Locomotive underframe detection and positioning method and device

Abstract

The application provides a locomotive underbody detection positioning method and device, and belongs to the technical field of image processing. The method comprises the following steps: obtaining a three-dimensional point cloud of a target detection component of a locomotive underbody to be detected collected by a camera at a composite robot body, to obtain a first three-dimensional point cloud; registering the first three-dimensional point cloud with a first preset point cloud template, to obtain a pose transformation matrix with a correction offset; and moving the composite robot to a position where the first three-dimensional point cloud coincides with the first preset point cloud template according to the correction offset of the pose transformation matrix, so as to perform underbody detection of the locomotive to be detected at the position where the first three-dimensional point cloud coincides with the first preset point cloud template. The device comprises a first point cloud acquisition module, a first point cloud registration module and a first positioning adjustment module. The application can greatly reduce the customization demand and reconstruction engineering quantity of the composite robot in a pit environment, and improve the positioning ability of the composite robot for special target detection components.

Description

Technical Field

[0001] This application belongs to the field of image processing technology, specifically relating to a method and apparatus for detecting and locating the underside of a locomotive. Background Technology

[0002] In existing technical solutions for vehicle undercarriage detection and positioning, one type of solution relies on the train's excellent positioning capabilities, such as subway cars. Their positioning is relatively accurate upon entering the depot for maintenance, and the vehicle types are relatively uniform. The combined inspection and maintenance robot only needs to be positioned at fixed coordinates on the map. Once the robot chassis reaches the fixed station, the robotic arm moves to the pre-set pose and performs image acquisition. Because the train itself has good positioning upon entering the depot, images acquired at any two different times will be essentially consistent for the same detection target. Through basic image matching and image differencing, positional changes or component defects of the target object can be detected. However, most locomotives and rolling stock maintenance vehicles lack accurate positioning capabilities, and their stopping positions within the depot are often quite arbitrary. Therefore, this type of vehicle undercarriage detection and positioning solution lacks universality.

[0003] Another type of vehicle undercarriage detection and positioning solution requires the wheelsets of a composite robot chassis to be mounted on steel rails. The composite robot chassis only performs one-dimensional movement, thus simplifying the navigation control dimension. The longitudinal contour of the vehicle undercarriage is obtained using line laser scanning, and the composite robot's position under the vehicle is determined by comparison. After reaching the positioning position, the robotic arm moves to the target pose and takes an image. This type of vehicle undercarriage detection and positioning solution requires laying steel rails in a pit, which inevitably leads to inconvenience for manual inspection. Summary of the Invention

[0004] Based on the above technical problems, this application proposes a locomotive undercarriage detection and positioning method and device.

[0005] Firstly, this application proposes a method for detecting and locating the undercarriage of a locomotive, comprising the following steps:

[0006] When the locomotive to be inspected arrives at the inspection position, the three-dimensional point cloud of the target inspection component under the locomotive to be inspected is obtained by the camera at the body of the composite robot, and the first three-dimensional point cloud is obtained.

[0007] The first three-dimensional point cloud is registered with the first preset point cloud template to obtain a pose transformation matrix with a correction offset. The first preset point cloud template is a three-dimensional point cloud containing the pose relationship between the composite robot body and the target detection component.

[0008] Based on the correction offset of the pose transformation matrix, the composite robot is moved to a position where the first three-dimensional point cloud and the first preset point cloud template coincide, so as to perform undercarriage detection of the vehicle to be inspected at the position where the first three-dimensional point cloud and the first preset point cloud template coincide.

[0009] The process of obtaining the first preset point cloud template includes:

[0010] When the locomotive to be inspected arrives at the inspection position, the activity space of the robotic arm of the composite robot and the shooting posture are adjusted to meet their respective preset conditions, and the position of the composite robot in the above state is determined as the measurement point position of the composite robot.

[0011] The first preset point cloud template is obtained by acquiring the three-dimensional point cloud of the target component under the vehicle body of the vehicle to be inspected, which is captured by the camera at the vehicle body of the composite robot.

