Method and device for automatic identification and positioning of switch rail

Abstract

The application provides a turnout rail automatic identification and positioning method and device, the method comprises a reference rail starting point positioning step, a non-reference rail pre-positioning step, and a non-reference rail starting point positioning step. After the starting point of the reference rail serving as a detection starting point is automatically identified based on mileage information and a matching template, the position range of the starting point of the non-reference rail can be inferred based on the starting point of the reference rail and the distance between the preset starting point of the reference rail and the starting points of other non-reference rails. Then, the corresponding rail profile data is extracted based on the position range, and the accurate starting point position of each non-reference rail is further detected. In this way, the type of each rail of the turnout and the specific position of the starting point can be automatically determined, that is, the automatic identification, classification and starting point positioning of the multiple rails of the turnout are realized. On this basis, the corresponding detection algorithm can be further automatically called for each rail, so that the highly automatic detection of the rails of the turnout is realized, the operation amount is small, and the on-site detection is facilitated.

Description

Technical Field

[0001] The present invention belongs to the technical field of turnout track detection, and in particular relates to a method and device for automatically identifying and positioning a turnout rail. Background Art

[0002] Turnouts are crucial track sections that allow trains to change tracks and cross tracks. Turnouts are complex structures with numerous rail types, including stock rails, connecting rails, point rails, guard rails, wing rails, and heart rails. The rail components themselves are also complex. For example, the point rail is a curved, variable-section rail that mates with the stock rail. Consequently, turnouts require numerous inspection tasks, including geometric data testing, profile inspection of each rail component, rail surface damage detection, and inspection of the hazardous space in front of the heart rail. This makes inspection time-consuming and challenging.

[0003] Currently, turnout inspections are primarily conducted manually, with inspectors carrying tools like measuring tape to the corresponding rail components, taking measurements and recording the data as required. This results in low inspection efficiency and places high demands on the inspectors. Errors and omissions can easily occur if the inspectors misremember the rail component type. While some automated turnout inspection methods exist, these methods still face challenges, hindering their widespread adoption in the field. A key issue is the lack of methods for automatically identifying, classifying, and accurately locating multiple rail components in turnout areas that can be used for on-site inspections. For example, some methods use neural networks based on camera images to identify and classify rail components. While these methods can achieve a certain degree of automatic rail component classification, they cannot accurately determine the physical location of the rail components (the pixel coordinates of camera images differ from their actual physical dimensions). Other methods collect track profile data or images to create a three-dimensional model of the turnout area, which is then matched against a standard three-dimensional model to obtain rail component classification results or location information. However, these methods are computationally intensive, making them difficult to implement for on-site inspections.

[0004] Therefore, in order to achieve a more automated, accurate and reliable turnout area detection, a method is needed that can automatically identify, classify and accurately locate multiple rail parts in the turnout area. Summary of the Invention

[0005] The present invention is made to solve the above problems and aims to provide a method and apparatus capable of automatically identifying, classifying and locating multiple rail members of a turnout based on collected track profile data, thereby enabling highly automated track turnout detection. The present invention adopts the following technical solutions:

[0006] The present invention provides a method for automatically identifying and locating turnout rails, which is used to automatically identify and locate multiple rails of the turnout based on multi-frame track profile data of the turnout and corresponding mileage information collected. The method is characterized in that one of the multiple rails is a reference rail used as a detection starting point, and the others are non-reference rails. The method comprises: a reference rail starting point locating step: based on the mileage information and a predetermined reference rail starting point matching template, the track profile data of each frame is matched with profile features in turn to identify the starting point of the reference rail and obtain the corresponding mileage information; a non-reference rail pre-positioning step: based on the mileage information of the starting point of the reference rail and the distance information between the predetermined starting point of the reference rail and the starting points of each of the other non-reference rails, the mileage range information of the starting point of each non-reference rail is obtained; a non-reference rail starting point locating step: for each non-reference rail, the corresponding track profile data is extracted based on the mileage range information, and based on the extracted track profile data and the profile features of the starting point of the non-reference rail, the starting point of the corresponding non-reference rail is identified and the corresponding rail type and the mileage information are obtained.

[0007] The automatic identification and positioning method for turnout rails provided by the present invention may also have such technical features, wherein the reference rail starting point matching template includes several predetermined position ranges and their corresponding rail types, feature profile interception algorithms, feature data acquisition algorithms, and standard feature data, and the reference rail starting point positioning step includes: extracting the track profile data of each frame in sequence as the current track profile data based on the mileage information; judging whether there is profile data within the predetermined position range in the current track profile data, and when the judgment is yes, performing profile interception on the profile data within the predetermined range based on the corresponding feature profile interception algorithm; calculating the feature data of the corresponding rail based on the intercepted profile data and the feature data acquisition algorithm; judging whether the feature data is consistent with the corresponding standard feature data, and when the judgment is yes, using the corresponding rail type as the rail type of the reference rail, and using the mileage information corresponding to the current track profile data as the mileage information of the starting point of the reference rail.

[0008] The automatic identification and positioning method for turnout rails provided by the present invention may also have such technical features, wherein the reference rail is a point rail, the characteristic profile interception algorithm is used to intercept the profile line of the inner rail waist of the point rail and the profile line of the upper end of the rail bottom, and the characteristic data acquisition algorithm is used to calculate the angle between the two sections of the profile line as the characteristic data of the starting point of the point rail.

[0009] The automatic identification and positioning method for turnout rails provided by the present invention may also have such technical features, wherein the reference rail is a guard rail, the characteristic profile interception algorithm is used to intercept the profile line at the upper end of the rail head of the guard rail and the profile line on the inner side of the rail head, and the characteristic data acquisition algorithm is used to calculate the similarity between the two profile lines and the predetermined guard rail characteristic profile as the characteristic data of the starting point of the guard rail.

[0010] The automatic identification and positioning method for turnout rails provided by the present invention may also have such technical features, wherein the non-reference rail includes at least a heart rail and a wing rail, and the starting point positioning step of the non-reference rail includes: based on the corresponding mileage range information, extracting the multiple frames of track profile data corresponding to the starting point of the heart rail and the multiple frames of track profile data corresponding to the starting point of the wing rail from the multiple frames of track profile data; stacking the track profile data corresponding to the starting point of the heart rail according to the corresponding mileage information to generate a point cloud, and based on the point cloud, identifying the starting point of the heart rail by fitting a plane and obtaining the corresponding mileage information; based on the mileage information and a predetermined wing rail starting point matching template, performing profile matching on the multiple frames of track profile data corresponding to the starting point of the wing rail in turn to identify the starting point of the wing rail and obtain the corresponding mileage information.

[0011] The automatic identification and positioning method for turnout rails provided by the present invention may also have such technical features, wherein the non-reference rail includes at least a center rail and a wing rail, and the starting point positioning step of the non-reference rail includes: extracting multiple frames of track profile data corresponding to the starting point of the center rail from multiple frames of track profile data based on the corresponding mileage range information; stacking the track profile data corresponding to the starting point of the center rail according to the corresponding mileage information to generate a point cloud, and based on the point cloud, identifying the starting point of the center rail by fitting a plane and obtaining the corresponding mileage information; obtaining the corrected mileage range information of the starting point of the wing rail based on the mileage information of the starting point of the center rail and the predetermined distance information between the starting point of the center rail and the starting point of the wing rail; extracting multiple frames of track profile data corresponding to the starting point of the wing rail from multiple frames of track profile data based on the corrected mileage range information; based on the mileage information and the predetermined wing rail starting point matching template, performing profile matching on the multiple frames of track profile data corresponding to the starting point of the wing rail in turn, identifying the starting point of the wing rail and obtaining the corresponding mileage information.

[0012] The automatic identification and positioning method for turnout rails provided by the present invention may also have such technical features, wherein the reference rail is a heart rail, the characteristic profile interception algorithm is used to intercept the profile line of the upper end face of the point rail between the cross-sectional starting point and the theoretical starting point of the heart rail, and the characteristic data acquisition algorithm is used to calculate the inclination angle and length of the profile line as the characteristic data of the cross-sectional starting point of the heart rail.

[0013] The automatic identification and positioning method for turnout rails provided by the present invention may also have such a technical feature, wherein the starting point positioning step of the reference rail also includes: judging whether the type of the identified reference rail is a heart rail; when the judgment is yes, extracting multiple frames of the track profile data corresponding to the theoretical starting point of the heart rail based on the mileage information corresponding to the current track profile data and the distance information from the predetermined cross-sectional starting point of the heart rail to its theoretical starting point; stacking the extracted multiple frames of the track profile data according to their mileage information to generate a point cloud, and based on the point cloud, identifying the theoretical starting point of the heart rail by fitting a plane, and obtaining the corresponding mileage information as the mileage information of the starting point of the reference rail.

[0014] The automatic identification and positioning method for turnout rails provided by the present invention may also have such a technical feature, wherein, identifying the starting point of the center rail by fitting a plane includes the following steps: fitting the inclined surface at the starting point of the center rail based on the RANSAC algorithm; fitting the plane before the starting point of the center rail based on the RANSAC algorithm; calculating the coordinates of the intersection line of the inclined surface and the plane, or calculating the center line of the inclined surface and the center line of the plane respectively, and calculating the coordinates of the intersection point of the two center lines; based on the coordinates of the intersection line or the coordinates of the intersection point, obtaining the corresponding mileage information as the mileage information of the starting point of the center rail.

