A vehicle pre-inspection device for a railway wagon and a pre-inspection method thereof

Abstract

The application provides a vehicle pre-inspection device and a pre-inspection method of a railway freight car. The device comprises: a truss track system, which comprises double-side overhead trusses arranged on both sides of a railway freight car body and a bottom truss arranged in a pit directly below the car body; a mechanical arm, which is respectively installed on the double-side overhead trusses and the bottom truss and can move along the corresponding track; and a scanning head, which is integrally installed at the tail end of each group of the mechanical arms and is used for collecting surface images and three-dimensional topographic data of the railway freight car. Through the cooperative matching of the truss track system, the mechanical arm and the scanning head, the detection dimension is expanded from the local static coverage of the traditional fixed camera to the full-car three-dimensional dynamic scanning.

Description

Technical Field

[0001] This invention relates to the field of railway engineering machinery technology, and in particular to a pre-inspection device and method for railway freight cars. Background Technology

[0002] In the railway freight car maintenance process, pre-inspection is the first and crucial step, and its quality directly determines the accuracy of subsequent maintenance plans and the efficiency of maintenance resource allocation. Currently, the industry mainstream still uses traditional manual inspection methods: inspectors use a hammer to tap components to listen for damage, combined with visual observation to determine the car's condition. This method has significant limitations—a complete inspection of a single car takes over twenty minutes, and the results are greatly affected by subjective and objective factors such as lighting conditions, personnel fatigue, and experience differences, leading to a high rate of misjudgment in damage location and characterization. The chalk marking method used during inspection is difficult to retain for long periods and cannot form structured data, causing an information gap between pre-inspection information and subsequent stages such as maintenance workshops and parts storage, severely hindering the intelligent upgrading of maintenance. The closest existing solution is a semi-automatic visual inspection system, which relies on a fixed camera array and can only cover local areas of the car body, such as side walls and ends. Critical blind spots such as the car body bottom and bogies still require manual supplementary inspection. Although the purchase cost of a single set of equipment is relatively low, two additional operators are required to operate the equipment and verify the results, resulting in persistently high labor costs. While the inspection time per vehicle is shorter than that of manual methods, it still takes nearly ten minutes. Furthermore, due to the limitations of image recognition algorithms, the accuracy rate for identifying complex damage, such as rust penetration and hidden cracks, only meets the basic industry requirements, and there is still a significant gap from the goal of precise maintenance.

[0003] Existing semi-automatic vision inspection systems suffer from three main shortcomings: First, significant blind spots. The static installation of fixed cameras has significant limitations, failing to effectively cover densely packed areas such as the underbody bogie assembly (e.g., bolsters, side frames) or enclosed areas like the inner walls and corners of the vehicle body, leading to the risk of missed detection of critical structural damage. Second, a lack of intelligent judgment capabilities. The system only provides basic image acquisition, and damage type determination still relies on human experience. This approach fails to establish unified damage assessment standards and lacks the ability to automatically diagnose and support decisions for complex faults. Third, data collaboration is fragmented. Inspection results are stored in unstructured formats, such as paper records or local images, preventing real-time data interaction with the vehicle maintenance management system. Repair plan formulation relies on manual secondary information transmission, which is prone to errors and hinders efficient scheduling of maintenance resources.

[0004] In summary, there is an urgent need to develop an inspection solution that can achieve full vehicle coverage, possess complex damage identification and diagnosis capabilities, and interface with the maintenance management system, in order to completely overcome the limitations of existing technologies. Summary of the Invention

[0005] This invention provides a pre-inspection device and method for railway freight cars, which enables full-area coverage scanning and automatic damage identification of the exterior, interior and bottom bogies of the car body.

[0006] In a first aspect, the present invention provides a pre-inspection device for railway freight cars, comprising: A truss track system, comprising double-sided elevated trusses disposed on both sides of the railway freight car body, and a bottom truss disposed in a pit directly below the car body; The robotic arms are respectively mounted on the double-sided elevated truss and the bottom truss, and can move along the corresponding tracks; the robotic arms mounted on the double-sided elevated truss are used to extend into the interior of the railway freight car body and the corner area for inspection, and the robotic arms mounted on the bottom truss are used to extend into the bottom bogie area of ​​the railway freight car body for inspection. The scanning head, integrated and installed at the end of each set of robotic arms, is used to collect surface images and three-dimensional topographic data of the railway freight cars.

[0007] Optionally, the bottom truss adopts a pit-embedded sealed cavity structure.