[0012] The step of acquiring the three-dimensional point cloud of the target detection component under the locomotive to be inspected, collected by a camera at the body of the composite robot, to obtain the first three-dimensional point cloud includes:

[0013] When the composite robot reaches the measurement point, the three-dimensional point cloud of the target detection component under the locomotive to be inspected is obtained by the camera at the body of the composite robot.

[0014] The pose transformation matrix with corrected offset is as follows:

[0015]

[0016] Wherein, Rot is the rotation matrix obtained by rotating the first 3D point cloud to the first preset point cloud template, and Tran is the translation vector obtained by translating the rotated 3D point cloud to the first preset point cloud template, the translation vector including a correction offset, T A2B This is the pose transformation matrix with corrected offset.

[0017] If, based on the correction offset of the pose transformation matrix, the composite robot cannot be moved to a position where the first 3D point cloud coincides with the first preset point cloud template, the locomotive undercarriage detection and positioning method further includes:

[0018] When the composite robot reaches the measurement point, the three-dimensional point cloud of the target detection component under the locomotive to be inspected is acquired by the camera at the robotic arm of the composite robot, and a second three-dimensional point cloud is obtained.

[0019] The second three-dimensional point cloud is registered with the second preset point cloud template to obtain the corrected pose of the robotic arm. The second preset point cloud template is a three-dimensional point cloud containing the pose relationship between the robotic arm of the composite robot and the target detection component.

[0020] Based on the corrected pose of the robotic arm, the composite robot is moved to a position where the second three-dimensional point cloud and the second preset point cloud template coincide, so as to perform undercarriage inspection of the vehicle to be inspected at the position where the second three-dimensional point cloud and the second preset point cloud template coincide.

[0021] The process of obtaining the second preset point cloud template includes:

[0022] When the locomotive to be inspected arrives at the inspection position, the activity space of the robotic arm of the composite robot and the shooting posture are adjusted to meet their respective preset conditions, and the position of the composite robot in the above state is determined as the measurement point position of the composite robot.

[0023] A second preset point cloud template is obtained by acquiring the three-dimensional point cloud of the target detection component under the locomotive to be inspected, which is captured by the camera at the robotic arm of the composite robot.

[0024] The corrected pose of the robotic arm is calculated using the following formula:

[0025]

[0026] in, To correct the pose of the robotic arm; This is the transformation matrix between the camera coordinate system of the camera on the vehicle body and the end effector coordinate system of the robotic arm; The pose of the robotic arm of the composite robot when obtaining the second preset point cloud template; The pose of the robotic arm of the composite robot when acquiring the second 3D point cloud.

[0027] Secondly, this application proposes a locomotive undercarriage detection and positioning device, comprising:

[0028] The first point cloud acquisition module is used to acquire the three-dimensional point cloud of the target detection component under the locomotive under the locomotive under the inspection, which is collected by the camera at the body of the composite robot when the locomotive under inspection arrives at the inspection position, and obtain the first three-dimensional point cloud.

[0029] The first point cloud registration module is used to register the first three-dimensional point cloud with the first preset point cloud template to obtain a pose transformation matrix with a correction offset. The first preset point cloud template is a three-dimensional point cloud containing the pose relationship between the composite robot body and the target detection component.

[0030] The first positioning adjustment module is used to move the composite robot to a position where the first three-dimensional point cloud and the first preset point cloud template coincide, based on the correction offset of the pose transformation matrix, so as to perform undercarriage detection of the vehicle to be inspected at the position.

[0031] Thirdly, this application proposes an electronic device, including: one or more processors, and a memory, the memory storing instructions, which, when executed by the one or more processors, cause the one or more processors to perform the locomotive undercarriage detection and positioning method.

[0032] Fourthly, this application proposes a computer-readable storage medium storing executable instructions that, when executed, cause a processor to perform the locomotive undercarriage detection and positioning method.