[0015] The automatic identification and positioning method for turnout rails provided by the present invention may also have such technical features, that is, after the step of obtaining the starting point of the reference rail, it also includes: a detection direction determination step: based on the identified rail type of the reference rail, the rail orientation corresponding to the track profile data of the reference rail, and the predetermined detection direction determination rules, the current detection direction is determined, and the current detection direction is one of the multiple travel directions of the turnout. wherein, in the other rail pre-positioning step, based on the current detection direction, the mileage information of the starting point of the reference rail, and the distance information, the rail type of each non-reference rail and the mileage range information of the starting point of each non-reference rail in the current detection direction are obtained.

[0016] The method for automatically identifying and locating a turnout rail provided by the present invention may also have such a technical feature, wherein the turnout is a single-opening turnout with four said directions of travel, the track profile data is obtained by a track detection device, the track detection device includes a first detection unit and a second detection unit, which are respectively used to obtain the track profile data of the rails on both sides of the track, and the detection direction determination rule includes: when the rail type of the identified reference rail is a pointed rail in a repelling state, and the track profile data of the identified reference rail comes from the first detection unit, it is determined that the current detection direction is the first direction of straight-through passage through the turnout; when the identified reference rail When the type of the reference rail component is a closed point rail and the track profile data of the reference rail component is identified to be from the first detection unit, it is determined that the current detection direction is the second direction of switching from a straight track to a curved track; when the type of the reference rail component is identified to be a guard rail and the track profile data of the reference rail component is identified to be from the second detection unit, it is determined that the current detection direction is the third direction opposite to the first direction; when the type of the reference rail component is identified to be a guard rail and the track profile data of the reference rail component is identified to be from the first detection unit, it is determined that the current detection direction is the fourth direction opposite to the second direction.

[0017] The present invention provides a turnout rail automatic identification and positioning device, which is used to automatically identify and locate multiple rails of the turnout based on the collected multi-frame track profile data of the turnout and the corresponding mileage information. It is characterized in that one of the multiple rails is a reference rail used as a detection starting point, and the others are non-reference rails. The device includes: a reference rail starting point positioning unit, which performs contour feature matching on each frame of the track profile data in sequence based on the mileage information and a predetermined reference rail feature matching template, identifies the starting point of the reference rail and obtains the corresponding mileage information; a non-reference rail pre-positioning unit, which obtains the mileage range information of the starting point of each non-reference rail based on the mileage information of the starting point of the reference rail and the predetermined distance information between the starting point of the reference rail and the starting point of each non-reference rail; and a non-reference rail starting point positioning unit, which extracts the corresponding track profile data for each non-reference rail based on the mileage range information, and identifies the starting point of the non-reference rail based on the extracted track profile data and the profile feature of the starting point of the non-reference rail, and obtains the corresponding rail type and the mileage information.

[0018] Functions and effects of the invention

[0019] According to the present invention, the method and apparatus for automatically identifying and locating turnout rails include a reference rail starting point location step, a non-reference rail pre-location step, and a non-reference rail starting point location step. After automatically identifying the starting point of the reference rail as the detection starting point based on mileage information and a matching template, the position range of the non-reference rail starting point can be inferred based on the distance between the reference rail starting point and the preset reference rail starting point and the starting points of other non-reference rails. Based on this position range, the corresponding track profile data is extracted to further detect the accurate starting position of each non-reference rail. In this way, the type of each turnout rail and the specific location of its starting point can be automatically determined, that is, the automatic identification and classification of multiple turnout rails and the starting point location can be achieved. Based on this benchmark, the corresponding detection algorithm can be further automatically called for each rail, thereby achieving highly automated detection of turnout rails. In addition, since the non-reference rails are pre-located, only the track profile data within the inferred small range needs to be extracted to further detect the accurate starting position of the non-reference rails. Therefore, the method has a low computational load and is very suitable for on-site detection applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of the method for automatically identifying and positioning a turnout rail member in the first embodiment of the present invention;

[0021] Figure 2 is a perspective view of a track inspection vehicle in a first embodiment of the present invention;

[0022] Figure 3 Schematic diagram of the distribution of sensors in the vehicle body unit in the first embodiment of the present invention;

[0023] Figure 4 is a perspective view of the load-bearing wheel assembly in the first embodiment of the present invention;

[0024] Figure 5 is a cross-sectional view of the load-bearing wheel assembly in the first embodiment of the present invention;

[0025] Figure 6 is a side view of the compression assembly in the first embodiment of the present invention;

[0026] Figure 7 is a three-dimensional diagram of the compression assembly in the first embodiment of the present invention;

[0027] Figure 8 This is a flow chart of the steps for locating the starting point of the reference rail member in the first embodiment of the present invention;

[0028] Figure 9 Schematic diagram of the predetermined position range in the template matching the starting point feature of the reference rail in the first embodiment of the present invention;

[0029] Figure 10 Schematic diagram of profile feature matching of the starting point of the switch rail in the fitted state in the first embodiment of the present invention;

[0030] Figure 11 Schematic diagram of profile feature matching of the starting point of the switch rail in the repulsive state in the first embodiment of the present invention;

[0031] Figure 12 Schematic diagram of the profile feature matching of the guardrail starting point in the first embodiment of the present invention;

[0032] Figure 13 This is a flow chart of the steps for locating the starting point of the non-reference rail member in the first embodiment of the present invention;

[0033] Figure 14 is a side view of the center rail near the starting point of the center rail according to the first embodiment of the present invention;

[0034] Figure 15 This is a flow chart of identifying the starting point of the center track by fitting a plane in the first embodiment of the present invention;

[0035] Figure 16 This is a block diagram of a device for automatically identifying and positioning a turnout rail member in a first embodiment of the present invention;

[0036] Figure 17 1 is a schematic structural diagram of a single turnout in the first embodiment of the present invention;

[0037] Figure 18 This is a flow chart of the steps for locating the starting point of the non-reference rail member in the second embodiment of the present invention;

[0038] Figure 19 It is a structural schematic diagram of the pointed rail and the slide bed plate in the third embodiment of the present invention.

[0039] Figure 20 This is a flow chart of the reference rail starting point positioning step in the third embodiment of the present invention;

[0040] Figure 21 This is a flow chart of the steps for locating the starting point of the reference rail member in the fourth embodiment of the present invention;

[0041] Figure 22 Schematic diagram of the profile feature matching of the starting point of the center rail section in the fourth embodiment of the present invention.

[0042] Reference numerals:

[0043] Track inspection vehicle 100; vehicle body 20; vehicle body unit 21; housing 211; oblique end portion 2111a; first line laser sensor 217a; second line laser sensor 217b; third line laser sensor 217c; fourth line laser sensor 217d; carrying mechanism 30; load-bearing wheel assembly 31; load-bearing wheel bracket 311; load-bearing wheel shaft 312; load-bearing wheel 313; encoder accommodating groove 3133; bearing 314; brake component 315; mileage encoder 317; pressing assembly 32; pressing bracket 321; mounting block 3211; guide rod 3212; guide component 322; guide rail 3221; slider 3222; elastic member 323; wheel body bracket 324; The clamping wheel 326; the locking component 327; the locking fitting 3271; the locking component 3272; the wrench component 328; the pushing mechanism 40; the switch rail automatic identification and positioning device 200; the positioning side information storage unit 201; the positioning side communication unit 202; the reference rail starting point positioning unit 203; the detection direction determination unit 204; the non-reference rail pre-positioning unit 205; the non-reference rail starting point positioning unit 206; the positioning side control unit 207; the single turnout 90; the basic rail 91; the point rail 92; the straight connecting rail 93; the guide curve rail 94; the guard rail 95; the wing rail 96; the heart rail 97; the heart rail theoretical starting point 971; the plane 972; the inclined surface 973; the heart rail section starting point 974; the slide bed plate 98. DETAILED DESCRIPTION

[0044] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the following is a detailed description of the automatic identification and positioning method and equipment for turnout rails of the present invention in combination with embodiments and drawings.

[0045] <Example 1>

[0046] This embodiment provides a method and device for automatically identifying and locating a turnout rail member, which is used to automatically identify and locate various turnout rail members (hereinafter referred to as rail members) at a turnout.

[0047] In this embodiment, a single turnout is taken as an example for specific description.

[0048] Figure 17 It is a structural schematic diagram of a single turnout in this embodiment.

[0049] like Figure 17As shown, a single-opening turnout 90 has three openings A, B, and C. Turnout 900 has four travel directions. Accordingly, four inspection directions, namely the first to fourth directions, can be used when inspecting the turnout rail. In the first direction, a train travels straight through the turnout from left to right in the figure without switching; in the second direction, a train travels straight through the turnout from right to left in the figure; in the third direction, a train switches at the turnout; and in the fourth direction, a train travels at the turnout and merges onto the mainline track.

[0050] The turnout rail components at the single turnout 90 include a stock rail 91, a point rail 92, a straight connecting rail 93, a guide rail 94, a guard rail 95, a wing rail 96, and a heart rail 97 (frog heart).

[0051] Along the first direction D1 (opening A to opening B), it passes through the positive starting point of the point rail 92 (the positive direction is from left to right in the figure, and the negative direction is from right to left), the straight connecting rail 93, the positive starting point of the wing rail 96, the positive starting point of the heart rail 97, and the positive starting point of the guard rail 95.