[0008] Optionally, the scanning head includes a visible light camera and a laser 3D scanner; The visible light camera is used to acquire images of the surface in order to capture surface texture details; The laser 3D scanner is used to acquire the 3D topographic data.

[0009] Optionally, the end effector of the robotic arm integrates a force sensor to monitor the distance between the scanning head and the railway freight car in real time, and dynamically adjust the posture of the robotic arm according to the distance between the car body to avoid collision.

[0010] Optionally, the robotic arm is a six-axis robotic arm.

[0011] In a second aspect, the present invention provides a pre-inspection method for railway freight cars, applied to the pre-inspection device for railway freight cars described in the first aspect, the method comprising: When the railway freight car to be inspected is towed to the scanning station, the origin of the car body coordinate system of the railway freight car to be inspected is automatically calibrated by the laser locator, and the detection parameter template of the corresponding model is loaded. The robotic arms on the double-sided elevated trusses are controlled to move synchronously along both sides of the railway freight car to be inspected, and the scanning head is used to collect surface images of the outer surface and inner cavity wall of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data. The robotic arm on the bottom truss is controlled to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and the surface image of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, are collected through the scanning head. The surface image and the corresponding three-dimensional topography data are fused together, and the damage type and damage level are identified based on a pre-trained damage recognition model. Based on the damage type and the damage level, a structured inspection result is generated, which includes a damage distribution heatmap, a maintenance level assessment, and a spare parts requirement list.

[0012] Optionally, the robotic arm on the bottom truss is controlled to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and the surface image of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, are acquired through the scanning head, including: The robotic arm on the bottom truss is controlled to rotate around the bogie at multiple angles to obtain surface images and three-dimensional topographic data of the brake beam pin holes and bolster stop of the braking system. The distance between the robotic arm and the components of the braking system is monitored in real time by force sensors, and a close-range fine scan is performed when the distance between the components is less than a preset distance threshold.

[0013] Optionally, the surface image and the corresponding three-dimensional topography data are fused, and the damage type and damage level are identified based on a pre-trained damage recognition model, including: The three-dimensional coordinates of the three-dimensional topography data are mapped and aligned with the pixel space of the surface image. The viewpoint difference is eliminated by coordinate transformation to obtain aligned multimodal data. The aligned multimodal data is input into the damage identification model to obtain the damage type and the damage level.

[0014] Optionally, the aligned multimodal data is input into the damage identification model to obtain the damage type and the damage level, including: If the inspected railway freight car is found to have side wall concavity or end wall structural cracks, a first-level structural integrity alarm will be triggered. If the inspected railway freight car is found to have a detached brake shoe or a missing connecting pin, a level two emergency alarm will be triggered. If scratches are detected on the railway freight car to be inspected and the depth exceeds a preset depth threshold, or if there is rust and the rust area exceeds a preset area threshold, a maintenance priority suggestion will be generated.

[0015] Optionally, during the process of controlling the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the railway freight car to be inspected, the method further includes: The surface image and the three-dimensional morphology data are used for defect detection in real time. If a preset major damage type is detected, the scan is immediately interrupted and an emergency alarm is triggered.

[0016] Thirdly, the present invention provides a pre-inspection device for railway freight cars, applied to the pre-inspection device for railway freight cars described in the first aspect, the device comprising: The calibration module is used to automatically calibrate the origin of the car body coordinate system of the railway freight car to be inspected by means of a laser locator when the freight car to be inspected is towed to the scanning station, and to load the detection parameter template of the corresponding model. The first acquisition module is used to control the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the railway freight car body to be inspected, and to acquire surface images of the outer surface and inner cavity wall of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, through the scanning head. The second acquisition module is used to control the robotic arm on the bottom truss to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and to acquire surface images of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, through the scanning head. The damage prediction module is used to fuse the surface image and the corresponding three-dimensional morphology data, and to identify the damage type and damage level based on a pre-trained damage recognition model. The test result generation module is used to generate structured test results, including a damage distribution heatmap, a maintenance level assessment, and a spare parts requirement list, based on the damage type and the damage level.

[0017] Fourthly, the present invention provides an electronic device including a processor and a memory, the memory storing computer-readable instructions that, when executed by the processor, perform the steps of the method provided in the first aspect above.

[0018] Fifthly, the present invention provides a storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method provided in the first aspect above.

[0019] In a sixth aspect, the present invention provides a computer program product comprising a computer program that, when executed by a processor, performs the steps of the method provided in the first aspect above.