[0033] Beneficial effects:

[0034] This application proposes a method and apparatus for locomotive undercarriage inspection and positioning. The method involves registering the 3D point cloud to be inspected with a pre-defined template point cloud to obtain a pose transformation matrix with a correction offset. Based on the correction offset of the pose transformation matrix, a composite robot is moved to a position where the 3D point cloud coincides with the pre-defined template point cloud, enabling undercarriage inspection of the locomotive at this position. This method significantly reduces the customization requirements and modification workload of the composite robot's pit environment and improves the relative positioning capability of the composite robot for special target inspection components. Attached Figure Description

[0035] Figure 1 This is a flowchart of the locomotive undercarriage detection and positioning method in this embodiment;

[0036] Figure 2 This is a schematic diagram of the three-dimensional point cloud of the target detection component acquired by the composite robot in this embodiment;

[0037] Figure 3 This is a schematic diagram of the various parts of the composite robot in this embodiment;

[0038] Figure 4 This is a schematic diagram of a local area of ​​the target detection part in this embodiment;

[0039] Figure 5 This is a schematic diagram of the locomotive undercarriage detection and positioning device in this embodiment;

[0040] Among them, 1-maintenance pit, 2-composite robot, 3-track, 4-carriage to be inspected, 21-chassis of composite robot, 22-camera at the vehicle body, 23-robotic arm, 24-camera at the robotic arm. Detailed Implementation

[0041] The present disclosure will be further described below with reference to the embodiments shown in the accompanying drawings.

[0042] In existing technical solutions for vehicle undercarriage detection and positioning, one type of solution relies on the train's excellent positioning capabilities, such as subway cars. Their positioning is relatively accurate upon entering the depot for maintenance, and the vehicle types are relatively uniform. The combined inspection and maintenance robot only needs to be positioned at fixed coordinates on the map. Once the robot chassis reaches the fixed station, the robotic arm moves to the pre-set pose and performs image acquisition. Because the train itself has good positioning upon entering the depot, images acquired at any two different times will be essentially consistent for the same detection target. Through basic image matching and image differencing, positional changes or component defects of the target object can be detected. However, most locomotives and rolling stock maintenance vehicles lack accurate positioning capabilities, and their stopping positions within the depot are often quite arbitrary. Therefore, this type of vehicle undercarriage detection and positioning solution lacks universality.

[0043] Another type of vehicle undercarriage detection and positioning solution requires the wheelsets of a composite robot chassis to be mounted on steel rails. The composite robot chassis only performs one-dimensional motion, thus simplifying the navigation control dimension. The longitudinal contour of the vehicle undercarriage is obtained using line laser scanning, and the composite robot's position under the vehicle is determined by comparison. After reaching the positioning position, the robotic arm moves to the target pose and takes an image. This type of vehicle undercarriage detection and positioning solution requires laying steel rails in a pit, which inevitably leads to inconvenience for manual inspection.

[0044] This application proposes a method and apparatus for detecting and locating the undercarriage of a locomotive. By using a registration technique between a 3D point cloud acquired by a depth camera and a preset point cloud template, the pose relationship between the end effector of a composite robot and the target component under the vehicle is obtained. Specifically, the pose relationship between the detection equipment fixedly mounted at the end effector and the target component under the vehicle is determined. This drives the chassis or joints of the composite robot to approximate the human-defined target pose. The method significantly reduces the customization requirements and modification workload of the composite robot's pit environment and improves the relative positioning capability of the composite robot for detecting special target components.

[0045] Example 1:

[0046] This embodiment proposes a method for detecting and locating the undercarriage of a locomotive, such as... Figure 1 As shown, it includes:

[0047] Step S1: When the locomotive to be inspected arrives at the inspection position, acquire the three-dimensional point cloud of the target inspection component under the locomotive to be inspected, which is collected by the camera at the body of the composite robot, and obtain the first three-dimensional point cloud.