[0052] Along the second direction D2 (opening B to opening A), it passes the positive starting point of the point rail 92, the guide rail 94, the positive starting point of the wing rail 96, the positive starting point of the center rail 97, and the positive starting point of the guard rail 95. The second direction is the opposite of the first direction.

[0053] Along the third direction D3 (opening A to opening C), it passes through the reverse starting point of the guard rail 95, the reverse starting point of the wing rail 96, the reverse starting point of the heart rail 97, the straight connecting rail 93, and the reverse starting point of the point rail 92 in sequence.

[0054] Along the fourth direction D4 (opening C to opening A), it passes the reverse starting point of the guard rail 95, the reverse starting point of the wing rail 96, the starting point of the heart rail 97, the guide curve rail 94, and the starting point of the point rail 92. The fourth direction is the opposite of the third direction.

[0055] Figure 1 Flowchart of the method for automatic identification and positioning of turnout rail members in this embodiment.

[0056] like Figure 1 As shown, the method for automatically identifying and positioning a turnout rail member specifically includes the following steps:

[0057] Turnout data collection step S1: collecting track profile data and corresponding mileage information at the turnout.

[0058] Reference rail starting point positioning step S2: Based on the mileage information and the predetermined reference rail feature matching template, contour feature matching is performed on each frame of track profile data in turn to identify the starting point of the reference rail and obtain the corresponding rail type and mileage information;

[0059] Detection direction determination step S3: Determine the current detection direction based on the identified rail type of the reference rail, the track side corresponding to the identified track profile data of the reference rail, and a predetermined detection direction determination rule.

[0060] Non-reference rail pre-positioning step S4: Based on the current detection direction, the mileage information of the starting point of the reference rail and the predetermined distance information between the starting point of the reference rail and the starting points of other non-reference rails in the current detection direction, the mileage range information of the starting points of each non-reference rail is obtained.

[0061] Non-reference rail starting point positioning step S5: For each non-reference rail, based on the mileage range information and the corresponding predetermined rail feature matching template, the corresponding track profile data is matched with the contour features to identify the starting point of the corresponding non-reference rail and obtain the corresponding mileage information.

[0062] The above steps will be described in detail below.

[0063] Turnout data collection step S1: collecting track profile data and corresponding mileage information at the turnout.

[0064] As one example, a track inspection vehicle equipped with a wired laser sensor and a mileage encoder for collecting turnout track profile data on site is shown below.

[0065] Figure 2 It is a three-dimensional diagram of the track inspection vehicle in this embodiment.

[0066] like Figure 2 As shown, the track inspection vehicle 100 includes a vehicle body 20, a supporting mechanism 30 and a pushing mechanism 40. The vehicle body 20 is movably placed on two main line rails 9 through the supporting mechanism 30 at the bottom. The pushing mechanism 40 is installed on the vehicle body 20 for inspection personnel to push the vehicle body 20.

[0067] The vehicle body 20 is a two-part structure, comprising two mirror-symmetrical vehicle body units 21. Each vehicle body unit 21 comprises a housing 211 and sensors, electronic control components, etc. disposed within the housing 211.

[0068] Figure 3 Schematic diagram of the distribution of sensors in the vehicle body unit in this embodiment.

[0069] like Figure 3 As shown, four line laser sensors are installed within the housing 211 of each vehicle body unit 21. For ease of description, these are designated as a first line laser sensor 217a, a second line laser sensor 217b, a third line laser sensor 217c, and a fourth line laser sensor 217d. The housing 211 has an opening below, through which each of the four line laser sensors can detect the switch rail member below.

[0070] The first line laser sensor 217a is mounted within the housing 211 via a corresponding bracket and is located within the oblique end 2111a of the housing 211. The first line laser sensor 217a is mounted horizontally. For example, when the track inspection vehicle 100 is placed on the stock rail, the first line laser sensor 217a is located outside the stock rail, slightly above the stock rail, and facing the stock rail. The projected line laser covers the inactive edge of the stock rail, the lower jaw portion of the inactive edge, and a portion of the rail surface, enabling the acquisition of corresponding point position data.

[0071] The second line laser sensor 217b is tilted and mounted on the top of the housing 211 via a corresponding bracket. When the track inspection vehicle 100 is placed on the track, the second line laser sensor 217b is located above and outside the base rail and facing the base rail. The projected line laser covers the inactive side of the base rail, the waist of the inactive side, and the rail surface, and can obtain corresponding point position data.

[0072] The third line laser sensor 217c is tilted and mounted on the top of the housing 211 via a corresponding bracket, closer to the center of the vehicle body 20 than the second line laser sensor 217b. When the track inspection vehicle 100 is mounted on the base rail, the third line laser sensor 217c is located above and inside the base rail and facing the base rail. The line laser it projects covers the active edge of the base rail, the waist of the active edge, and the rail surface, enabling the acquisition of corresponding point position data. If other turnout rail components are located inside the base rail, the line laser projected by the third line laser sensor 217c can also cover or partially cover the upper end surface and inner side surface of the turnout rail component.

[0073] The fourth line laser sensor 217d is also mounted at an angle on the top of the housing 211 via a corresponding bracket. Its angle of inclination is approximately the same as that of the third laser sensor 217c, and its mounting position is closer to the center of the vehicle body. When the track inspection vehicle 100 is mounted on the stock rail, the fourth line laser sensor 217d is located above and inside the stock rails, between the two stock rails, and toward the base of the stock rails. If other turnout rail components are located inside the stock rails, the line laser projected by the fourth line laser sensor 217d covers the upper end surface and inner side surface of the turnout rail component.

[0074] In this embodiment, the multiple line laser sensors on the track inspection vehicle 100 are of the same model. In addition, before starting the inspection, the line laser sensors have been calibrated to obtain corresponding calibration parameters.

[0075] like Figure 2As shown, the supporting mechanism 30 includes four supporting wheel assemblies 31 and four clamping assemblies 32. The four supporting wheel assemblies 31 are mounted on the lower portions of the vehicle body 20 in the width direction, and are arranged in pairs near one end of the vehicle body 20 in the length direction, corresponding to the two rails 9. The four clamping assemblies 32 are also mounted on the lower portions of the vehicle body 20 in the width direction, and are located next to the four supporting wheel assemblies 31. The track inspection device 100 is movably mounted on the two rails 9 via the four supporting wheel assemblies 31. At this time, the four clamping assemblies 32 are located on the inner side of the rails 9 on the corresponding side.

[0076] Figure 4 3D is a perspective view of the load-bearing wheel assembly in this embodiment.

[0077] Figure 5 2 is a cross-sectional view of the load-bearing wheel assembly in this embodiment.

[0078] like Figures 4 and 5 As shown, the load-bearing wheel assembly 31 includes a load-bearing wheel bracket 311 , a load-bearing wheel shaft 312 , a load-bearing wheel 313 , a bearing 314 and a brake component 315 .

[0079] The ends of the load-bearing wheel shaft 312 are mounted on the load-bearing wheel bracket 311 via two bearings 314. The load-bearing wheel 313 is mounted on the load-bearing wheel shaft 312 and is restrained by the two bearings 314. In this embodiment, the load-bearing wheel 313 is made of an insulating material, preferably ceramic or POM plastic, and has an insulation resistance of no less than 1 MΩ. This prevents the detection device 100 from connecting the two rails and energizing them, thereby preventing interference with the track's electrical system.

[0080] The load-bearing wheels 313 have encoder accommodating slots 3133 within them. Two of the load-bearing wheels 313 have distance detection encoders 317 installed within the encoder accommodating slots 3133. These encoders are used to obtain mileage information as the load-bearing wheels 313 roll along the rails 9. The encoder accommodating slots 3133 are slightly larger than the outer dimensions of the distance detection encoders 317. Therefore, the distance detection encoders 317 and the load-bearing wheels 313 do not directly contact each other, preventing friction from affecting the rotation of the load-bearing wheels 313.

[0081] It should be noted that the structure of the vehicle body 20 and the installation position of the line laser sensors enable multiple line laser sensors to scan and obtain profile data for the same cross-section of the mainline rail 9. Under the control of the corresponding controller, during the movement of the cart, the multiple line laser sensors perform a synchronous scan after a predetermined time or distance. Each frame of point data (point cloud) collected by the line laser sensors contains hundreds to thousands of coordinate points, each of which contains mileage information (Y-axis), height information (Z-axis), width information (X-axis), and brightness information.

[0082] Figure 6 2 is a side view of the clamping assembly in this embodiment.

[0083] Figure 7 It is a three-dimensional diagram of the clamping assembly in this embodiment.

[0084] like Figures 6 and 7 As shown, the clamping assembly 32 includes a clamping bracket 321, a guide component 322, an elastic member 323, a wheel body bracket 324, a clamping wheel shaft (not shown in the figure), a clamping wheel 326, a locking component 327, and a wrench component 328.

[0085] The pressing bracket 321 includes a mounting block 3211 and a guide rod 3212. The guide rod 3212 is a cylindrical rod, and its extending direction is consistent with the length direction of the vehicle body 20.