[0020] As can be seen from the above technical solutions, the present invention has the following advantages: This invention provides a pre-inspection device and method for railway freight cars. The device includes: a truss track system comprising double-sided elevated trusses disposed on both sides of the freight car body and a bottom truss disposed in a pit directly below the car body; robotic arms respectively mounted on the double-sided elevated trusses and the bottom truss, and movable along corresponding tracks; the robotic arms mounted on the double-sided elevated trusses are used to extend into the interior and corner areas of the freight car body for inspection, and the robotic arms mounted on the bottom truss are used to extend into the bottom bogie area of ​​the freight car body for inspection; and a scanning head integrated at the end of each set of robotic arms for acquiring surface images and three-dimensional topographic data of the freight car. Through the coordinated operation of the truss track system, robotic arms, and scanning head, the inspection dimension is expanded from the local static coverage of traditional fixed cameras to full-vehicle three-dimensional dynamic scanning, providing a precise pose control basis for subsequent image acquisition and fault identification. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the structure of a pre-inspection device for railway freight cars according to the present invention; Figure 2 This is a schematic diagram of the robotic arm and scanning head structure of a pre-inspection device for railway freight cars according to the present invention; Figure 3 This is a flowchart illustrating the steps of a method for inspecting the side of a railway freight car according to an embodiment of the present invention. Figure 4 This is a flowchart illustrating the steps of a second embodiment of the inspection method for the side of a railway freight car according to the present invention. Figure 5 This is a structural block diagram of an embodiment of a railway freight car side inspection device according to the present invention. Detailed Implementation

[0023] This invention provides a pre-inspection device and method for railway freight cars, which enables full-area coverage scanning and automatic damage identification of the exterior, interior and bottom bogies of the car body.

[0024] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0025] Example 1 Please see Figure 1 , Figure 1 This is a schematic diagram of the pre-inspection structure of a railway freight car according to the present invention, including: a truss track system 1, wherein the truss track system 1 includes double-sided elevated trusses disposed on both sides of the railway freight car body 5, and a bottom truss disposed in the pit directly below the car body 5; The robotic arms are respectively mounted on the double-sided elevated truss and the bottom truss, and can move along the corresponding tracks; the left robotic arm 2 and the right robotic arm 4 mounted on the double-sided elevated truss are used to extend into the interior and corner areas of the railway freight car body 5 for inspection, and the bottom robotic arm 3 mounted on the bottom truss is used to extend into the bottom bogie area of ​​the railway freight car body 5 for inspection; The scanning head is integrated and mounted at the end of each robotic arm, specifically as follows: Figure 2 The schematic diagram of the robotic arm and scanning head structure of the pre-inspection device for railway freight cars of the present invention is shown, wherein 4 is the right robotic arm and 6 is the scanning head connected to the right robotic arm. The scanning head is used to collect surface images and three-dimensional morphological data of the railway freight car.

[0026] In this embodiment, the truss track system 1 consists of double-sided elevated trusses and a bottom truss. The double-sided elevated trusses are arranged along the length of the vehicle body 5, and their tops integrate high-precision linear guides to provide a stable horizontal motion reference for the robotic arms mounted on them. The bottom truss adopts a pit-embedded design, placing the scanning unit in the working area directly below the vehicle body 5. The robotic arms are driven from bottom to top to approach the undercarriage components through a lifting or translation mechanism. The double-sided trusses and the bottom truss share a unified spatial coordinate system, and the three sets of robotic arms can operate independently or synchronously along their respective tracks, achieving seamless coverage of the external facade, internal cavity, and bottom bogie area of ​​the vehicle body 5.

[0027] In an alternative embodiment, the bottom truss adopts a pit-embedded sealed cavity structure.

[0028] In this embodiment, the main structure of the bottom truss is embedded inside the pit, and a sealed cavity is provided externally to isolate the robotic arm, transmission mechanism, and electrical wiring from the external environment. The cavity as a whole achieves an IP68 protection rating. Through multiple sealing rings and pressure balancing components, it maintains internal cleanliness and dryness under the common conditions of water vapor, oil, and dust in railway maintenance plants, ensuring the long-term stable operation of the bottom robotic arm and scanning unit, thereby supporting the reliability of unmanned undercarriage inspection.

[0029] In one alternative embodiment, the scanning head includes a visible light camera and a laser 3D scanner; The visible light camera is used to acquire images of the surface in order to capture surface texture details; The laser 3D scanner is used to acquire the 3D topographic data.

[0030] In this embodiment, the scanning head integrates a visible light camera and a laser 3D scanner into the same housing, achieving coaxial or paraxial acquisition through internal optical path design. The visible light camera acquires surface texture details such as color, texture, rust, and scratches on the vehicle body surface, while the laser 3D scanner simultaneously projects line laser or structured light to acquire three-dimensional morphological data of the component surface.