[0048] In this embodiment, the locomotive tractor pulls the locomotive 4 to be inspected to the inspection position in the maintenance pit 1, such as... Figure 2As shown, because the locomotive tractor is manually controlled, the positioning of the locomotive 4 under inspection is inaccurate, with a front-to-back error of approximately ±500mm. Since the locomotive 4 moves on track 3, the undercarriage inspection robot 2, used to collect the 3D point cloud of the target inspection components under the locomotive 4, will only have positioning errors in the locomotive's traveling direction, with smaller offsets in the lateral direction. Figure 2 As shown, the composite robot 2, which is used for undercarriage inspection, uses a depth camera at the vehicle body to obtain the offset of the 3D point cloud of the key structural components under the vehicle body and the standard template point cloud in the vehicle body coordinate system. This corrects the front-to-back and lateral offset of the vehicle 4 to be inspected, ensuring the accuracy of the relative position between the chassis of the composite robot and the undercarriage of the vehicle 4 to be inspected.

[0049] In this embodiment, the chassis of the composite robot 2 is navigated using navigation software. The navigation method employs laser navigation with reflective pillars, and the body of the composite robot 2 is equipped with a depth camera to capture images of the underside of the vehicle under inspection. The navigation software includes a site position offset correction interface.

[0050] Step S2: Register the first 3D point cloud with the first preset point cloud template to obtain a pose transformation matrix with corrected offset. The first preset point cloud template is a 3D point cloud containing the pose relationship between the composite robot body and the target detection component.

[0051] In practice, it is necessary to pre-capture and save a first preset point cloud template. The process of acquiring the first preset point cloud template includes:

[0052] Step a: When the vehicle to be inspected arrives at the inspection position, adjust the activity space of the robotic arm of the composite robot and the shooting pose to meet their respective preset conditions, and determine the position of the composite robot in the state as the measurement point position of the composite robot.

[0053] Step b: Obtain the 3D point cloud of the target detection component under the vehicle body of the vehicle to be inspected, which is captured by the camera at the vehicle body of the composite robot, and obtain the first preset point cloud template.

[0054] like Figure 2 As shown, after the vehicle to be inspected 4 reaches the inspection position, it is necessary to check whether the working space and shooting posture of the robotic arm of the composite robot 2 meet their respective preset conditions. If not, corresponding adjustments need to be made until the working space and shooting posture of the robotic arm of the composite robot 2 both meet their respective preset conditions. Record the coordinate position of the composite robot 2 in the vehicle coordinate system at this time. The vehicle coordinate system is as follows: Figure 3As shown, the vehicle body coordinate system is a three-dimensional Cartesian coordinate system. The origin of the vehicle body coordinate system is on the vehicle body cover. The positive direction of the x-axis is the forward direction of the vehicle body, the positive direction of the z-axis is perpendicular to the x-axis and points towards the sky, and the positive direction of the y-axis is determined by the right-hand rule, with the thumb pointing in the positive direction of the x-axis, the middle finger pointing in the positive direction of the z-axis, and the index finger pointing in the positive direction of the y-axis. Three-dimensional point cloud data is acquired through camera 22 at the vehicle body. After appropriate trimming, filtering, and downsampling, it is stored as a first preset point cloud template (also called a reference point cloud). The first preset point cloud template is the three-dimensional point cloud acquired by the camera at the composite robot body, representing the pose relationship between the composite robot body and the target detection component under standard conditions.

[0055] Having already obtained the first preset point cloud template, the first 3D point cloud is registered with the first preset point cloud template. The registration process is as follows:

[0056] Step S2.1: Calculate the PPF (Point Pair Feature) features of the first preset point cloud template A according to the specified step size, obtain the first point pair feature set, and construct the first hash table for the first point pair feature set;

[0057] Step S2.2: Calculate the PPF features of the first 3D point cloud A' according to the specified step size to obtain the second point pair feature set. Calculate the corresponding hash value of each point pair in the second point pair feature set. According to the first hash table, find the point pair feature corresponding to the hash value. Each point pair in the second point pair feature set and the point pair feature corresponding to the hash value are regarded as a feature pair.

[0058] Step S2.4: Calculate the pose transformation matrix between the first preset point cloud template A and the first 3D point cloud A' for each feature pair;

[0059] Step S2.5: Cluster all pose transformation matrices and sort them according to the number of feature pairs they belong to, to obtain the top N candidate poses to be determined;

[0060] In this embodiment, N = 8. It can be understood that N can be selected with other values ​​depending on the actual situation.