[0086] The guide member 322 includes a guide rail 3221 and a slider 3222. The guide rail 3221 is fixedly mounted on the housing 211 and extends in the same direction as the guide rod 3212. The slider 3222 is slidably mounted on the guide rail 3221.

[0087] The elastic member 323 is a spring, which is also mounted on the guide rod 3212 , and one end of the spring abuts against the mounting block 3211 , and the other end is embedded in the cylindrical groove on the wheel body bracket 324 and abuts against the bottom of the groove.

[0088] The pinch wheel shaft is mounted in the notch at the end of the wheel body bracket 324, and the pinch wheel 326 is rotatably mounted on the pinch wheel shaft and partially embedded in the notch. The pinch wheel 326 is also made of the above-mentioned insulating material.

[0089] Therefore, under the action of the spring force of the elastic member 323, the wheel support 324 and the clamping wheel 326 thereon can be pressed toward the inner side of the rail, so that the clamping wheel 326 is closely attached to the active edge of the rail. In this embodiment, when the track inspection vehicle 100 is placed on the rail, the clamping wheel 326 is located 16 cm below the rail surface.

[0090] The locking component 327 includes a locking fitting 3271 and a locking member 3272 .

[0091] The locking member 3272 is a locking screw, which is installed at the corresponding locking member hole on the mounting block 3211, and the end of the screw can move along the hole.

[0092] The locking fitting 3271 is in the shape of a strip plate with multiple circular holes along its length. One end of the locking fitting 3271 is mounted on the wheel support 324, and the other end passes through the locking fitting hole in the mounting block 3211 and can move along the hole. The locking hole is connected to the locking fitting hole. When the screw end of the locking member 3272 extends downward, the screw end passes through the locking hole to the locking fitting hole and then penetrates the circular hole on the end of the locking fitting 3271, thereby fixing (locking) the relative position of the wheel support 324 and the mounting block 3211.

[0093] In this embodiment, the locking members 3272 of the two clamping assemblies 32 corresponding to one rail 9 are unlocked, and their elastic members 323 are free to move. In the two clamping assemblies 32 on the other side, the locking members 3272 are locked, and the positions of the wheel brackets 324 and the mounting blocks 3211 are fixed, that is, the positions of the clamping wheels 326 relative to the vehicle body 20 are fixed. In other words, in this embodiment, the two clamping assemblies 32 on one side are in the form of fixed side wheels, and the two on the other side are in the form of spring side wheels.

[0094] The wrench component 328 includes two combined connecting rods, which are used for the inspection personnel to adjust the relative position of the clamping wheel 326 with respect to the vehicle body 20, so as to make it easier to place the track inspection vehicle 100 on the rail.

[0095] The pushing mechanism 40 comprises a push rod and a notebook support installed at the end of the push rod. The angle of the push rod is adjustable, and the angle of the notebook support relative to the push rod is also adjustable.

[0096] In this embodiment, the inspection personnel place the above-mentioned track inspection vehicle 100 on the basic rail at the switch, start the on-board sensors and data collector, and then push the track inspection vehicle 100 along one of the above-mentioned four inspection directions. The track inspection vehicle 100 collects track profile data and corresponding mileage information while moving along the inspection direction.

[0097] Furthermore, in this embodiment, it is assumed that the line laser sensors in the two vehicle body units 21 acquire the track profile data synchronously.

[0098] Reference rail starting point positioning step S2: Based on the mileage information and the predetermined reference rail feature matching template, contour feature matching is performed on each frame of track profile data in turn to identify the starting point of the reference rail and obtain the corresponding rail type and mileage information.

[0099] In this embodiment, according to the detection direction, the rail type of the reference rail is a point rail in a contact state, a point rail in a repelling state, or a guard rail. Accordingly, the reference rail feature matching template includes:

[0100] Position range information of the fitted switch rail profile in the profile point data, switch rail feature profile interception algorithm, and switch rail feature data acquisition algorithm;

[0101] Position range information of the tip rail profile in the repelling state in the profile point data, tip rail feature profile interception algorithm, and tip rail feature data acquisition algorithm;

[0102] The position range information of the guardrail contour in the contour point data, the guardrail feature contour interception algorithm, and the guardrail feature data acquisition algorithm.

[0103] The following description will be made by taking the first direction and the reference rail being a pointed rail as an example.

[0104] Figure 8 It is a flow chart of the steps for positioning the starting point of the reference rail in this embodiment.

[0105] like Figure 8 As shown, the reference rail starting point positioning step S2 specifically includes the following steps:

[0106] Step S2-0: Set k=1, that is, specify the first frame of contour point data based on mileage information.

[0107] Step S2-1: for the k-th frame of contour point data, based on the reference rail starting point matching template, determine whether there is profile data within a plurality of predetermined position ranges in the frame of contour point data.

[0108] Step S2-2: When the judgment in step S2-1 is no, set k=k+1 and return to step S2-1, that is, specify the next frame of contour point data based on the mileage information.

[0109] Step S2-3: When the judgment in step S2-1 is yes, the silhouette data within the position range where the silhouette data exists is extracted, and the silhouette data is subjected to silhouette clipping based on the corresponding feature silhouette clipping algorithm;

[0110] Step S2-4: Calculate the characteristic data of the corresponding rail member based on the extracted profile and the corresponding characteristic data acquisition algorithm.

[0111] Step S2-5: Determine whether the calculated feature data is consistent with the corresponding predetermined standard feature data, and return to step S2-2 if the determination is no.

[0112] Step S2-6: When the judgment in step S2-5 is yes, the type of the reference rail is obtained, and the mileage information corresponding to the frame contour point data is used as the mileage information of the starting point of the reference rail.

[0113] In this embodiment, the track inspection vehicle 100 is pushed in a first direction to collect track profile data. Between the start point of the switch rail, the track profile collected only includes the profile of the stock rail. At the start point of the switch rail, the track profile includes the profile of the stock rail and the profile of the start point of the switch rail. Because the relative positions of the stock rail and the switch rail are generally fixed, and the profile features of the start point of the switch rail are known (with corresponding drawing data), the profile of the switch rail can be extracted from the measured track profile based on the relative positions of the stock rail and the switch rail. The profile features of the start point of the switch rail can then be used to determine whether the extracted profile is the profile of the start point of the switch rail.

[0114] Along the first direction, in the profile data collected by the left car body unit 21, the starting point of the point rail fits the base rail, and in the profile data collected by the right car body unit 21, the starting point of the point rail repels the base rail.

[0115] Figure 9 Schematic diagram of the predetermined position range of the reference rail feature matching template in this embodiment.

[0116] like Figure 9 As shown, in this embodiment, the reference rail feature matching template includes three predetermined position ranges and a contour interception template corresponding to each position range.

[0117] Among them, the position range P1 is located on the inner side of the basic rail profile (on the right side in the figure) and close to the basic rail, and the profile of the pointed rail in the fitted state is within this position range P1; accordingly, the profile cutting template corresponding to the position range P1 is used to cut the inner side rail waist profile line of the pointed rail and the profile line of the upper end face of the rail bottom.

[0118] Position range P2 is located inside the basic rail profile and relatively far away from the basic rail, with a predetermined distance between the basic rail and the point rail profile in the repulsive state within this position range P2; accordingly, the profile cutting template corresponding to position range P2 is the same as P1.

[0119] Position range P3 is also located inside the basic rail profile and close to the basic rail. Position range P3 partially overlaps with position range P1, and the guard rail profile is within this position range P3. Accordingly, the profile cutting template corresponding to position range P3 is used to cut the upper end profile line and the inner side profile line of the rail head of the guard rail.

[0120] Figure 10 It is a schematic diagram of feature matching and identification of the starting point of the pointed rail in the fitted state in this embodiment.

[0121] like Figure 10As shown, in this embodiment, in step S2-1, it is determined that profile data exists within position range P1. In step S2-3, the profile data within position range P1 is extracted and then intercepted using the switch rail profile interception template to obtain the profile line of the inner rail web of the switch rail (a roughly vertical straight line segment) and the profile line of the upper end surface of the switch rail base (a horizontal, slightly inclined straight line segment). The angle between these two straight lines is an obtuse angle slightly greater than 90 degrees. In step S2-4, it is determined whether these two straight lines exist. If the determination is yes, it is further determined whether the angle between the two straight lines is a predetermined value. If the determination is yes, it indicates that the profile of the switch rail starting point has been identified. In step S2-5, the reference rail type is determined to be a switch rail, and the mileage information corresponding to the profile point data in this frame is used as the mileage information of the switch rail starting point.

[0122] Figure 11 It is a schematic diagram of feature matching and identification of the starting point of the pointed rail in the repulsive state in this embodiment.

[0123] like Figure 11 As shown, similarly, for the pointed rail in the repulsive state, it is determined in step S2-1 that there is profile data within the position range P2, and in step S2-3, the profile data within the position range P2 is extracted, and the remaining steps are the same as above.

[0124] Figure 12 Schematic diagram of feature matching and recognition of the guardrail starting point in this embodiment.

