A method and apparatus for measuring the distance between a train undercarriage component and the rail surface.

Abstract

This application proposes a method and apparatus for measuring the distance between a train undercarriage component and the rail surface, belonging to the field of image processing technology. The method includes: acquiring three-dimensional point clouds of a first train guide rail, a second train guide rail, and a target train undercarriage component; calculating the common tangent plane between the three-dimensional point clouds of the first and second train guide rails; calculating the distance from all points in the three-dimensional point cloud of the target train undercarriage component to the common tangent plane; and taking the minimum value among all distances from the common tangent plane as the distance between the train undercarriage component and the rail surface. The apparatus includes: a three-dimensional point cloud acquisition module, a common tangent plane acquisition module, a distance calculation module, and a rail surface distance output module. This application significantly improves the automation rate and measurement stability of measuring the distance between train undercarriage components and the rail surface, reduces the labor intensity of personnel, and significantly improves the measurement accuracy of the distance between train undercarriage components and the rail surface.

Description

Technical Field

[0001] This application relates to the field of image processing technology, specifically to a method and apparatus for measuring the distance between a train undercarriage component and the rail surface. Background Technology

[0002] In existing scenarios for measuring the distance between train undercarriage components and the rail surface, contact measurement methods, such as using height calipers, are often employed. The caliper's base is a long horizontal beam that is mounted on the left and right rail surfaces of the train. The sliding vernier of the caliper contacts the corresponding surface under the train undercarriage, and the distance between the train undercarriage component and the rail surface is then obtained by reading the scale on the caliper. The drawbacks of this method are: 1. The contact between the bottom of the height caliper and the relative rail surface is unstable, resulting in low measurement accuracy; 2. Contact measurement is difficult to perform for components with complex structures. Summary of the Invention Based on the above technical problems, this application proposes a method and device for measuring the distance between train undercarriage components and the rail surface.

[0003] In a first aspect, this application proposes a method for measuring the distance between a train undercarriage component and the rail surface, comprising the following steps: Collect the 3D point cloud of the first guide rail of the train, the 3D point cloud of the second guide rail of the train, and the 3D point cloud of the target component under the train. Calculate the common tangent plane between the 3D point cloud of the first guide rail and the 3D point cloud of the second guide rail of the train; Calculate the distance from all points in the 3D point cloud of the target component under the train to the common tangent plane; The minimum value among the distances from all points to the common tangent plane is taken as the distance between the train undercarriage component and the rail surface.

[0004] If the area corresponding to the three-dimensional point cloud of the first guide rail of the train, the three-dimensional point cloud of the second guide rail of the train, or the three-dimensional point cloud of the target component under the train is less than the set area threshold, then the relative position between the camera used to collect the three-dimensional point cloud and the first guide rail of the train, the second guide rail of the train, or the target component under the train is adjusted, and the corresponding three-dimensional point cloud is collected again.

[0005] Before calculating the common tangent plane between the three-dimensional point cloud of the first train guide rail and the three-dimensional point cloud of the second train guide rail, the method further includes: resampling the three-dimensional point cloud of the first train guide rail, the three-dimensional point cloud of the second train guide rail, and the three-dimensional point cloud of the target component under the train, respectively, in order to reduce the density of the three-dimensional point cloud.

[0006] The resampling was performed using a Gaussian regression algorithm.

[0007] The kernel function of the Gaussian regression algorithm is the radial basis function.

