is lower than the set threshold. The acquired profile point cloud data is adjusted so that the corrected first circle center coordinates overlap with the third circle center coordinates corresponding to the rail waist point cloud data in the standard profile point cloud data in the second coordinate system, and the corrected second circle center coordinates overlap with the fourth circle center coordinates corresponding to the rail bottom point cloud data in the standard profile point cloud data in the second coordinate system. The total wear of the rail to be tested is determined based on the distance between the rail head point cloud data of the adjusted profile point cloud data and the rail head point cloud data of the standard profile point cloud data.

2. The method according to claim 1, characterized in that, The step of correcting the coordinates of the first and second center circles according to the LM algorithm to obtain the corrected coordinates of the second center circle includes: The distance from each point in the rail waist point cloud data to the coordinates of the first circle center is determined according to the LM algorithm, and the sum of the first differences with the first standard radius is calculated; and the distance from each point in the rail bottom point cloud data to the coordinates of the second circle center is determined, and the sum of the second differences with the second standard radius is calculated. Determine whether the sum of the first differences is less than the set threshold and whether the sum of the second differences is less than the set threshold, respectively; When the sum of the first differences is greater than the set threshold, the first center coordinates are corrected, and the process returns to determine the distance from each point in the orbital point cloud data to the first center coordinates, and the sum of the first differences with the first standard radius, until the sum of the first differences is less than the set threshold. When the sum of the second differences is greater than the set threshold, the second center coordinates are corrected, and the process returns to determine the distance from each point in the orbital bottom point cloud data to the second center coordinates, and the sum of the second differences with the second standard radius, until the sum of the second differences is less than the set threshold.

3. The method according to claim 1, characterized in that, Determining the coordinates of the first center of the orbital waist point cloud data and the coordinates of the second center of the orbital bottom point cloud data includes: The coordinates of the first center of the orbital waist point cloud data and the coordinates of the second center of the orbital bottom point cloud data are determined using the least squares method.

4. The method according to claim 3, characterized in that, The step of determining the coordinates of the first center of the orbital waist point cloud data and the coordinates of the second center of the orbital bottom point cloud data according to the least squares method includes: Based on the coordinates (x, y) of the point in the i-th orbital waist point cloud data BCi y BCi ), and the formula: Get F(A) BC B BC R BC Find the first minimum value of ), and obtain the center coordinates (A) of the circle at the first minimum value. BC R BC The value of ) and the radius R corresponding to the orbital waist point cloud data. BC The value of ; Among them, F(A) BC B BC R BC ) represents the sum of the squared differences between the distances from each point in the orbital waist point cloud data to the center of the circle corresponding to the orbital waist point cloud data and the radius corresponding to the orbital waist point cloud data, and n represents the number of points in the orbital waist point cloud data. Based on the coordinates (x, y) of the point in the i-th orbital base point cloud data DEi y DEi ), and the formula: Get F(A) DE B DE R DE The second minimum value of ) is obtained, and the coordinates of the center of the circle at the second minimum value are obtained (A). DE B SE The value of ) and the radius R corresponding to the orbital base point cloud data. SE The values ​​of F(A) are given by F(A). DE B DE R DE ) represents the sum of the squared differences between the distances from each point in the track bottom point cloud data to the center of the corresponding circle in the track bottom point cloud data and the radius corresponding to the corresponding circle in the track bottom point cloud data, and m is the number of points in the track bottom point cloud data. The center coordinates (A) of the circle corresponding to the first minimum value BC B BC The value of ) is taken as the coordinate of the first center of the circle, and the coordinate of the center of the circle corresponding to the second minimum value (A) is taken as the coordinate of the first center of the circle. DE B DE The value of ) is used as the coordinate of the second center of the circle.

5. The method according to claim 1, characterized in that, The adjustment of the acquired profile point cloud data includes: Determine the first slope and first position of the first line connecting the corrected first center coordinates and the corrected second center coordinates; and determine the second slope and second position of the second line connecting the third center coordinates corresponding to the rail waist point cloud data and the fourth center coordinates corresponding to the rail bottom point cloud data in the standard profile point cloud data. The obtained profile point cloud data is rotated so that the first slope is consistent with the second slope; The contour point cloud data obtained by translation is made so that the first position is consistent with the second position.