[0125] like Figure 12 As shown, for the guardrail, in step S2-1, it is determined that there is profile data within the position range P3. In step S2-3, the profile data within the position range P3 is extracted, and the extracted profile data is intercepted according to the guardrail profile interception template to obtain the profile line of the upper end face of the rail head of the guardrail (a roughly horizontal straight line segment) and the profile line of the inner side of the rail head (a vertical and slightly inclined straight line segment). In step S2-4, it is determined whether these two straight line segments exist. If the judgment is yes, the two straight line segments are matched with the predetermined standard straight line segments and the similarity is calculated; it is further determined whether the calculated similarity is greater than the preset similarity threshold. If the judgment is yes, it means that the profile of the guardrail starting point has been identified. In step S2-5, the type of the reference rail is determined to be a guardrail, and the mileage information corresponding to the frame contour point data is used as the mileage information of the guardrail starting point.

[0126] It should be noted that the inner side of the guardrail head may be worn, causing the profile line on the inner side of the guardrail head to move outward. Therefore, when matching the two intercepted straight line segments, the profile line on the upper end face of the guardrail head is used as a reference, and the profile line on the inner side of the rail head can be appropriately shifted for matching.

[0127] Detection direction determination step S3: Determine the current detection direction based on the identified type of the reference rail, the track side corresponding to the identified track profile data of the reference rail, and a predetermined detection direction determination rule.

[0128] Specifically, for the single turnout of this embodiment, the detection direction determination rule is:

[0129] When the identified reference rail is a pointed rail in a repelled state, and the track profile data of the pointed rail is detected by the sensor from the left vehicle body unit (the left side based on the direction of the detecting person during pushing), the current detection direction is determined to be the first direction.

[0130] When the identified reference rail is a closed point rail and the track profile data detecting the point rail is from the sensor of the left vehicle body unit, it is determined that the current detection direction is the second direction.

[0131] When the identified reference rail is a guard rail and the rail profile data detecting the guard rail is from the sensor of the right vehicle body unit, it is determined that the current detection direction is the third direction.

[0132] When the identified reference rail is a guard rail and the rail profile data detecting the guard rail is from the sensor of the left vehicle body unit, it is determined that the current detection direction is the fourth direction.

[0133] Non-reference rail pre-positioning step S4: Based on the current detection direction, the mileage information of the starting point of the reference rail, the predetermined rail types of each non-reference rail, and the distance information between the starting point of the reference rail and each non-reference rail in the current detection direction, the rail types of other rails and the mileage range information of the starting points are obtained.

[0134] Specifically, the reference rail type, the types of other rails, and the distances between the reference rail's starting point and the starting points of other rails are preset for each detection direction. After obtaining the mileage value of the reference rail's starting point, this mileage value is added to the corresponding distance value to infer the rail type of each other rail and its mileage value in the current detection direction. This mileage value is then added to and subtracted from a predetermined value to generate mileage range information. Due to wear, drift, and other factors, the rail's starting point may not be at the calculated mileage value, but rather near it. Therefore, a mileage range near this location is inferred based on the calculated mileage value, and the corresponding rail starting point is searched for within this process range.

[0135] When the detection direction is the first direction, the reference rail is the point rail, and the non-reference rails include the center rail, guard rail, and wing rail. The distance from the point rail as the detection starting point to the center rail, guard rail, and wing rail in the first direction is preset.

[0136] After obtaining the mileage value of the point rail starting point, we add the corresponding predetermined distance values ​​to obtain the mileage values ​​of the center rail starting point, the mileage values ​​of the wing rail starting point, and the mileage values ​​of the guard rail starting point. These mileage values ​​are then added to the predetermined floating value to generate the corresponding mileage value range. The center rail starting point positioning algorithm requires plane fitting, so a relatively large floating value is added to its mileage value to obtain a sufficient number of data frames for plane fitting.

[0137] The situation in the second direction is similar to that in the first direction and will not be described in detail.

[0138] When the inspection direction is the third direction, the reference rail is the guard rail, and the non-reference rails include the center rail, wing rail, and point rail. Similarly, after obtaining the mileage value of the guard rail starting point, add the predetermined distance value to obtain the mileage value of the other rail components. The other steps are the same.

[0139] The situation in the fourth direction is similar to that in the third direction and will not be described in detail.

[0140] Non-reference rail starting point positioning step S5: For each non-reference rail, the corresponding track profile data is extracted based on the mileage range information, and based on the corresponding predetermined rail feature matching template, the extracted track profile data is matched with contour features to identify the starting point of the corresponding rail and obtain the corresponding mileage information.

[0141] Figure 13 Flowchart of the steps for locating the starting point of the non-reference rail in this embodiment.

[0142] like Figure 13 As shown, taking the first direction as an example, the non-reference rail starting point positioning step S5 specifically includes the following steps:

[0143] Step S5-1: extracting multiple frames of data corresponding to the starting point of the center track from multiple frames of track profile data based on the corresponding mileage range information.

[0144] Step S5-2: Stack the multi-frame track profile data corresponding to the starting point of the center track according to its mileage information, and convert the stacked profiles into a three-dimensional point cloud.

[0145] Among them, each frame of point data is data in the XZ plane coordinate system, and each point contains corresponding mileage information. According to the mileage information, multiple frames of point data are stacked along the Y-axis direction and converted into a three-dimensional point cloud according to the corresponding algorithm.

[0146] Step S5-3: Based on the three-dimensional point cloud, identify the starting point of the center track by fitting a plane and obtain the corresponding mileage information.

[0147] Figure 14 It is a side view of the center rail near the starting point of the center rail in this embodiment.

[0148] like Figure 14 As shown, between the starting point 974 of the heart rail section and the theoretical starting point 971 of the heart rail, the upper end surface of the heart rail 97 is a plane 972. After the theoretical starting point 971 of the heart rail, the upper end surface of the heart rail is a smaller inclined surface 973. The intersection between the plane 972 and the inclined surface 973 is the theoretical starting point 971 of the heart rail.

[0149] Figure 15 This is a flow chart of identifying the starting point of the center track by fitting a plane in this embodiment.

[0150] like Figure 15 As shown, step S5-3 specifically includes the following sub-steps:

[0151] Step S5-3-1: Based on the three-dimensional point cloud, the slope at the starting point of the center track is fitted using the RANSAC algorithm.

[0152] Step S5-3-2: Based on the three-dimensional point cloud, the plane before the center track starting point is fitted using the RANSAC algorithm.

[0153] Step S5-3-3, calculate the Y-axis coordinate of the point on the intersection line of the fitted slope and the fitted plane, and obtain the corresponding mileage information based on the Y-axis coordinate of the point on the intersection line as the mileage information of the starting point of the center track.

[0154] Alternatively, the center line of the fitted slope and the center line of the fitted plane (both are center lines along the Y-axis direction) can be calculated separately, and the Y-axis coordinate of the intersection of the two center lines can be calculated to obtain the corresponding mileage information.

[0155] Among them, the plane fitting by RANSAC algorithm is specifically as follows:

[0156] Randomly sample at least three points (the minimum number of points that can fit the plane model) in the corresponding 3D point cloud as the initial points. Assume the initial equation of the plane model AX+BY+CZ+D=0, and fit the plane model based on the multiple initial points to obtain the values ​​of the plane model parameters A, B, C, and D. In the 3D point cloud, traverse all points except the initial point, and calculate the distance d from each traversed point to the plane model:

[0157]

[0158] Where x0, y0, and z0 are the coordinates of the traversed points. Points whose distance d is less than a preset distance threshold are placed in the inner group. The ratio of the number of inner group points to the total number of points in the 3D point cloud is calculated. When this ratio is greater than a preset ratio threshold, the updated plane model parameters are obtained based on the inner group points, and the corresponding plane is fitted.

[0159] Step S5-4: extracting multiple frames of data corresponding to the starting point of the wing rail from the multiple frames of track profile data based on the corresponding mileage range information.

[0160] Step S5-5: Based on the mileage information and the predetermined wing rail starting point matching template, the extracted multi-frame track profile data is sequentially matched to identify the starting point of the wing rail and obtain the corresponding mileage information.

[0161] Step S5-6: extracting multiple frames of data corresponding to the starting point of the guardrail from the multiple frames of track profile data based on the corresponding mileage range information.

[0162] Step S5-7: Based on the mileage information and the predetermined guardrail starting point matching template, perform profile matching on the extracted multiple frames of track profile data in sequence to identify the starting point of the guardrail and obtain the corresponding mileage information.

[0163] The starting point matching and identification method of the guard rail has been described in detail above, and the starting point matching and identification method of the wing rail is similar.

[0164] Figure 16 It is a block diagram of the automatic identification and positioning device for turnout rail members in this embodiment.

[0165] like Figure 16 As shown, this embodiment also provides a rail automatic identification and positioning device 200 corresponding to the above method, which includes a positioning side information storage unit 201, a positioning side communication unit 202, a reference rail starting point positioning unit 203, a detection direction determination unit 204, a non-reference rail pre-positioning unit 205, a non-reference rail starting point positioning unit 206, and a positioning side control unit 207.

[0166] Among them, the positioning side information storage unit 201 stores the information required for the classification and positioning of the turnout rails, including the parameters of the above-mentioned templates, interception algorithms and matching algorithms, etc. The positioning side communication unit 202 is used to communicate with other devices, including obtaining the track profile data and corresponding mileage information collected by the line laser sensor from the data acquisition device of the track detection vehicle 100. The reference rail starting point positioning unit 203 obtains the type of the reference rail and the mileage information of the starting point according to the method of the above-mentioned step S2. The detection direction determination unit 204 determines the current detection direction according to the method of the above-mentioned step S3. The non-reference rail pre-positioning unit 205 performs pre-positioning of other rails according to the method of the above-mentioned step S4. The non-reference rail starting point positioning unit 206 calculates the starting position of multiple other rails according to the method of the above-mentioned step S5. The positioning side control unit 207 is used to control the operation of the above-mentioned functional units.