[0008] The collected 3D point clouds of both the first and second train guide rails consist of at least two layers; the calculation of the common tangent plane between the 3D point clouds of the first and second train guide rails includes: Step S102.1: At least two 3D point clouds of both the first and second train guide rails are collected. Two resampled 3D point clouds of the first train guide rail and one resampled 3D point cloud of the second train guide rail are taken, or one resampled 3D point cloud of the first train guide rail and two resampled 3D point clouds of the second train guide rail are taken, resulting in three 3D point clouds. A point is randomly selected from each of the three 3D point clouds, and a plane A is determined based on these three selected points. n ; Step S102.2: Take the plane A n The plane normal vector N n The plane A n The side facing the target component under the train is defined as the plane normal vector N. n The positive direction; Step S102.3: In the three 3D point clouds, in the plane normal vector N n Take the plane A in the positive direction respectively n The point furthest away is used to obtain three new points; Step S102.4: Redetermine plane A based on the three new points. n+1 Take the plane A n+1 The plane normal vector N n+1 The plane A n+1 The side facing the target component under the train is defined as the plane normal vector N. n+1 The positive direction; Step S102.5: Repeat steps S102.3 to S102.4 until the number of points on the positive direction of the normal vector of the currently determined plane in the three 3D point clouds is less than the set number threshold. Then the currently determined plane is the common tangent plane.

[0009] Secondly, this application proposes a measuring device for the distance between a train undercarriage component and the rail surface, comprising: The 3D point cloud acquisition module is used to acquire the 3D point cloud of the first guide rail of the train, the 3D point cloud of the second guide rail of the train, and the 3D point cloud of the target component under the train. The common tangent surface acquisition module is used to calculate the common tangent surface between the three-dimensional point cloud of the first guide rail and the three-dimensional point cloud of the second guide rail of the train. The distance calculation module is used to calculate the distance from all points in the three-dimensional point cloud of the target component under the train to the common tangent plane; The rail surface distance output module takes the minimum value among the distances from all points to the common tangent plane as the distance between the train undercarriage component and the rail surface.

[0010] The measuring device for the distance between the train undercarriage component and the rail surface further includes a resampling module, used to resample the three-dimensional point cloud of the first train guide rail, the three-dimensional point cloud of the second train guide rail, and the three-dimensional point cloud of the target component under the train, respectively, so as to reduce the density of the three-dimensional point cloud.

[0011] Thirdly, this application proposes an electronic device, comprising: one or more processors, and a memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform the method for measuring the distance between a train undercarriage component and the rail surface.

[0012] Fourthly, this application proposes a computer-readable storage medium storing executable instructions that, when executed, cause a processor to perform the method for measuring the distance between a train undercarriage component and the rail surface.

[0013] Beneficial effects: This application proposes a method and apparatus for measuring the distance between a train undercarriage component and the rail surface. The method involves calculating the common tangent plane between the three-dimensional point clouds of the first and second train guide rails; then calculating the distances from all points in the three-dimensional point cloud of the target component under the train to the common tangent plane; finally, the minimum value among the calculated distances is taken as the distance between the train undercarriage component and the rail surface. This method significantly improves the automation rate and stability of measuring the distance between the train undercarriage component and the rail surface, while reducing the labor intensity of personnel. Simultaneously, it greatly improves the measurement accuracy of the distance between the train undercarriage component and the rail surface. Attached Figure Description

[0014] Figure 1 This is a flowchart of the method for measuring the distance between a train undercarriage component and the rail surface according to Embodiment 1 of this application; Figure 2 This is a schematic diagram of the process of acquiring three-dimensional point clouds according to an embodiment of this application; Figure 3 This is a flowchart of the method for measuring the distance between a train undercarriage component and the rail surface according to Embodiment 2 of this application; Figure 4 Point cloud of the locomotive front obstacle clearer in this embodiment of the application; Figure 5 This refers to the curved surface of the traction motor portion under the vehicle in this embodiment of the application; Figure 6 This is a comparison image of the original 3D point cloud and the resampled 3D point cloud of the vehicle undercarriage motor in an embodiment of this application. Figure 7This is a comparison image of the original 3D point cloud and the resampled 3D point cloud of the coupler portion in an embodiment of this application. Figure 8 This is a schematic diagram showing the positive direction of the plane normal vector in an embodiment of this application; Figure 9 This is a schematic block diagram of the measuring device for the distance between the train undercarriage component and the rail surface according to an embodiment of this application; Among them, 1-car bottom inspection composite robot, 2-inspection pit, 3-chassis of the train to be inspected, 4-track, 5-first guide rail, 6-second guide rail, 7-target component under the train. Detailed Implementation

[0015] The present disclosure will be further described below with reference to the embodiments shown in the accompanying drawings.