[0061] Step S 2.6: Using the N candidate poses to be determined as the initial poses, perform M registration calculations and obtain the score for each fine registration. The score represents the degree of closeness between the first preset point cloud template A and the first 3D point cloud A'. The pose transformation matrix corresponding to the highest score is determined as the optimal pose transformation matrix.

[0062] In this embodiment, M = 8. It can be understood that M can be selected with other values ​​depending on the actual situation.

[0063] The process involves M registration operations using the N candidate poses as initial poses. The registration method can be either coarse or fine registration. The coarse registration method uses the `match` function of the `PPF3DDetector` class within the `surface match` module of OpenCV (OpenCV is a cross-platform computer vision and machine learning software library released under the Apache 2.0 license). The fine registration method uses the relevant functions of the `ICP` class within the `surface match` module of OpenCV.

[0064] Step S3: Based on the correction offset of the pose transformation matrix, move the composite robot to the position where the first three-dimensional point cloud and the first preset point cloud template coincide, so as to perform the undercarriage detection of the vehicle to be detected at the position where the first three-dimensional point cloud and the first preset point cloud template coincide.

[0065] The pose transformation matrix with corrected offset is as follows:

[0066]

[0067] Wherein, Rot is the rotation matrix obtained by rotating the first 3D point cloud to the first preset point cloud template, and Tran is the translation vector obtained by translating the rotated 3D point cloud to the first preset point cloud template, the translation vector including a correction offset, T A2B This is the pose transformation matrix with corrected offset.

[0068] Under normal circumstances, the above process can achieve the purpose of detecting and locating the undercarriage of the locomotive. However, if the pose of the composite robot cannot be corrected, or if the structure of the target detection component has certain dimensional changes (the installation position is not completely determined according to the drawings), such as pipe joints or flexible slings of components, then the robotic arm needs to make small-range pose adjustments based on the current state of the target detection component. The adjustment process is as follows:

[0069] Step S4: When the composite robot reaches the measurement point position, acquire the three-dimensional point cloud of the target detection component under the locomotive to be inspected, which is collected by the camera at the robotic arm of the composite robot, to obtain the second three-dimensional point cloud.

[0070] Step S5: Register the second three-dimensional point cloud with the second preset point cloud template to obtain the corrected pose of the robotic arm. The second preset point cloud template is a three-dimensional point cloud containing the pose relationship between the robotic arm of the composite robot and the target detection component.

[0071] Step S6: Based on the corrected pose of the robotic arm, move the composite robot to the position where the second three-dimensional point cloud and the second preset point cloud template coincide, so as to perform undercarriage inspection of the vehicle to be inspected at the position where the second three-dimensional point cloud and the second preset point cloud template coincide.

[0072] The process of obtaining the second preset point cloud template includes:

[0073] Step c: When the vehicle to be inspected arrives at the inspection position, adjust the activity space of the robotic arm of the composite robot and the shooting pose to meet their respective preset conditions, and determine the position of the composite robot in the state as the measurement point position of the composite robot.

[0074] Step d: Obtain the 3D point cloud of the target detection component under the locomotive to be inspected, captured by the camera at the robotic arm of the composite robot, to obtain the second preset point cloud template.

[0075] The processes for steps c and d can be referenced from those for steps a and b, and will not be repeated here.

[0076] The corrected pose of the robotic arm is calculated using the following formula:

[0077]

[0078] in, To correct the pose of the robotic arm; This is the transformation matrix between the camera coordinate system of the camera on the vehicle body and the end effector coordinate system of the robotic arm; The pose of the robotic arm of the composite robot when obtaining the second preset point cloud template; The pose of the composite robot's arm is determined to acquire the second 3D point cloud. The camera coordinate system is a Cartesian coordinate system fixed to the camera body. Its specific location is determined by the camera manufacturer. The end effector coordinate system is fixed to the end effector joint, i.e., the location where the camera is mounted, and is by default at the center of the end effector joint. Its specific orientation and location are determined by the robot manufacturer.