6. The method according to claim 1, characterized in that, After adjusting the acquired profile point cloud data, the method further includes: Determine the third slope and third position of the non-working edge of the track head in the acquired profile point cloud data, and determine the fourth slope and fourth position of the non-working edge of the track head in the standard profile point cloud data; When the difference between the third slope and the fourth slope is greater than a set slope threshold, the profile point cloud data is rotated so that the difference between the third slope and the fourth slope is less than the slope threshold. When the distance between the third position and the fourth position is greater than the first distance threshold, the profile point cloud data is translated and rotated so that the distance between the third position and the fourth position is less than the first distance threshold.

7. The method according to claim 1, characterized in that, After acquiring the profile point cloud data of the rail to be detected in the first coordinate system, the method further includes: Noise in the acquired profile point cloud data is removed using a two-dimensional amplitude limiting filtering algorithm.

8. The method according to claim 1, characterized in that, The acquisition of the profile point cloud data of the rail to be detected in the first coordinate system includes: Receive profile point cloud data in the first coordinate system from the laser sensing module.

9. The method according to claim 8, characterized in that, The laser sensing module includes a first laser camera located on the upper left of the rail to be inspected and a second laser camera located on the upper right of the rail to be inspected; Receive first sub-profile point cloud data in a first coordinate system from the first laser camera, wherein the first sub-profile point cloud data includes the rail head point cloud data, rail waist point cloud data and rail bottom point cloud data on the left side of the rail to be detected; Receive second sub-profile point cloud data in the first coordinate system from the second laser camera, wherein the second sub-profile point cloud data includes the rail head point cloud data, rail web point cloud data and rail bottom point cloud data on the right side of the rail to be detected; By fusing the first sub-profile point cloud data and the second sub-profile point cloud data, profile point cloud data in the first coordinate system is obtained.

10. The method according to claim 9, characterized in that, The step of segmenting the profile point cloud data to obtain the rail web point cloud data and rail base point cloud data of the rail to be inspected includes: In the profile point cloud data where the distance is greater than a second distance threshold, the point located above the other is the starting point of the rail jaw of the rail to be detected, and the point located above the other is the ending point of the rail jaw of the rail to be detected. Based on the preset distance relationship between the starting point of the rail jaw and each segment of the rail to be tested, or the preset distance relationship between the ending point of the rail jaw and each segment of the rail to be tested, the profile point cloud data of the rail to be tested in the first coordinate system is segmented to obtain the rail waist point cloud data and rail bottom point cloud data of the rail to be tested.

11. A rail wear detection device, characterized in that, The device includes: The data acquisition unit is used to acquire the profile point cloud data of the rail to be inspected in the first coordinate system; The profile segmentation unit is used to segment the profile point cloud data to obtain the rail web point cloud data and rail bottom point cloud data of the rail to be detected. The center determination unit is used to determine the first center coordinates of the rail waist point cloud data and the second center coordinates of the rail bottom point cloud data; The center correction unit is used to correct the first center coordinates and the second center coordinates according to the LM algorithm to obtain the corrected second center coordinates. The objective function of the LM algorithm is to correct the first center coordinates so that the sum of the first differences between the distances from each point in the track waist point cloud data to the first center coordinates and the first standard radius is lower than a set threshold; and to correct the second center coordinates so that the sum of the second differences between the distances from each point in the track bottom point cloud data to the second center coordinates and the second standard radius is lower than the set threshold. The profile matching unit is used to adjust the acquired profile point cloud data so that the corrected first circle center coordinates overlap with the third circle center coordinates corresponding to the rail waist point cloud data in the standard profile point cloud data in the second coordinate system, and so that the corrected second circle center coordinates overlap with the fourth circle center coordinates corresponding to the rail bottom point cloud data in the standard profile point cloud data in the second coordinate system. The wear determination unit is used to determine the total wear of the rail to be tested based on the distance between the rail head point cloud data of the adjusted profile point cloud data and the rail head point cloud data of the standard profile point cloud data.

12. A server comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it causes the server to perform the method as described in any one of claims 1 to 10.

13. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it causes the computer to perform the method as described in any one of claims 1 to 10.

Images