[0016] Typically, the distances between train undercarriage components and the track surface are large, often reaching the meter level, while the desired measurement accuracy is at or near the sub-millimeter level. Traditional contact measurement methods, such as height calipers, often suffer from cumbersome measurement processes and unstable measurement benchmarks leading to poor accuracy. This application employs a composite robot arm carrying a high-precision structured light depth camera to acquire 3D point clouds. Based on the acquired 3D point clouds, surface filtering fitting is performed using a Gaussian process. By calculating the common tangent plane between the 3D point clouds, high-precision measurement of the distances between train undercarriage components and the track surface over a large scale is achieved.

[0017] Example 1: This embodiment proposes a method for measuring the distance between a train undercarriage component and the rail surface, such as... Figure 1 As shown, it includes the following steps: Step S1: Collect the 3D point cloud of the first guide rail of the train, the 3D point cloud of the second guide rail of the train, and the 3D point cloud of the target component under the train. In some cases, 3D point cloud acquisition can be performed using a vehicle-mounted under-vehicle inspection composite robot with a camera, which can be as follows: Figure 2 As shown, the data acquisition process is as follows: Step a: The undercarriage inspection composite robot 1 first moves to the predetermined measurement position in the maintenance pit 2, where the train to be inspected ( Figure 2 To clearly show its internal structure and maintenance process, only the chassis 3 of the train to be inspected is shown. The chassis 3 moves along the track 4 above the maintenance pit 2 and stops at the predetermined maintenance position. The robotic arm of the undercarriage inspection composite robot, carrying a high-precision structured light depth camera, begins data acquisition of the 3D point cloud. The robotic arm first reaches the designated shooting position on the first guide rail 5 for data acquisition, and then reaches the designated shooting position on the second guide rail 6 for data acquisition, or vice versa. This embodiment does not limit the shooting order. It is understood that if the direction of travel of the train to be inspected is as follows... Figure 2As shown, if the direction of the undercarriage inspection composite robot 1 is such that the first guide rail 5 is the left guide rail of the train to be inspected, and the second guide rail 6 is the right guide rail of the train to be inspected, then... Conversely, if the direction of travel of the train to be inspected is as shown... Figure 2 As shown, the direction is opposite to that of the undercarriage inspection composite robot 1. Then, the first guide rail 5 is the right guide rail of the train to be inspected, and the second guide rail 6 is the left guide rail of the train to be inspected.

[0018] Step b: The robotic arm of the undercarriage inspection composite robot 1 rises up to capture the 3D point cloud of the target component 7 under the train.

[0019] If the area corresponding to the three-dimensional point cloud of the first train guide rail 5, the second train guide rail 6, or the target component under the train is less than the set area threshold, the relative position between the camera used to acquire the three-dimensional point cloud and the first train guide rail 5, the second train guide rail 6, or the target component under the train is adjusted, and the corresponding three-dimensional point cloud is acquired again. By adjusting the relative position between the camera and the subject, the area corresponding to the three-dimensional point cloud of the first train guide rail 5, the second train guide rail 6, and the target component under the train is ensured to be not less than the set area threshold.

[0020] Step c: After the image capture is complete, the robotic arm of the undercarriage inspection composite robot 1 retracts. The currently measured 3D point cloud data is temporarily stored in the edge controller of the undercarriage inspection composite robot. After the current measurement is completely completed, the remote server reads the corresponding data from the edge controller and performs the calculation process of steps S2 to S4.

[0021] Step S2: Calculate the common tangent plane between the 3D point cloud of the first guide rail and the 3D point cloud of the second guide rail of the train; Step S3: Calculate the distance from all points in the 3D point cloud of the target component under the train to the common tangent plane; Step S4: Take the minimum value among the distances from all points to the common tangent plane as the distance between the train undercarriage component and the rail surface.