[0079] A detailed explanation of the above process is provided, such as... Figure 4 As shown, the local region A of the target detection component represents its position in the vehicle coordinate system during measurement and input. This can also be understood as the position of the camera at the robotic arm in the camera coordinate system during input. When the position of local region A of the target detection component changes in the vehicle coordinate system, its position in the camera coordinate system at the robotic arm shifts to position A”. At this point, the original robotic arm pose needs to be corrected. The corrected homogeneous matrix of the robotic arm end-effector pose is... This ensures that the 3D point cloud obtained from the second photograph coincides with the second preset point cloud template.

[0080] This embodiment proposes a locomotive undercarriage detection and positioning method. It acquires a 3D point cloud of the target detection component on the undercarriage of the locomotive to be inspected, captured by a camera on the body of a composite robot, thus obtaining a first 3D point cloud. The first 3D point cloud is then registered with a first preset point cloud template to obtain a pose transformation matrix with a correction offset. Based on the correction offset of the pose transformation matrix, the composite robot is moved to a position where the first 3D point cloud and the first preset point cloud template coincide, so that the undercarriage of the locomotive to be inspected can be detected at this position. When the composite robot cannot be moved to a position where the first 3D point cloud coincides with the first preset point cloud template, a second 3D point cloud is obtained by acquiring the target detection component under the locomotive's undercarriage through a camera on the robot's arm. This second 3D point cloud is then registered with the second preset point cloud template to obtain the corrected pose of the robot arm. Based on the corrected pose, the composite robot is moved to a position where the second 3D point cloud coincides with the second preset point cloud template, allowing for undercarriage detection of the locomotive at this location. This method significantly reduces the customization requirements and modification workload for the composite robot's pit environment and improves the relative positioning capability of the composite robot for special target detection components.

[0081] Example 2:

[0082] This embodiment proposes a locomotive undercarriage detection and positioning device, such as... Figure 5 As shown, it includes:

[0083] The first point cloud acquisition module is used to acquire the three-dimensional point cloud of the target detection component under the locomotive under the locomotive under the inspection, which is collected by the camera at the body of the composite robot when the locomotive under inspection arrives at the inspection position, and obtain the first three-dimensional point cloud.

[0084] The first point cloud registration module is used to register the first three-dimensional point cloud with the first preset point cloud template to obtain a pose transformation matrix with a correction offset. The first preset point cloud template is a three-dimensional point cloud containing the pose relationship between the composite robot body and the target detection component.

[0085] The first positioning adjustment module is used to move the composite robot to a position where the first three-dimensional point cloud and the first preset point cloud template coincide, based on the correction offset of the pose transformation matrix, so as to perform undercarriage detection of the vehicle to be inspected at the position.

[0086] The first point cloud acquisition module is connected to the first point cloud registration module, and the first point cloud registration module is connected to the first positioning adjustment module.

[0087] If, based on the correction offset of the pose transformation matrix, the composite robot cannot be moved to a position where the first 3D point cloud coincides with the first preset point cloud template, the locomotive undercarriage detection and positioning device further includes:

[0088] The second point cloud acquisition module is used to acquire the three-dimensional point cloud of the target detection component under the locomotive to be inspected, which is captured by the camera at the robotic arm of the composite robot when the composite robot reaches the measurement point position, and obtain the second three-dimensional point cloud.

[0089] The second registration module is used to register the second three-dimensional point cloud with the second preset point cloud template to obtain the corrected pose of the robotic arm. The second preset point cloud template is a three-dimensional point cloud containing the pose relationship between the robotic arm of the composite robot and the target detection component.

[0090] The second positioning adjustment module is used to move the composite robot to a position where the second three-dimensional point cloud and the second preset point cloud template coincide, based on the corrected pose of the robotic arm, so as to perform undercarriage inspection of the vehicle to be inspected at the position where the second three-dimensional point cloud and the second preset point cloud template coincide.