[0031] In one alternative embodiment, the end effector of the robotic arm integrates a force sensor for real-time monitoring of the distance between the scanning head and the railway freight car body, and dynamically adjusting the robotic arm posture according to the distance to avoid collision.

[0032] In this embodiment, a force sensor is installed between the end flange of the robotic arm and the scanning head, continuously collecting contact force and torque signals as the robotic arm approaches the vehicle body. When the distance between the scanning head and the vehicle body is detected to be too close or the contact force exceeds a preset threshold, the robotic arm posture is automatically adjusted to increase the distance based on the direction and amplitude feedback from the sensor. This ensures that when the scanning head operates in narrow areas such as bogie gaps and vehicle interior cavities, it can both obtain high-precision images at close range and avoid equipment damage caused by trajectory deviation.

[0033] In one alternative embodiment, the robotic arm is a six-axis robotic arm.

[0034] In this embodiment of the application, the six-axis robotic arm has six rotary joints, providing translational degrees of freedom in the X, Y, and Z directions and attitude degrees of freedom in the three directions, enabling its end-effector scanning head to reach designated detection points on the exterior, interior, and bottom bogie of the vehicle body in any spatial orientation.

[0035] This application provides a pre-inspection device for railway freight cars, comprising: a truss track system, including double-sided elevated trusses disposed on both sides of the freight car body and a bottom truss disposed in a pit directly below the car body; robotic arms, respectively mounted on the double-sided elevated trusses and the bottom truss, and capable of moving along corresponding tracks; the robotic arms mounted on the double-sided elevated trusses are used to extend into the interior and corner areas of the freight car body for inspection, and the robotic arms mounted on the bottom truss are used to extend into the bottom bogie area of ​​the freight car body for inspection; and a scanning head, integrated at the end of each set of robotic arms, for acquiring surface images and three-dimensional topographic data of the freight car. Through the coordinated operation of the truss track system, robotic arms, and scanning head, the inspection dimension is expanded from the local static coverage of traditional fixed cameras to a full-vehicle three-dimensional dynamic scan, providing a precise pose control basis for subsequent image acquisition and fault identification.

[0036] Example 2 Please see Figure 3 , Figure 3 This is a flowchart illustrating the steps of a pre-inspection method for railway freight cars according to Embodiment 1 of the present invention. The method is applied to the pre-inspection device for railway freight cars in Embodiment 1; the method includes: Step S101: When the railway freight car to be inspected is towed to the scanning station, the origin of the car body coordinate system of the railway freight car to be inspected is automatically calibrated by the laser locator, and the detection parameter template of the corresponding model is loaded. In this embodiment, the railway freight car to be inspected is precisely pushed along the track by the traction system to a preset scanning station. Multiple sets of laser positioners are installed on the station floor and both sides. The laser positioners emit laser beams towards preset reference points on the car body. Through multi-point ranging and triangulation, the actual position and deflection angle of the car body in the spatial coordinate system are calculated, automatically calibrating the origin of the car body coordinate system and eliminating scanning errors caused by deviations in the car's parking position. Simultaneously, a detection parameter template corresponding to the corresponding car model is loaded from the detection parameter library. The template includes preset parameters such as the robotic arm's motion trajectory, scanning point distribution, camera exposure parameters, laser scanning resolution, and damage discrimination threshold.

[0037] Step S102: Control the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the railway freight car to be inspected, and collect surface images of the outer surface and inner cavity wall of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, through the scanning head. In this embodiment, the scanning head at the end of the robotic arm integrates a high-resolution visible light camera and a laser 3D scanner, continuously acquiring data during movement. The visible light camera uses linear or area array imaging to obtain surface images of the outer surface and inner cavity walls of the railway freight car under inspection, capturing texture details, rust marks, scratches, and other surface features of the outer surface and inner cavity walls. The laser 3D scanner projects linear laser or structured light to acquire three-dimensional morphological data of the component surface in real time, recording morphological features and spatial coordinates. The two robotic arms move in tandem, simultaneously constructing digital images and 3D models of both sides of the car body and the inner cavity walls, providing complete basic data for subsequent damage identification.

[0038] Step S103: Control the robotic arm on the bottom truss to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and collect the surface image of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, through the scanning head; In this embodiment, the bottom truss adopts a pit-embedded sealed cavity structure. The bottom robotic arm, carrying a scanning head, performs multi-angle circumferential motion around the periphery of the bogie, sequentially aligning with each key component of the braking system. A visible light camera and a laser 3D scanner simultaneously acquire surface images and 3D topographic data of components such as brake shoes, pins, bolsters, and side frames. During the acquisition process, the robotic arm can automatically switch to a close-range fine scanning mode according to a preset path to ensure the data acquisition quality of concealed parts.