[0022] The method for measuring the distance between train undercarriage components and the rail surface in this embodiment involves acquiring three-dimensional point clouds of the first and second train guide rails, as well as the three-dimensional point cloud of the target component under the train. It then calculates the common tangent plane between the three-dimensional point clouds of the first and second train guide rails. Finally, it takes the minimum distance from all points in the three-dimensional point cloud of the target component under the train to the common tangent plane as the distance between the train undercarriage component and the rail surface. This method significantly improves the automation rate and stability of measuring the distance between train undercarriage components and the rail surface, while reducing the workload of personnel.

[0023] Example 2: This embodiment proposes a method for measuring the distance between train undercarriage components and the rail surface. In practice, it was found that due to the large number of points in the original three-dimensional point cloud image, the calculation of the common tangent between the three-dimensional point clouds of the first guide rail and the second guide rail of the train is prone to inaccuracy. Therefore, it is necessary to use the Gaussian regression algorithm for resampling to reduce the density of the three-dimensional point cloud. Furthermore, a method for calculating the common tangent is proposed, which significantly improves the measurement accuracy of the distance between train undercarriage components and the rail surface.

[0024] A method for measuring the distance between a train undercarriage component and the rail surface, such as... Figure 3 As shown, it includes the following steps: Step S100: Collect the 3D point cloud of the first guide rail of the train, the 3D point cloud of the second guide rail of the train, and the 3D point cloud of the target component under the train. The process of acquiring 3D point clouds in this embodiment is the same as that in Embodiment 1, so it will not be described again.

[0025] Step S101: Resample the three-dimensional point cloud of the first guide rail of the train, the three-dimensional point cloud of the second guide rail of the train, and the three-dimensional point cloud of the target component under the train to reduce the density of the three-dimensional point cloud. The original 3D point cloud contains outlier noise points and has a high density, requiring resampling to reduce the point cloud density after fitting. Before resampling, the original 3D point cloud needs to be cropped, denoised, and have outlier removed. These basic operations are already mature algorithms or software functions and will not be elaborated upon. These basic operations can be implemented using the PCL (Point Cloud Library) point cloud processing library or the open-source software CloudCompare. The point cloud of the locomotive front obstacle clearer after the above processing is shown below. Figure 4 As shown, the bottom gray 3D point cloud represents the portion of the locomotive's front obstacle clearer closest to the track surface. From... Figure 4 It is not difficult to see that a 3D point cloud is actually a curved surface, even though it appears to be planar. However, fitting it with a planar equation will inevitably lead to a large measurement error. Figure 5 The curved surface of the traction motor section under the vehicle, as shown in Figures 4 and 5, reveals that the geometry of the underbody structural components is quite complex. Therefore... Figure 4 and Figure 5The required kernel function structure scale parameters for fitting also vary. Gaussian process-based regression algorithms are more suitable for fitting curved surfaces of vehicle underbody structural components. For large-scale, relatively flat curved surfaces, the kernel function for Gaussian process regression algorithms can be the Radial Basis Function (RBF), used to fit the obstacle remover surface shown in Figure 4. This also filters out some small-scale structural noise, which may be caused by ambient light interference from the depth camera itself, or by dust or paint cracking noise on the obstacle remover surface. The hyperparameters of the kernel function adapted to the surface patch are calculated through the training process. Since the structure scale of each surface patch is different, a corresponding parameter training process is required for each surface patch to be tested on the vehicle underbody.

[0026] like Figure 6 As shown, the black point cloud is the original 3D point cloud of the undercarriage motor, and its hyperparameters are obtained through training calculations. These parameters are related to the structural curvature of the surface and the scale of change. The gray 3D point cloud is the 3D point cloud obtained after resampling using a Gaussian regression algorithm. Figure 7 The image shows the point cloud of the coupler section, where black represents the original 3D point cloud of the coupler section, and gray represents the 3D point cloud obtained after resampling using a Gaussian regression algorithm.