[0091] This embodiment proposes a locomotive undercarriage detection and positioning device. A first point cloud acquisition module obtains a first 3D point cloud. A first point cloud registration module registers the first 3D point cloud with a first preset point cloud template to obtain a pose transformation matrix with corrected offset. Finally, a first positioning adjustment module moves a composite robot to a position where the first 3D point cloud and the first preset point cloud template coincide, based on the corrected offset of the pose transformation matrix, to perform undercarriage detection of the locomotive at this position. Furthermore, if the composite robot cannot be moved to the position where the first 3D point cloud and the first preset point cloud template coincide based on the corrected offset of the pose transformation matrix, a second point cloud acquisition module is activated to obtain a second 3D point cloud. A second registration module registers the second 3D point cloud with a second preset point cloud template to obtain a corrected pose for the robotic arm. Then, a second positioning adjustment module moves the composite robot to the position where the second 3D point cloud and the second preset point cloud template coincide, to perform undercarriage detection of the locomotive at this position. The device in this embodiment can significantly reduce the customization requirements and modification workload of the pit environment for composite robots; and improve the relative positioning capability of composite robots for special target detection components.

[0092] Example 3:

[0093] This embodiment proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the locomotive undercarriage detection and positioning method described above.

[0094] Example 4:

[0095] 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 locomotive undercarriage detection and positioning method.

[0096] 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 locomotive undercarriage detection and positioning method as described in the embodiments. It is understood that the electronic device may also include an input / output (I / O) interface and communication components.

[0097] The processor is used to execute all or part of the steps in the locomotive undercarriage detection and positioning method 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.

[0098] 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 locomotive undercarriage detection and positioning method described in the above embodiments.

[0099] Example 5:

[0100] This embodiment proposes a computer-readable storage medium storing executable instructions, which, when executed, cause a processor to perform the locomotive undercarriage detection and positioning method.

[0101] 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.

[0102] 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 locomotive undercarriage detection and positioning method described in the various embodiments of this application.

[0103] 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, on which computer programs are stored. When the computer program is executed by a processor, it can implement the various steps of the locomotive undercarriage detection and positioning method described above.

[0104] 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.

[0105] 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 of inspecting and positioning a locomotive underframe, comprising: Includes the following steps: When the locomotive to be inspected arrives at the inspection position, the three-dimensional point cloud of the target inspection component under the locomotive to be inspected is obtained by the camera at the body of the composite robot, and the first three-dimensional point cloud is obtained. The first three-dimensional point cloud is registered with the first preset point cloud template to obtain a pose transformation matrix with a correction offset. The first preset point cloud template is a three-dimensional point cloud containing the pose relationship between the composite robot body and the target detection component. Based on the correction offset of the pose transformation matrix, the composite robot is moved to the position where the first three-dimensional point cloud and the first preset point cloud template coincide, so as to perform the undercarriage detection of the vehicle to be detected at the position where the first three-dimensional point cloud and the first preset point cloud template coincide. If, based on the correction offset of the pose transformation matrix, the composite robot cannot be moved to a position where the first 3D point cloud coincides with the first preset point cloud template, the locomotive undercarriage detection and positioning method further includes: when the composite robot reaches the measuring point position, acquiring a 3D point cloud of the target detection component of the locomotive undercarriage to be detected, captured by a camera at the robotic arm of the composite robot, to obtain a second 3D point cloud; registering the second 3D point cloud with the second preset point cloud template to obtain a corrected pose of the robotic arm, wherein the second preset point cloud template is a 3D point cloud containing the pose relationship between the robotic arm of the composite robot and the target detection component; and, based on the corrected pose of the robotic arm, moving the composite robot to a position where the second 3D point cloud coincides with the second preset point cloud template, so as to perform undercarriage detection of the locomotive undercarriage at the position where the second 3D point cloud coincides with the second preset point cloud template.