[0039] Step S104: The surface image and the corresponding three-dimensional topography data are fused, and the damage type and damage level are identified based on the pre-trained damage recognition model. In this embodiment, by using pre-calibrated camera intrinsic and extrinsic parameters, the coordinates of the 3D point cloud acquired by the laser 3D scanner are transformed and mapped to the pixel space of the image acquired by the visible light camera. This eliminates the viewpoint deviation caused by differences in the robot arm's posture, ensuring that each pixel corresponds to accurate spatial depth information. The aligned multimodal data is then input into a pre-trained damage recognition model, which outputs the damage type and damage level.

[0040] Step S105: Based on the damage type and the damage level, generate a structured inspection result that includes a damage distribution heatmap, a maintenance level assessment, and a spare parts requirement list.

[0041] In this embodiment, the damage distribution heatmap marks the spatial location and severity of each damage on the three-dimensional model of the vehicle body, and intuitively presents the concentrated damage area in a color gradient manner; the maintenance level assessment divides maintenance tasks into different priorities such as emergency maintenance, planned maintenance or condition monitoring according to the damage type and level, supporting the scheduling of maintenance resources; the spare parts requirement list automatically matches the required spare parts, such as brake shoes, connecting pins, seals, etc., according to the damage type, and outputs the specifications and quantity of spare parts.

[0042] The present invention provides a pre-inspection method for railway freight cars, comprising: when the railway freight car to be inspected is towed to the scanning station, automatically calibrating the origin of the car body coordinate system of the railway freight car to be inspected by a laser locator, and loading the detection parameter template of the corresponding car model; controlling the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the car body of the railway freight car to be inspected, and acquiring surface images of the outer surface and inner cavity walls of the railway freight car to be inspected, as well as the corresponding three-dimensional morphological data, through the scanning head; controlling the robotic arms on the bottom truss to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and acquiring surface images of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional morphological data, through the scanning head; performing fusion processing on the surface images and the corresponding three-dimensional morphological data, and identifying the damage type and damage level based on a pre-trained damage recognition model; and generating a structured inspection result including a damage distribution heatmap, maintenance level assessment, and spare parts requirement list based on the damage type and the damage level. By using a robotic arm with two elevated sides and a bottom truss to scan the surface image and three-dimensional shape data of the railway freight car to be inspected in a three-dimensional collaborative manner, and through data fusion and deep learning recognition, the system can achieve full-area coverage scanning and automatic damage identification of the exterior, interior and bottom bogies of the car body, solving the problems of many blind spots, low efficiency and inconsistent standards in traditional manual inspection.

[0043] Example 3 Please see Figure 4 , Figure 4 This is a flowchart illustrating a second embodiment of a pre-inspection method for railway freight cars according to the present invention. The method is applied to the pre-inspection device for railway freight cars in the first embodiment. The method includes: Step S201: When the railway freight car to be inspected is towed to the scanning station, the origin of the car body coordinate system of the railway freight car to be inspected is automatically calibrated by the laser locator, and the detection parameter template of the corresponding model is loaded. In this embodiment, after the vehicle to be inspected is precisely delivered into the scanning station by the traction system, the laser locator automatically projects onto a preset reference point on the vehicle body. Multi-point ranging is used to determine the actual position and attitude of the vehicle body in space, completing the automatic calibration of the origin of the vehicle coordinate system. Based on the identified vehicle model information, a corresponding detection parameter template is loaded from the detection parameter library. This template includes preset parameters such as the robotic arm's motion trajectory, scanning point distribution, imaging parameters, and damage discrimination thresholds.

[0044] Step S202: Control the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the railway freight car to be inspected, and collect surface images of the outer surface and inner cavity wall of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, through the scanning head; In this embodiment, the robotic arms on the dual elevated trusses move synchronously along the guide rails on both sides of the vehicle body, driven by servo motors and a rack and pinion transmission mechanism. The scanning head at the end of the robotic arm integrates a visible light camera and a laser 3D scanner, continuously acquiring surface images and 3D topographic data of the vehicle's outer surface and inner cavity walls during movement. The visible light camera captures surface texture details, while the laser 3D scanner simultaneously acquires 3D point clouds.