[0167] As one example, the rail automatic identification and positioning device 200 is a laptop computer equipped with a corresponding computer program. This laptop computer is placed on a laptop holder on the track inspection vehicle 100 and connected to the vehicle's onboard data acquisition device via a cable. Inspectors can use this laptop computer to conveniently obtain the classification and positioning results of turnout rail components.

[0168] In this embodiment, parts not described in detail are well-known technologies in the art.

[0169] Functions and effects of embodiment 1

[0170] According to the method and device for automatically identifying and locating turnout rails provided in this embodiment, the method includes a reference rail starting point locating step, a non-reference rail pre-locating step, and a non-reference rail starting point locating step. After automatically identifying the starting point of the reference rail as the detection starting point based on mileage information and a matching template, the position range of the non-reference rail starting point can be inferred based on the distance between the starting point of the reference rail and the preset reference rail starting point and the starting points of other non-reference rails. Based on this position range, the corresponding track profile data is extracted to further detect the accurate starting position of each non-reference rail. In this way, the type of each turnout rail and the specific location of its starting point can be automatically determined, that is, the automatic identification and classification of multiple turnout rails and the starting point location are achieved. Based on this benchmark, the corresponding detection algorithm can be further automatically called for each rail, thereby achieving highly automated detection of turnout rails. In addition, since the non-reference rails are pre-located, only the track profile data within the inferred small range needs to be extracted to further detect the accurate starting position of the non-reference rails. Therefore, the method has a small amount of computation and is very conducive to the application of on-site detection.

[0171] In the embodiment, the reference rail is a point rail or a guard rail at a switch. The corresponding profile is found according to the relative position of the point rail or guard rail and the base rail, and the profile of the inner waist of the point rail and the profile of the upper end of the rail bottom, the profile of the upper end of the guard rail head and the profile of the inner side of the rail head are cut out as their characteristic profiles to calculate the characteristic data, so as to identify the starting point of the point rail and the starting point of the guard rail based on the characteristic data. Therefore, not only can the type and starting position of the reference rail be effectively identified based on the relative position of the rail parts and the profile characteristics of the rail parts, but also only a small amount of profiles are extracted to calculate the characteristic data, the amount of calculation is small, and it is suitable for on-site detection.

[0172] In the embodiment, the mileage range information of other rail members is calculated based on the identified starting point of the reference rail member, and then the track profile data corresponding to the mileage range is extracted to calculate the mileage information of the starting point of the corresponding rail member, that is, based on the identified starting point of the reference rail member and the structural characteristics of the switch, the positions of other rail members are first inferred, and then the starting points of each corresponding rail member are specifically located near the inferred position, so that the automatic classification and starting point positioning of multiple rail members in a switch with a complex structure can be achieved with a relatively simple and clear algorithm; and similarly, since only the track profile data within the inferred mileage range is extracted to calculate the starting position of the rail member, without processing a large amount of profile data, the amount of calculation is small, and the results of rail classification and starting point positioning can be quickly obtained during on-site inspection.

[0173] Furthermore, when calculating the starting point of the center rail, based on the structural characteristics of the center rail starting point, a plane fitting method is used to fit the plane before the starting point and the inclined plane at the starting point. The coordinates of the intersection line or center line of the two planes are calculated. Based on these coordinates, the corresponding mileage information is obtained as the mileage information of the center rail starting point. Due to the unique structure of the center rail starting point, it is difficult to accurately identify its theoretical starting point based on a single frame of profile data. However, the center rail is an extremely important rail component in the turnout. In this embodiment, the plane fitting method can be used to accurately obtain the starting point of the center rail.

[0174] In the embodiment, the direction of the current cart inspection can also be automatically determined based on the type of reference rail member identified and the track side to which the track profile data of the reference rail member corresponds (the sensor of which vehicle body unit the data comes from). Therefore, even for a turnout with a complex structure, multiple openings, and multiple directions of travel, the inspector only needs to start the inspection equipment mounted on the track inspection vehicle and push the track inspection vehicle along the track to collect data. The method of this embodiment will automatically determine the inspection direction based on the collected data and correspondingly identify and locate each rail member under the inspection direction; the inspector does not need to set the inspection direction, set the inspection algorithm based on the rail member distribution in the inspection direction, etc., which can avoid errors and omissions caused by manual settings, and has a high degree of automation, which can reduce the burden on the inspector and improve the efficiency of turnout inspection.

[0175] Furthermore, based on the identified type and starting position of each turnout rail, profile data at a specified turnout position or within a specified range can be accurately extracted for geometric detection based on the starting position and mileage information; alternatively, based on the starting position of each rail, the corresponding profile data of the rail or the profile data of the specified position of the rail can be extracted, and the corresponding detection algorithm can be called to perform profile detection, relative position detection, etc. of the rail. It can be seen that the method of this embodiment is of great significance for achieving highly automated turnout structure detection.

[0176] <Example 2>

[0177] This embodiment provides a method and device for automatically identifying and positioning a switch rail. Compared with the first embodiment, the second embodiment assigns the same symbols to the components having the same structure as the first embodiment and omits the corresponding descriptions.

[0178] Figure 18 Flowchart of the steps for locating the starting point of the non-reference rail in this embodiment.

[0179] like Figure 18 As shown, in this embodiment, the non-reference rail starting point positioning step S5 specifically includes the following sub-steps:

[0180] Step S5-1: extracting multiple frames of data corresponding to the starting point of the center track from the multiple frames of track profile data based on the corresponding mileage range information.

[0181] Step S5-2: Stack the multi-frame track profile data corresponding to the starting point of the center track according to its mileage information, and convert the stacked profiles into a three-dimensional point cloud.

[0182] Step S5-3: Based on the three-dimensional point cloud, identify the starting point of the center track by fitting a plane and obtain the corresponding mileage information.

[0183] Step S5-4: Based on the mileage information of the center rail starting point and the distance information from the predetermined center rail starting point to the starting points of other nearby rail members, the corrected mileage range information of other rail members is calculated.

[0184] Taking the first direction as an example, after calculating the mileage value of the starting point of the center rail, the mileage range information of the starting points of the wing rails and the guard rails near the center rail is recalculated based on the mileage value.

[0185] Step S5-5: Based on the corresponding corrected mileage range information, extract multiple frames of data corresponding to the starting points of other rail members from the multiple frames of track profile data.

[0186] Step S5-6: Based on the mileage information and the corresponding rail starting point matching template, perform profile matching on the extracted multi-frame track profile data of each other rail, identify the starting point of each other rail, and obtain the corresponding rail type and mileage information.

[0187] That is, in this embodiment, the starting points of the wing rails and guard rails nearby are located based on the starting point of the center rail.

[0188] In this embodiment, other method steps are the same as those in the first embodiment and will not be described again.

[0189] Functions and effects of Example 2

[0190] According to the method and device for automatically identifying and locating turnout rails provided in this embodiment, on the basis of the first embodiment, other rails in the vicinity of the identified center rail starting point are further located, thereby correcting the positioning of other rails in the vicinity of the center rail. Since the starting point of the point rail is relatively far from these rails, the mileage calculated based on the point rail may have a certain deviation, and it is necessary to set a relatively large mileage range to extract profile data to prevent the starting point of the rail from falling outside the mileage range. However, through the method of this embodiment, the positioning of the center rail is used to further locate the rails in the vicinity, and the positioning is more accurate. Therefore, a relatively smaller mileage range can be set to extract profile data, and profile matching and identification of the starting points of other rails can be performed, which is conducive to the accurate positioning of other rails and further reduces the amount of calculation.

[0191] <Example 3>

[0192] This embodiment provides a method and device for automatically identifying and positioning a switch rail. Compared with the first embodiment, the third embodiment assigns the same symbols to the components having the same structure as the first embodiment and omits the corresponding descriptions.

[0193] Figure 19 Schematic diagram of the structure of the point rail and the slide bed in this embodiment.

[0194] like Figure 19 As shown, below the point rail 92 are a slide bed plate 98 and a sleeper (not shown in the figure), the point rail 92 is supported on the slide bed plate 98, and the slide bed plate 98 is fixed on the sleeper.

[0195] In this embodiment, taking the first direction as an example, after the starting point of the point rail is identified, the starting point of the slide bed plate below it is found, and the starting point of the slide bed plate is used as the detection starting point.

[0196] Figure 20 Flowchart of the method for automatic identification and positioning of turnout rail members in this embodiment.

[0197] like Figure 20 As shown, in this embodiment, after the reference rail starting point positioning step S2, it also includes a detection starting point positioning step S2a: based on the mileage information of the identified reference rail starting point, multiple frames of track profile data within the corresponding mileage range are extracted, and based on the mileage information and the predetermined slide plate starting point matching template, the extracted multiple frames of track profile data are matched with profile features to identify the starting point of the slide plate under the reference rail and obtain the corresponding mileage information as the mileage information of the detection starting point.