[0027] Step S102: Calculate the common tangent plane between the three-dimensional point cloud of the first guide rail of the train and the three-dimensional point cloud of the second guide rail of the train after resampling; It is understood that at least two 3D point clouds of the first and second train guide rails are collected. Collecting the 3D point cloud of a specified location on the first train guide rail once yields one 3D point cloud image of the first train guide rail; collecting the 3D point cloud of a specified location on the first train guide rail twice yields two 3D point clouds image images of the first train guide rail. The calculation of the common tangent plane between the 3D point clouds of the first and second train guide rails includes: Step S102.1: Take two resampled 3D point clouds of the first train guide rail and one resampled 3D point cloud of the second train guide rail, or take one resampled 3D point cloud of the first train guide rail and two resampled 3D point clouds of the second train guide rail to obtain three 3D point clouds. Randomly select one point from each of the three 3D point clouds, and determine a plane A based on the three selected points. n ; Step S102.2: Take the plane A n The plane normal vector N n The plane A n The side facing the target component under the train is defined as the plane normal vector N. n The positive direction; Plane A nThe equation can be written as:

[0028] in, For plane A n Any point A0 on the ( x , y , z () is the plane equation A n The variable, ( ) is the plane normal vector N n Wherein, the plane normal vector N n positive direction, such as Figure 8 As shown, plane A n The side facing the target component under the train is defined as the plane normal vector N. n The positive direction, and conversely, the plane A n The side of the target component facing away from the train undercarriage is defined as the plane normal vector N. n The negative direction.

[0029] Step S102.3: In the three 3D point clouds, in the plane normal vector N n Take the plane A in the positive direction respectively n The point furthest away is used to obtain three new points; Step S102.4: Redetermine plane A based on the three new points. n+1 Take the plane A n+1 The plane normal vector N n+1 The plane A n+1 The side facing the target component under the train is defined as the plane normal vector N. n+1 The positive direction; Similarly, the plane A defined by the three newly obtained points n+1 It is also necessary to determine the positive direction of the plane normal vector, which is determined as follows: Figure 8 As shown, the plane A n+1 The side facing the target component under the train is defined as the plane normal vector N. n+1 The positive direction of the plane A n+1 The side of the target component facing away from the train undercarriage is defined as the plane normal vector N. n+1 The negative direction of the plane normal vector can be understood as follows: regardless of how the Cartesian coordinate system is constructed, the method for determining the positive direction of the plane normal vector conforms to the above principle.

[0030] Step S102.5: Repeat steps S102.3 to S102.4 until the number of points on the positive direction of the normal vector of the currently determined plane in the three 3D point clouds is less than the set number threshold. Then the currently determined plane is the common tangent plane.

[0031] The detailed process is as follows: For each 3D point cloud, it can be generated from a plane. It is divided into two parts. As the number of iterations increases, the 3D point cloud in this area is located at the normal vector N. n+1 The number of points on each side will continuously decrease. When the three point clouds are at N... n+1 If the number of points on both sides is less than the set threshold M, then the plane can be considered as a plane. Let n be the common tangent of the three surfaces, where n is the number of iterations and n+1 indicates that the number of iterations increases by 1.

[0032] Step S103: Calculate the distance from all points in the 3D point cloud of the target component under the train undercarriage to the common tangent plane after resampling; Step S104: Take the minimum value among the distances from all points to the common tangent plane as the distance between the train undercarriage component and the rail surface.

[0033] Understandably, once the common tangent plane is determined, it is necessary to calculate the distance from all points in the 3D point cloud of the target component under the train to the common tangent plane after resampling. The minimum value among all the distances from the common tangent plane is taken as the distance between the train undercarriage component and the rail surface, and the measurement of the distance between the train undercarriage component and the rail surface is completed.

[0034] The method for measuring the distance between the train undercarriage component and the rail surface in this embodiment addresses the issue of inaccurate calculation of the common tangent plane between the three-dimensional point cloud of the first guide rail and the three-dimensional point cloud of the second train guide rail. On one hand, it employs a Gaussian regression algorithm for resampling to reduce the density of the three-dimensional point cloud; on the other hand, it uses a point-based planar iteration method to calculate the common tangent plane, taking the minimum distance to the common tangent plane as the distance between the train undercarriage component and the rail surface. This method significantly improves the measurement accuracy of the distance between the train undercarriage component and the rail surface.