2. The method of claim 1, wherein, The process of obtaining the first preset point cloud template includes: When the locomotive to be inspected arrives at the inspection position, the activity space of the robotic arm of the composite robot and the shooting posture are adjusted to meet their respective preset conditions, and the position of the composite robot in the above state is determined as the measurement point position of the composite robot. The first preset point cloud template is obtained by acquiring the three-dimensional point cloud of the target component under the vehicle body of the vehicle to be inspected, which is captured by the camera at the vehicle body of the composite robot.

3. The method of claim 2, wherein, The step of acquiring the three-dimensional point cloud of the target detection component under the locomotive to be inspected, collected by a camera at the body of the composite robot, to obtain the first three-dimensional point cloud includes: When the composite robot reaches the measurement point, the three-dimensional point cloud of the target detection component under the locomotive to be inspected is obtained by the camera at the body of the composite robot.

4. The method of claim 1, wherein, The pose transformation matrix with corrected offset is as follows: in, Rot The rotation matrix is ​​used to rotate the first 3D point cloud to a first preset point cloud template. Tran To translate the rotated 3D point cloud to a first preset point cloud template using a translation vector, the translation vector includes a correction offset. This is the pose transformation matrix with corrected offset.

5. The locomotive undercarriage detection and positioning method according to claim 1, characterized in that, The process of obtaining the second preset point cloud template includes: When the locomotive to be inspected arrives at the inspection position, the activity space of the robotic arm of the composite robot and the shooting posture are adjusted to meet their respective preset conditions, and the position of the composite robot in the above state is determined as the measurement point position of the composite robot. A second preset point cloud template is obtained by acquiring the three-dimensional point cloud of the target detection component under the locomotive to be inspected, which is captured by the camera at the robotic arm of the composite robot.

6. The locomotive undercarriage detection and positioning method according to claim 1, characterized in that, The corrected pose of the robotic arm is calculated using the following formula: in, , To correct the pose of the robotic arm; This is the transformation matrix between the camera coordinate system of the camera on the vehicle body and the end effector coordinate system of the robotic arm; The pose of the robotic arm of the composite robot when obtaining the second preset point cloud template; The pose of the robotic arm of the composite robot when acquiring the second 3D point cloud.

7. A locomotive undercarriage detection and positioning device, characterized in that, include: The first point cloud acquisition module is used to acquire the three-dimensional point cloud of the target detection component under the locomotive under the locomotive under the inspection, which is collected by the camera at the body of the composite robot when the locomotive under inspection arrives at the inspection position, and obtain the first three-dimensional point cloud. The first point cloud registration module is used to register the first three-dimensional point cloud with the first preset point cloud template to obtain a pose transformation matrix with a correction offset. The first preset point cloud template is a three-dimensional point cloud containing the pose relationship between the composite robot body and the target detection component. The first positioning adjustment module is used to move the composite robot to a position where the first three-dimensional point cloud and the first preset point cloud template coincide, based on the correction offset of the pose transformation matrix, so as to perform undercarriage detection of the vehicle to be inspected at the position. If, based on the correction offset of the pose transformation matrix, the composite robot cannot be moved to a position where the first 3D point cloud coincides with the first preset point cloud template, the locomotive undercarriage detection and positioning device further includes: a second point cloud acquisition module, used to acquire a 3D point cloud of the target detection component of the locomotive undercarriage to be inspected, captured by a camera at the robotic arm of the composite robot, when the composite robot reaches the measuring point position, to obtain a second 3D point cloud; a second registration module, used to register the second 3D point cloud with the second preset point cloud template to obtain a corrected pose of the robotic arm, wherein the second preset point cloud template is a 3D point cloud containing the pose relationship between the robotic arm of the composite robot and the target detection component; and a second positioning adjustment module, used to move the composite robot to a position where the second 3D point cloud coincides with the second preset point cloud template, based on the corrected pose of the robotic arm, so as to perform undercarriage detection of the locomotive undercarriage at the position where the second 3D point cloud coincides with the second preset point cloud template.

8. 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 locomotive undercarriage detection and positioning method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed, cause the processor to perform the locomotive undercarriage detection and positioning method according to any one of claims 1-6.

Images