[0045] In step S203, during the process of controlling the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the railway freight car to be inspected, the surface images and the three-dimensional topography data are used in real time to detect defects. If a preset major damage type is detected, the scanning is immediately interrupted and an emergency alarm is triggered. In this embodiment of the application, when a preset major damage type is identified, such as severe concave-convex deformation of the side wall, structural cracking of the end wall, or obvious loss of key components, the current scanning task is immediately interrupted, an audible and visual emergency alarm is triggered, and a first-time alarm message containing the location and type of damage is generated so that on-site personnel can intervene in a timely manner to avoid missing major safety hazards.

[0046] Step S204: Control the robotic arm on the bottom truss to rotate around the bogie at multiple angles to obtain surface images and three-dimensional topographic data of the brake beam pin hole and bolster stop of the braking system; In this embodiment, the bottom truss adopts a pit-embedded sealed cavity structure, and its built-in robotic arm is lifted to the working position by a lifting mechanism. The robotic arm, carrying a scanning head, performs multi-angle circumferential motion around the bogie, sequentially aligning with blind spots in traditional manual inspections such as brake beam pin holes and bolster stops. Through the coordinated control of multi-degree-of-freedom joints, the robotic arm can penetrate the bogie link gaps and collect surface images and three-dimensional topographic data of key parts of the braking system from different directions, ensuring full coverage inspection of concealed structures.

[0047] Step S205: The distance between the robotic arm and the components of the braking system is monitored in real time by force sensors, and a close-range fine scan is performed when the distance between the components is less than a preset distance threshold. In this embodiment, a force sensor integrated at the end of the robotic arm monitors the distance and contact force between the scanning head and the braking system components in real time. When the detected distance is less than a preset distance threshold, the controller determines that it is entering a close-range fine scanning mode, automatically reduces the feed speed of the robotic arm, and dynamically adjusts the posture based on sensor feedback, so that the scanning head approaches the surface of the component while maintaining a safe distance, thereby acquiring higher resolution images and three-dimensional morphology data. This is suitable for fine detection scenarios such as crack depth and wear amount.

[0048] Step S206: Map and align the three-dimensional coordinates of the three-dimensional topography data with the pixel space of the surface image, and eliminate the viewpoint difference through coordinate transformation to obtain the aligned multimodal data; In this embodiment, the coordinates of the 3D point cloud acquired by the laser 3D scanner are transformed and mapped to the pixel space of the image acquired by the visible light camera. By using calibrated camera intrinsic and extrinsic parameters, a correspondence between the 3D coordinates and 2D pixels is established, eliminating the viewing angle deviation caused by differences in the robotic arm's posture. The aligned multimodal data allows the same detection point to simultaneously possess surface texture information and depth topography information.

[0049] Step S207: Input the aligned multimodal data into the damage identification model to obtain the damage type and the damage level; In this embodiment of the application, if the side wall of the railway freight car to be inspected is found to be deformed or the end wall is structurally cracked, a first-level structural integrity alarm is triggered; if the brake shoe is found to be detached or the connecting pin is lost, a second-level emergency alarm is triggered; if the scratches on the railway freight car to be inspected exceed a preset depth threshold, or if there is corrosion and the corrosion area exceeds a preset area threshold, a maintenance priority suggestion is generated.

[0050] In the implementation, the aligned multimodal data is input into a damage recognition model built on a convolutional neural network. This model, trained on a large-scale database of railway freight car damage samples, is capable of automatically segmenting and classifying rusted areas, crack orientations, and deformation degrees. The model outputs damage type and damage level, providing a decision-making basis for a graded alarm mechanism.

[0051] Step S208: Based on the damage type and the damage level, generate a structured inspection result that includes a damage distribution heatmap, a maintenance level assessment, and a spare parts requirement list.

[0052] In this embodiment, a structured inspection report is generated by summarizing all identified damage information. The report includes a damage distribution heatmap, marking the location of each damage on the vehicle body's 3D model; it outputs a maintenance level assessment, classifying damages into emergency repair, planned repair, or condition monitoring levels according to severity; and it automatically matches a spare parts requirement list based on the damage type, such as brake shoes and connecting pins.