[0035] Example 3: This embodiment proposes a device for measuring the distance between a train undercarriage component and the rail surface, such as... Figure 9 As shown, it includes: a 3D point cloud acquisition module, a common tangent surface acquisition module, a distance calculation module, and a track surface distance output module; The 3D point cloud acquisition module is connected to the common tangent surface acquisition module, the common tangent surface acquisition module is connected to the distance calculation module, and the distance calculation module is connected to the track surface distance output module. The 3D point cloud acquisition module is used to acquire the 3D point cloud of the first guide rail of the train, the 3D point cloud of the second guide rail of the train, and the 3D point cloud of the target component under the train. The common tangent surface acquisition module is used to calculate the common tangent surface between the three-dimensional point cloud of the first guide rail and the three-dimensional point cloud of the second guide rail of the train. The distance calculation module is used to calculate the distance from all points in the three-dimensional point cloud of the target component under the train to the common tangent plane; The rail surface distance output module takes the minimum value among the distances from all points to the common tangent plane as the distance between the train undercarriage component and the rail surface.

[0036] The measuring device for the distance between the train undercarriage component and the rail surface further includes: a resampling module, used to resample the three-dimensional point cloud of the first train guide rail, the three-dimensional point cloud of the second train guide rail, and the three-dimensional point cloud of the target component under the train, respectively, in order to reduce the density of the three-dimensional point cloud. The three-dimensional point cloud acquisition module is connected to the resampling module, and the resampling module is connected to the common tangent surface acquisition module.

[0037] The device for measuring the distance between a train undercarriage component and the rail surface proposed in this embodiment employs a 3D point cloud acquisition module to acquire 3D point clouds of the first and second train guide rails, as well as the 3D point cloud of the target component under the train. A resampling module resamples these 3D point clouds, and a common tangent plane acquisition module calculates the common tangent plane between the 3D point clouds of the first and second train guide rails. A distance calculation module then calculates the distance from all points in the 3D point cloud of the target component under the train to the common tangent plane. Finally, a rail surface distance output module uses the minimum value among all distances from the common tangent plane as the distance between the train undercarriage component and the rail surface. This device significantly improves the automation rate and stability of measuring the distance between the train undercarriage component and the rail surface, while reducing the labor intensity of personnel. It also significantly improves the measurement accuracy of the distance between the train undercarriage component and the rail surface.

[0038] Example 4: This embodiment proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the method for measuring the distance between the train undercarriage component and the rail surface.

[0039] Example 5: This embodiment proposes an electronic device, including: one or more processors, and a memory, wherein the memory stores instructions, and when the instructions are executed by the one or more processors, the one or more processors perform the method for measuring the distance between a train undercarriage component and the rail surface.

[0040] The electronic device may be a mobile phone, computer, or tablet computer, etc., and includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements the method for measuring the distance between the train undercarriage components and the rail surface as described in the embodiment. It is understood that the electronic device may also include an input / output (I / O) interface and communication components.

[0041] The processor is used to execute all or part of the steps in the method for measuring the distance between the train undercarriage component and the rail surface as described in the above embodiments. The memory is used to store various types of data, which may include, for example, instructions for any application or method in the electronic device, as well as application-related data.

[0042] The processor can be implemented as an Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor, or other electronic components, and is used to execute the method for measuring the distance between the train undercarriage components and the rail surface as described in the above embodiments.

[0043] Example 6: This embodiment proposes a computer-readable storage medium storing executable instructions, which, when executed, cause a processor to perform the method for measuring the distance between a train undercarriage component and the rail surface.

[0044] In the various embodiments of the present invention, the functional units can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0045] Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method for measuring the distance between the train undercarriage component and the rail surface described in the various embodiments of this application.

[0046] The aforementioned storage media include: flash memory, hard disk, multimedia card, card-type memory (e.g., SD (Secure Digital Memory Card) or DX (Memory Data Register, MDR) memory, random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, disk, optical disk, server, APP (Application) application store, and other media capable of storing program verification codes. These media store computer programs, and when executed by a processor, they can implement the various steps of the aforementioned method for measuring the distance between the train undercarriage components and the rail surface.