[0198] Similarly, in the multiple frames of data acquired along the first direction, there is only the profile of the basic rail at the beginning, and the profile of the slide bed plate further appears in the corresponding data frame at the starting position of the slide bed plate. Based on the template, the track profile data frame is searched to see if there is a profile at the position corresponding to the slide bed plate. If there is a profile, the profile of that position is extracted, and the profile line (horizontal straight line segment) of the upper end face of the slide bed plate and the profile line (vertical straight line segment) on the inner side are intercepted. The angle between the two straight line segments is calculated to determine whether the angle is approximately a right angle. If the judgment is yes, the starting point of the slide bed plate is identified, and the corresponding mileage information is obtained as the mileage information of the detection starting point.

[0199] In the subsequent non-reference rail pre-positioning step S4, the mileage range information of the starting points of other non-reference rails is obtained based on the mileage information of the detection starting point.

[0200] In this embodiment, other method steps are the same as those in the first embodiment and will not be described again. In addition, the solution of this embodiment can also be combined with the solution of the second embodiment.

[0201] Functions and effects of embodiment 3

[0202] According to the method and apparatus for automatically identifying and locating turnout rails provided in this embodiment, based on the first embodiment, the slide plate below the identified point rail starting point is further located, and the starting point or centerline position of the slide plate is used as the detection starting point. The position range of other rail components is inferred based on the mileage information of the slide plate starting point. Because the slide plate is fixed to the sleeper, and the position of the sleeper is not easily changed, that is, the starting position of the slide plate is not easily changed, using the slide plate starting point as the detection starting point can further improve the accuracy of the inferred position range of other rail components. Therefore, a relatively smaller mileage range can be set to extract profile data and perform profile matching and identification of the starting points of other rail components, which is conducive to the accurate positioning of other rail components and further reduces the amount of calculation.

[0203] <Example 4>

[0204] This embodiment provides a method and device for automatically identifying and positioning a switch rail. Compared with the first embodiment, the fourth embodiment assigns the same symbols to the components having the same structure as the first embodiment and omits the corresponding descriptions.

[0205] In this embodiment, the reference rail includes a point rail and a center rail.

[0206] Figure 21 It is a flow chart of the steps for positioning the starting point of the reference rail in this embodiment.

[0207] like Figure 21 As shown, in this embodiment, the reference rail starting point positioning step S2 includes the following sub-steps:

[0208] Step S2-0: Set k=1, that is, specify the first frame of contour point data based on mileage information.

[0209] Step S2-1: for the k-th frame of contour point data, based on the reference rail starting point matching template, determine whether there is profile data within a plurality of predetermined position ranges in the frame of contour point data.

[0210] Step S2-2: When the judgment in step S2-1 is no, set k=k+1 and return to step S2-1, that is, specify the next frame of contour point data based on the mileage information.

[0211] Step S2-3: When the judgment in step S2-1 is yes, the silhouette data within the position range where the silhouette data exists is extracted, and the silhouette data is subjected to silhouette clipping based on the corresponding feature silhouette clipping algorithm.

[0212] Step S2-4: Calculate the characteristic data of the corresponding rail member based on the extracted profile and the corresponding characteristic data acquisition algorithm.

[0213] Step S2-5: Determine whether the calculated feature data is consistent with the corresponding predetermined standard feature data, and return to step S2-2 if the determination is no.

[0214] Step S2-6: When the answer in step S2-5 is yes, the type of the reference rail is obtained.

[0215] The profile matching and identification method of the starting point of the point rail is the same as that in the first embodiment and will not be described again.

[0216] Figure 22 Schematic diagram of the profile feature matching of the starting point of the center rail section in this embodiment.

[0217] like Figure 22 As shown, for the starting point of the heart rail section, based on the corresponding heart rail section starting point profile matching template, it is determined whether there is a profile at a predetermined position between the wing rail profiles on both sides in the track profile data. When it is determined that there is a profile, the profile is intercepted to obtain the profile line of the upper end face of the heart rail; the inclination angle and length of the intercepted profile line are calculated as feature data, and it is further determined whether the profile line is a roughly horizontal straight line segment with a predetermined length. When the judgment is yes, the starting point of the heart rail section is identified.

[0218] Step S2-7: Determine whether the type of the reference rail is a center rail.

[0219] Step S2-8: When the judgment in step S2-7 is no, that is, when the starting point of the point rail is detected, the mileage information corresponding to the k-th frame track profile data is used as the mileage information of the starting point of the reference rail (the starting point of the point rail).

[0220] Step S2-9, when the judgment in step S2-7 is yes, that is, when the starting point of the center track section is detected, based on the mileage information corresponding to the k-th frame track profile data and the distance information between the predetermined center track section starting point and the center track theoretical starting point, extract multiple frames of track profile data corresponding to the center track theoretical starting point.

[0221] In step S2-10, the extracted multi-frame track profile data are stacked according to their mileage information to generate a point cloud, and based on the point cloud, the theoretical starting point of the center rail is identified by fitting a plane and the corresponding mileage information is obtained as the mileage information of the starting point of the reference rail (theoretical starting point of the center rail).

[0222] The algorithm for identifying the theoretical starting point of the center track by fitting a plane is the same as that in the first embodiment, and therefore will not be described again.

[0223] In the detection direction determination step S3, the determination rules for the first direction and the second direction are the same as those in the first embodiment. The determination rules for the third direction and the fourth direction are:

[0224] When the identified reference rail is a center rail, and the track profile data of the center rail is detected by the sensor of the left vehicle body unit, it is determined that the current detection direction is the third direction;

[0225] When the identified reference rail is the center rail, and the track profile data detecting the center rail is from the sensor of the right vehicle body unit, it is determined that the current detection direction is the fourth direction.

[0226] In this embodiment, other method steps are the same as those in the first embodiment and will not be described again. In addition, the solution of this embodiment can also be combined with the solution of the third embodiment.

[0227] Functions and effects of embodiment 4

[0228] According to the method and device for automatically identifying and locating turnout rails provided in this embodiment, the reference rails include the point rail and the center rail. Since the wing rails, guard rails, etc. at the switch are all near the center rail, using the center rail as the reference rail is beneficial for improving the identification and positioning accuracy of these turnout rails. In addition, since the starting point of the center rail section is first detected by profile matching, after the starting point of the section is identified, the position range of the theoretical starting point of the center rail is inferred based on the starting point of the section, and only the profile data within this position range is extracted to further fit the plane to calculate the theoretical starting point of the center rail. Therefore, not only can the accurate theoretical starting point position of the center rail and the accurate starting positions of other nearby turnout rails be obtained, but the amount of computation can still be kept small, making this method applicable to on-site testing.

[0229] The above embodiments are only used to illustrate specific implementations of the present invention, and the present invention is not limited to the description scope of the above embodiments.

[0230] In the above embodiment, a single-opening turnout is used as an example for specific description. In fact, it can be understood that other types of turnouts, such as crossover turnouts (also known as diamond turnouts) and compound crossover turnouts, can also similarly adopt variations of the method of the present application to perform multiple rail components and automatic identification and positioning. Taking the crossover turnout as an example, it has four openings and corresponding multiple directions of passage, each of which is a point rail. Similarly, after detecting the starting point of the point rail, the position range of other rail components is inferred and the specific position of other rail components is further detected; and the current detection direction can be automatically determined based on the detected state of the point rail (whether it is in contact or repelled state) and the detected orientation of the guard rail (i.e., which side of the base rail the guard rail is located on).

[0231] In the above steps, for ease of explanation, the step of starting point positioning is performed after the step of turnout data collection. In fact, it can be understood that turnout data collection and starting point positioning of each rail member can also be performed alternately. For example, after starting data collection, the starting point of the reference rail member is detected and the position range of other rail members is inferred. Data collection is continued. After the data of the inferred position range is collected, the specific starting point position of the corresponding rail member is detected. In this way, some calculations can be completed during the data collection process, and the efficiency of on-site detection is higher.

[0232] In the above embodiment, basic track profile data is used to automatically classify and locate rail components. In an alternative solution, the accuracy of automatic classification can be further improved by combining two-dimensional track images. For example, a two-dimensional camera can be further installed on the track inspection vehicle to capture a two-dimensional track image from above. Based on the mileage range information (the inferred starting position of the rail component), the corresponding two-dimensional image is input into a trained neural network model to obtain a rail type classification result. If the classification result does not match the preset rail type at that location, an automatic re-inspection can be performed for verification, and a corresponding warning message can be generated to prompt the inspection personnel.

[0233] In the above embodiments, the point rail, heart rail, wing rail, and guard rail are used as examples for specific description. It can be understood that the above method can also be used to automatically classify and locate other types of rails. For example, non-reference rails can further include straight connecting rails, fork root rails, etc.