[0053] The present invention provides a pre-inspection method for railway freight cars, comprising: when the railway freight car to be inspected is towed to the scanning station, automatically calibrating the origin of the car body coordinate system of the railway freight car to be inspected by a laser locator, and loading the detection parameter template of the corresponding car model; controlling the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the car body of the railway freight car to be inspected, and acquiring surface images of the outer surface and inner cavity walls of the railway freight car to be inspected, as well as the corresponding three-dimensional morphological data, through the scanning head; controlling the robotic arms on the bottom truss to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and acquiring surface images of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional morphological data, through the scanning head; performing fusion processing on the surface images and the corresponding three-dimensional morphological data, and identifying the damage type and damage level based on a pre-trained damage recognition model; and generating a structured inspection result including a damage distribution heatmap, maintenance level assessment, and spare parts requirement list based on the damage type and the damage level. By employing a three-dimensional collaborative scanning system using robotic arms on both sides of the elevated structure and the bottom truss, the system scans the surface images and 3D topographic data of the railway freight cars under inspection. Through data fusion and deep learning recognition, it achieves full-area coverage scanning and automatic damage identification of the car body's exterior, interior, and bottom bogies, solving the problems of numerous blind spots, low efficiency, and inconsistent standards associated with traditional manual inspection. Simultaneously, a real-time defect screening and emergency interruption mechanism during the scanning process ensures timely response to major damage. Furthermore, force sensors enable close-range, fine scanning, improving inspection accuracy.

[0054] Example 4 Please see Figure 5 , Figure 5 This is a structural block diagram of an embodiment of a pre-inspection device for railway freight cars according to the present invention. The device is applied to the pre-inspection device for railway freight cars in Embodiment 1; the device includes: The calibration module 301 is used to automatically calibrate the origin of the car body coordinate system of the railway freight car to be inspected by means of a laser locator when the railway freight car to be inspected is towed to the scanning station, and load the detection parameter template of the corresponding model. The first acquisition module 302 is used to control the robotic arms on the double-sided elevated truss to move synchronously along both sides of the railway freight car body to be inspected, and to acquire surface images of the outer surface and inner cavity wall of the railway freight car to be inspected, as well as the corresponding three-dimensional morphological data, through the scanning head. The second acquisition module 303 is used to control the robotic arm on the bottom truss to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and to acquire the surface image of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, through the scanning head. The damage prediction module 304 is used to fuse the surface image and the corresponding three-dimensional morphology data, and to identify the damage type and damage level based on a pre-trained damage recognition model. The test result generation module 305 is used to generate a structured test result that includes a damage distribution heat map, a maintenance level assessment, and a spare parts requirement list based on the damage type and the damage level.

[0055] In an optional embodiment, the second acquisition module 303 includes: The surface data acquisition submodule is used to control the robotic arm on the bottom truss to rotate around the bogie at multiple angles to acquire surface images and three-dimensional topographic data of the brake beam pin holes and bolster stops of the braking system. The close-range scanning submodule is used to monitor the distance between the robotic arm and the components of the braking system in real time through force sensors, and to perform close-range fine scanning when the distance between the components is less than a preset distance threshold.

[0056] In an optional embodiment, the damage prediction module 304 includes: The fusion submodule is used to map and align the three-dimensional coordinates of the three-dimensional topography data with the pixel space of the surface image, eliminate the viewpoint difference through coordinate transformation, and obtain the aligned multimodal data. The prediction submodule is used to input the aligned multimodal data into the damage identification model to obtain the damage type and the damage level.

[0057] In an optional embodiment, the prediction submodule includes: The first prediction unit is used to trigger a primary structural integrity alarm if the side wall of the railway freight car under inspection is found to be deformed or the end wall is structurally cracked. The second prediction unit is used to trigger a level two emergency alarm if it detects that the brake shoe has fallen off or the connecting pin is missing in the railway freight car under inspection. The third prediction unit is used to generate maintenance priority suggestions if it detects scratches on the railway freight car to be inspected that exceed a preset depth threshold, or if it detects rust that exceeds a preset area threshold.

[0058] In an optional embodiment, the first acquisition module 302 further includes: The emergency alarm submodule is used to perform defect detection on the surface image and the three-dimensional morphology data in real time. If a preset major damage type is detected, the scanning is immediately interrupted and an emergency alarm is triggered.

[0059] Example 5 This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of a pre-inspection method for railway freight cars according to any embodiment.

[0060] Example 6 This invention also provides a computer storage medium storing a computer program thereon, which, when executed by the processor, implements the steps of a pre-inspection method for railway freight cars according to any embodiment.

[0061] Example 7 This invention also provides a computer program product having a computer program stored thereon, wherein the computer program, when executed by the processor, implements the steps of a pre-inspection method for railway freight cars according to any embodiment.

[0062] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0063] In the several embodiments provided in this application, it should be understood that the methods, apparatuses, electronic devices, and storage media disclosed in this invention can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0064] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0065] Furthermore, the functional units in the various embodiments of the present invention 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. The integrated unit can be implemented in hardware or as a software functional unit.