[0047] The various embodiments in this disclosure are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0048] The scope of protection of this disclosure is not limited to the embodiments described above. Obviously, those skilled in the art can make various modifications and variations to this disclosure without departing from its scope and spirit. If such modifications and variations fall within the scope of the claims of this disclosure and their equivalents, then the intent of this disclosure also includes such modifications and variations.

Claims

1. A method for measuring the distance between a train undercarriage component and the rail surface, characterized in that, Includes the following steps: Collect the 3D point cloud of the first guide rail of the train, the 3D point cloud of the second guide rail of the train, and the 3D point cloud of the target component under the train. The three-dimensional point cloud of the first guide rail of the train, the three-dimensional point cloud of the second guide rail of the train, and the three-dimensional point cloud of the target component under the train are resampled to reduce the density of the three-dimensional point cloud. Calculate the common tangent plane between the 3D point cloud of the first guide rail and the 3D point cloud of the second guide rail of the train; Calculate the distance from all points in the 3D point cloud of the target component under the train to the common tangent plane; The minimum value among the distances from all points to the common tangent plane is taken as the distance between the train undercarriage component and the rail surface; 3D point cloud acquisition is performed using a vehicle under-vehicle inspection composite robot carrying a camera. The acquisition process is as follows: Step a: The undercarriage inspection composite robot first moves to the predetermined measurement position in the maintenance pit. The train to be inspected runs along the track above the maintenance pit and stops at the predetermined maintenance position. The robotic arm of the undercarriage inspection composite robot, carrying a structured light depth camera, begins to collect 3D point cloud data. The robotic arm first reaches the designated shooting position on the first guide rail for data collection, and then reaches the designated shooting position on the second guide rail for data collection. Step b: The robotic arm of the undercarriage inspection composite robot rises up to capture the 3D point cloud of the target component under the train. If the acquired 3D point cloud of the first guide rail, the second guide rail, or the target component under the train is within the specified range, the data will be collected. If the area corresponding to the 3D point cloud of the target component is less than the set area threshold, the relative position between the camera used to acquire the 3D point cloud and the first guide rail, the second guide rail, or the target component under the train is adjusted, and the corresponding 3D point cloud is acquired again. By adjusting the relative position between the camera and the object being photographed, the areas corresponding to the 3D point clouds of the first guide rail, the second guide rail, and the target component under the train are all not less than the set area threshold. Step c: After the shooting is completed, the robotic arm of the undercarriage inspection composite robot is retracted, and the currently measured 3D point cloud data is temporarily stored in the edge controller of the undercarriage inspection composite robot. The collected 3D point clouds of both the first and second train guide rails consist of at least two layers; the calculation of the common tangent plane between the 3D point clouds of the first and second train guide rails includes: Step S102.1: Take two resampled 3D point clouds of the first train guide rail and one resampled 3D point cloud of the second train guide rail, or take one resampled 3D point cloud of the first train guide rail and two resampled 3D point clouds of the second train guide rail to obtain three 3D point clouds. Randomly select one point from each of the three 3D point clouds, and determine a plane A based on the three selected points. n ; Step S102.2: Take the plane A n The plane normal vector N n The plane A n The side facing the target component under the train is defined as the plane normal vector N. n The positive direction; Step S102.3: In the three 3D point clouds, in the plane normal vector N n Take the plane A in the positive direction respectively n The point furthest away is used to obtain three new points; Step S102.4: Redetermine plane A based on the three new points. n+1 Take the plane A n+1 The plane normal vector N n+1 The plane A n+1 The side facing the target component under the train is defined as the plane normal vector N. n+1 The positive direction; Step S102.5: Repeat steps S102.3 to S102.4 until the number of points on the positive direction of the normal vector of the currently determined plane in the three 3D point clouds is less than the set number threshold. Then the currently determined plane is the common tangent plane.

2. The method for measuring the distance between a train undercarriage component and the rail surface according to claim 1, characterized in that, The resampling was performed using a Gaussian regression algorithm.