Claims

1. A method for automatically identifying and locating a turnout rail member, for automatically identifying and locating a plurality of turnout rail members based on the collected multi-frame track profile data of the turnout and the corresponding mileage information, characterized in that: One of the plurality of rails is a reference rail used as a detection starting point, and the others are non-reference rails. The method includes: Reference rail starting point positioning step: Based on the mileage information and a predetermined reference rail starting point matching template, profile feature matching is performed on each frame of track profile data in sequence to identify the starting point of the reference rail and obtain the corresponding mileage information; a detection direction determination step: determining a current detection direction based on the identified rail type of the reference rail, the track side corresponding to the track profile data of the reference rail, and a predetermined detection direction determination rule, wherein the current detection direction is one of the multiple travel directions of the turnout; A non-reference rail pre-positioning step: based on the mileage information of the starting point of the reference rail and the predetermined distance information between the starting point of the reference rail and the starting point of each non-reference rail, obtaining the mileage range information of the starting point of each non-reference rail, wherein based on the current detection direction, the mileage information of the starting point of the reference rail and the distance information, the rail type of each non-reference rail and the mileage range information of the starting point of each non-reference rail in the current detection direction are obtained; The step of locating the starting point of the non-reference rail is as follows: for each non-reference rail, the corresponding track profile data is extracted based on the mileage range information, and based on the extracted track profile data and the profile characteristics of the starting point of the rail, the starting point of the corresponding rail is identified and the corresponding rail type and mileage information are obtained. The reference rail starting point matching template includes several predetermined position ranges and their corresponding rail types, feature profile interception algorithms, feature data acquisition algorithms, and standard feature data. The reference rail starting point positioning step includes: Extracting the track profile data of each frame in sequence based on the mileage information as current track profile data; determining whether there is profile data within the predetermined position range in the current track profile data, and if so, performing profile interception on the profile data within the predetermined range based on the corresponding feature profile interception algorithm; Calculating characteristic data of the corresponding rail member based on the extracted profile data and the characteristic data acquisition algorithm; Determine whether the characteristic data is consistent with the corresponding standard characteristic data. If it is determined to be consistent, use the corresponding rail type as the rail type of the reference rail, and use the mileage information corresponding to the current track profile data as the mileage information of the starting point of the reference rail.

2. The method for automatically identifying and positioning a turnout rail according to claim 1, characterized in that: in, The reference rail is a pointed rail. The characteristic profile interception algorithm is used to intercept the profile line of the inner rail waist of the point rail and the profile line of the upper end of the rail bottom, The characteristic data acquisition algorithm is used to calculate the angle between the two sections of the profile line as the characteristic data of the starting point of the point rail.

3. The method for automatically identifying and positioning a turnout rail according to claim 1, characterized in that: in, The reference rail is a guard rail. The characteristic profile interception algorithm is used to intercept the profile line of the upper end of the rail head and the profile line of the inner side of the rail head of the guard rail. The characteristic data acquisition algorithm is used to calculate the similarity between the two profile lines and the predetermined guardrail characteristic profile. The characteristic data as the starting point of the guard rail.

4. The method for automatically identifying and positioning a turnout rail according to claim 1, Its characteristics are: Wherein, the non-reference rail comprises at least a center rail and a wing rail, The non-reference rail starting point positioning step includes: extracting, from the multiple frames of track profile data, the multiple frames of track profile data corresponding to the theoretical starting point of the center rail and the multiple frames of track profile data corresponding to the starting point of the wing rail based on the corresponding mileage range information; The track profile data corresponding to the theoretical starting point of the center track is stacked according to the corresponding mileage information to generate a point cloud, and based on the point cloud, the theoretical starting point of the center track is identified by plane fitting and the corresponding mileage information is obtained; Based on the mileage information and a predetermined wing rail starting point matching template, profile matching is performed on multiple frames of track profile data corresponding to the starting point of the wing rail in sequence to identify the starting point of the wing rail and obtain the corresponding mileage information.

5. The method for automatically identifying and positioning a turnout rail according to claim 1, Its characteristics are: Wherein, the non-reference rail comprises at least a center rail and a wing rail, The non-reference rail starting point positioning step includes: Extracting multiple frames of track profile data corresponding to the theoretical starting point of the center track from multiple frames of track profile data based on the corresponding mileage range information; The track profile data corresponding to the theoretical starting point of the center track is stacked according to the corresponding mileage information to generate a point cloud, and based on the point cloud, the theoretical starting point of the center track is identified by plane fitting and the corresponding mileage information is obtained; Obtaining corrected mileage range information of the starting point of the wing rail based on the mileage information of the theoretical starting point of the center rail and predetermined distance information between the theoretical starting point of the center rail and the starting point of the wing rail; Extracting a plurality of frames of track profile data corresponding to the starting point of the wing rail from the plurality of frames of track profile data based on the corrected mileage range information; Based on the mileage information and a predetermined wing rail starting point matching template, profile matching is performed on multiple frames of track profile data corresponding to the starting point of the wing rail in sequence to identify the starting point of the wing rail and obtain the corresponding mileage information.

6. The method for automatically identifying and positioning a turnout rail according to claim 1, characterized in that: The reference rail is a center rail. The characteristic profile interception algorithm is used to intercept the profile line of the upper end surface of the center rail between the cross-section starting point and the theoretical starting point of the center rail. The characteristic data acquisition algorithm is used to calculate the inclination angle and length of the profile line as the characteristic data of the cross-section starting point of the center rail.

7. The method for automatically identifying and positioning a turnout rail according to claim 6, Its characteristics are: Wherein, the reference rail starting point positioning step further includes: Determining whether the identified reference rail is a center rail, and if so, extracting multiple frames of track profile data corresponding to the center rail's theoretical starting point based on the mileage information corresponding to the current track profile data and the distance information from the predetermined starting point of the center rail's cross section to its theoretical starting point; The extracted multiple frames of track profile data are stacked according to their mileage information to generate a point cloud, and based on the point cloud, the theoretical starting point of the center rail is identified by fitting a plane, and the corresponding mileage information is obtained as the mileage information of the starting point of the reference rail.

8. The method for automatically identifying and positioning a turnout rail according to claim 4, 5 or 7, Its characteristics are: The process of identifying the theoretical starting point of the center track by plane fitting includes the following steps: Fitting the slope at the starting point of the center track based on the RANSAC algorithm; Fitting the plane before the starting point of the center track based on the RANSAC algorithm; Calculating the coordinates of the intersection line of the inclined surface and the plane, or calculating the center line of the inclined surface and the center line of the plane respectively, and calculating the coordinates of the intersection point of the two center lines; Based on the coordinates of the intersection line or the coordinates of the intersection point, the corresponding mileage information is obtained as the mileage information of the starting point of the center track.

9. The method for automatically identifying and positioning a turnout rail according to claim 1, Its characteristics are: Wherein, the turnout is a single-opening turnout with four passing directions. The track profile data is obtained through track detection equipment, The track detection device includes a first detection unit and a second detection unit, each used to obtain the track profile data on both sides of the track. The detection direction determination rules include: When the identified rail type of the reference rail is a point rail in a repelling state, and the track profile data of the identified reference rail comes from the first detection unit, determining that the current detection direction is the first direction of going straight through the switch; When the identified reference rail is a closed point rail and the track profile data identifying the reference rail comes from the first detection unit, it is determined that the current detection direction is a second direction from the straight track of the turnout to the curved track; When the identified type of the reference rail is a guard rail, and the track profile data of the identified reference rail comes from the second detection unit, determining that the current detection direction is a third direction opposite to the first direction; When the identified type of the reference rail is a guard rail, and the track profile data of the identified reference rail comes from the first detection unit, it is determined that the current detection direction is a fourth direction opposite to the second direction.

10. A turnout rail automatic identification and positioning device for automatically identifying and positioning multiple rails of a turnout based on collected multi-frame track profile data of the turnout and corresponding mileage information, characterized in that: One of the plurality of rails is a reference rail used as a detection starting point, and the others are non-reference rails. The device includes: A reference rail starting point positioning unit, based on the mileage information and a predetermined reference rail feature matching template, sequentially performs contour feature matching on the track profile data of each frame, identifies the starting point of the reference rail and obtains the corresponding mileage information; a detection direction determination unit, configured to determine a current detection direction based on the identified rail type of the reference rail, the track side corresponding to the track profile data of the reference rail, and a predetermined detection direction determination rule, wherein the current detection direction is one of the plurality of travel directions of the turnout; a non-reference rail pre-positioning unit, which obtains mileage range information of the starting point of each non-reference rail based on the mileage information of the starting point of the reference rail and the predetermined distance information between the starting point of the reference rail and the starting point of each non-reference rail, wherein the rail type of each non-reference rail and the mileage range information of the starting point of each non-reference rail in the current detection direction are obtained based on the current detection direction, the mileage information of the starting point of the reference rail, and the distance information; and The non-reference rail starting point positioning unit extracts the corresponding track profile data for each non-reference rail based on the mileage range information, and identifies the starting point of the corresponding rail based on the extracted track profile data and the profile characteristics of the starting point of the rail, and obtains the corresponding rail type and the mileage information. The reference rail starting point matching template includes several predetermined position ranges and their corresponding rail types, feature profile interception algorithms, feature data acquisition algorithms, and standard feature data. The reference rail starting point positioning unit identifies the starting point of the reference rail and obtains the corresponding mileage information in the following manner: Extracting the track profile data of each frame in sequence based on the mileage information as current track profile data; determining whether there is profile data within the predetermined position range in the current track profile data, and if so, performing profile interception on the profile data within the predetermined range based on the corresponding feature profile interception algorithm; Calculating characteristic data of the corresponding rail member based on the extracted profile data and the characteristic data acquisition algorithm; Determine whether the characteristic data is consistent with the corresponding standard characteristic data. If it is determined to be consistent, use the corresponding rail type as the rail type of the reference rail, and use the mileage information corresponding to the current track profile data as the mileage information of the starting point of the reference rail.

Images