[0066] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0067] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A pre-inspection device for railway freight cars, characterized in that, include: A truss track system, comprising double-sided elevated trusses disposed on both sides of the railway freight car body, and a bottom truss disposed in a pit directly below the car body; The robotic arms are respectively mounted on the double-sided elevated truss and the bottom truss, and can move along the corresponding tracks; the robotic arms mounted on the double-sided elevated truss are used to extend into the interior of the railway freight car body and the corner area for inspection, and the robotic arms mounted on the bottom truss are used to extend into the bottom bogie area of ​​the railway freight car body for inspection. The scanning head, integrated and installed at the end of each set of robotic arms, is used to collect surface images and three-dimensional topographic data of the railway freight cars.

2. The pre-inspection device for railway freight cars according to claim 1, characterized in that, The bottom truss adopts a pit-embedded sealed cavity structure.

3. The pre-inspection device for railway freight cars according to claim 1, characterized in that, The scanning head includes a visible light camera and a laser 3D scanner; The visible light camera is used to acquire images of the surface in order to capture surface texture details; The laser 3D scanner is used to acquire the 3D topographic data.

4. The pre-inspection device for railway freight cars according to claim 1, characterized in that, The robotic arm has a force sensor integrated at its end, which is used to monitor the distance between the scanning head and the railway freight car in real time, and dynamically adjust the posture of the robotic arm according to the distance between the car body to avoid collision.

5. The pre-inspection device for railway freight cars according to claim 1, characterized in that, The robotic arm is a six-axis robotic arm.

6. A method for pre-inspection of railway freight cars, characterized in that, The method of the pre-inspection device for railway freight cars according to any one of claims 1-5 includes: When the railway freight car to be inspected is towed to the scanning station, the origin of the car body coordinate system of the railway freight car to be inspected is automatically calibrated by the laser locator, and the detection parameter template of the corresponding model is loaded. The robotic arms on the double-sided elevated trusses are controlled to move synchronously along both sides of the railway freight car to be inspected, and the scanning head is used to collect surface images of the outer surface and inner cavity wall of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data. The robotic arm on the bottom truss is controlled to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected, and the surface image of the braking system of the railway freight car to be inspected, as well as the corresponding three-dimensional topographic data, are collected through the scanning head. The surface image and the corresponding three-dimensional topography data are fused together, and the damage type and damage level are identified based on a pre-trained damage recognition model. Based on the damage type and the damage level, a structured inspection result is generated, which includes a damage distribution heatmap, a maintenance level assessment, and a spare parts requirement list.

7. The pre-inspection method for railway freight cars according to claim 6, characterized in that, The robotic arm on the bottom truss is controlled to lift and approach the undercarriage bogie area of ​​the railway freight car to be inspected. The scanning head then acquires surface images of the braking system of the railway freight car to be inspected, as well as corresponding three-dimensional topographic data, including: The robotic arm on the bottom truss is controlled to rotate around the bogie at multiple angles to obtain surface images and three-dimensional topographic data of the brake beam pin holes and bolster stop of the braking system. The distance between the robotic arm and the components of the braking system is monitored in real time by force sensors, and a close-range fine scan is performed when the distance between the components is less than a preset distance threshold.

8. The pre-inspection method for railway freight cars according to claim 6, characterized in that, The surface image and the corresponding three-dimensional topography data are fused, and the damage type and damage level are identified based on a pre-trained damage recognition model, including: The three-dimensional coordinates of the three-dimensional topography data are mapped and aligned with the pixel space of the surface image. The viewpoint difference is eliminated by coordinate transformation to obtain aligned multimodal data. The aligned multimodal data is input into the damage identification model to obtain the damage type and the damage level.

9. The pre-inspection method for railway freight cars according to claim 8, characterized in that, The aligned multimodal data is input into the damage identification model to obtain the damage type and the damage level, including: If the inspected railway freight car is found to have side wall concavity or end wall structural cracks, a first-level structural integrity alarm will be triggered. If the inspected railway freight car is found to have a detached brake shoe or a missing connecting pin, a level two emergency alarm will be triggered. If scratches are detected on the railway freight car to be inspected and the depth exceeds a preset depth threshold, or if there is rust and the rust area exceeds a preset area threshold, a maintenance priority suggestion will be generated.

10. The pre-inspection method for railway freight cars according to claim 6, characterized in that, The process of controlling the robotic arms on the double-sided elevated trusses to move synchronously along both sides of the railway freight car to be inspected also includes: The surface image and the three-dimensional morphology data are used for defect detection in real time. If a preset major damage type is detected, the scan is immediately interrupted and an emergency alarm is triggered.

Images