3. The method for measuring the distance between a train undercarriage component and the rail surface according to claim 2, characterized in that, The kernel function of the Gaussian regression algorithm is the radial basis function.

4. A measuring device for the distance between a train undercarriage component and the rail surface, characterized in that, include: The 3D point cloud acquisition module is used to acquire the 3D point cloud of the first guide rail of the train, the 3D point cloud of the second guide rail of the train, and the 3D point cloud of the target component under the train. The resampling module is used to resample the three-dimensional point cloud of the first guide rail of the train, the three-dimensional point cloud of the second guide rail of the train, and the three-dimensional point cloud of the target component under the train, so as to reduce the density of the three-dimensional point cloud. The common tangent surface acquisition module is used to calculate the common tangent surface between the three-dimensional point cloud of the first guide rail and the three-dimensional point cloud of the second guide rail of the train. The distance calculation module is used to calculate the distance from all points in the three-dimensional point cloud of the target component under the train to the common tangent plane; The rail surface distance output module takes the minimum value among the distances from all points to the common tangent plane as the distance between the train undercarriage component and the rail surface; 3D point cloud acquisition is performed using a composite robot for undercarriage inspection, equipped with a camera. The acquisition process is as follows: The composite robot first moves to the predetermined measurement position in the maintenance pit. The train to be inspected travels along the track above the maintenance pit and stops at the predetermined position. The robotic arm of the composite robot, carrying a structured light depth camera, begins data acquisition of the 3D point cloud. The robotic arm first reaches the designated shooting position on the first guide rail for data acquisition, and then reaches the designated shooting position on the second guide rail for data acquisition. The robotic arm of the composite robot then rises to capture the 3D point cloud of the target component under the train. If the acquired 3D point cloud of the first guide rail or the second guide rail is within the target area, the data acquisition is performed. If the area corresponding to the 3D point cloud of the guide rail or the 3D point cloud of the target component under the train is less than the set area threshold, the relative position between the camera used to acquire the 3D point cloud and the first or second guide rail of the train or the target component under the train is adjusted, and the corresponding 3D point cloud is acquired again. By adjusting the relative position between the camera and the object being photographed, the areas corresponding to the 3D point clouds of the first guide rail, the second guide rail, and the target component under the train are all guaranteed to be not less than the set area threshold. After the photographing is completed, the robotic arm of the undercarriage inspection composite robot is retracted, and the currently measured 3D point cloud data is temporarily stored in the edge controller of the undercarriage inspection composite robot. The collected 3D point clouds of both the first and second train guide rails consist of at least two layers; the calculation of the common tangent plane between the 3D point clouds of the first and second train guide rails includes: Step S102.1: Take two resampled 3D point clouds of the first train guide rail and one resampled 3D point cloud of the second train guide rail, or take one resampled 3D point cloud of the first train guide rail and two resampled 3D point clouds of the second train guide rail to obtain three 3D point clouds. Randomly select one point from each of the three 3D point clouds, and determine a plane A based on the three selected points. n ; Step S102.2: Take the plane A n The plane normal vector N n The plane A n The side facing the target component under the train is defined as the plane normal vector N. n The positive direction; Step S102.3: In the three 3D point clouds, in the plane normal vector N n Take the plane A in the positive direction respectively n The point furthest away is used to obtain three new points; Step S102.4: Redetermine plane A based on the three new points. n+1 Take the plane A n+1 The plane normal vector N n+1 The plane A n+1 The side facing the target component under the train is defined as the plane normal vector N. n+1 The positive direction; Step S102.5: Repeat steps S102.3 to S102.4 until the number of points on the positive direction of the normal vector of the currently determined plane in the three 3D point clouds is less than the set number threshold. Then the currently determined plane is the common tangent plane.

5. An electronic device, characterized in that, include: One or more processors, and a memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform the method for measuring the distance between a train undercarriage component and the rail surface as described in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed, cause the processor to perform the method for measuring the distance between a train undercarriage component and the rail surface as described in any one of claims 1-3.

Images