Running train wheel tread flaw detection device

By arranging the flaw detection unit and image acquisition components on both sides of the rail, the existing wheel tread detection methods have been solved, and efficient and accurate detection and monitoring of wheel tread defects have been achieved.

CN222728002UActive Publication Date: 2025-04-08SHIJIAZHUANG TIEDAO UNIV
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202421684897.X
Authority / Receiving Office
CN · China
Patent Type
Utility models(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2025-04-08
Estimated Expiration
2034-07-17

AI Technical Summary

Technical Problem

The existing wheel tread detection methods have low detection accuracy and low patrol efficiency. They cannot detect defects in the rail-side wheel tread in time, resulting in safety hazards and equipment damage.

Method used

A wheel tread detection device for driving trains is designed, and four sets of detection units are arranged along the rails. Each group of units includes multiple sets of image acquisition components. Through the combination of sensors, cameras and light sources, continuous detection and image acquisition of the entire circumference of the wheel are realized.

Benefits of technology

It realizes timely detection of wheel tread defects, improves detection accuracy and efficiency, meets visualization and dynamic monitoring requirements, and reduces detection costs in harsh environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN222728002U_ABST
    Figure CN222728002U_ABST
Patent Text Reader

Abstract

The utility model discloses a running train wheel tread flaw detection device, which belongs to the technical field of wheel flaw detection and comprises four groups of flaw detection units symmetrically arranged along rails on two sides, and the two groups of flaw detection units on the same side are sequentially arranged front and back to respectively perform flaw detection on front and rear wheels on two sides of a train bogie; the plurality of sets of image acquisition assemblies of each group of flaw detection unit are arranged at equal intervals along the rail on the same side and are used for performing flaw detection on the whole circumference of the wheel; and the image acquisition assembly transmits a detection result to the main control computer, so that the track-side wheel tread defects can be conveniently found in time. And the plurality of image acquisition assemblies of each group of flaw detection unit are used for detecting different parts of the wheel in sequence when the wheel passes through, so that flaw detection on the whole circumference of the wheel is realized. According to the utility model, the detection work efficiency and safety can be improved, the detection quality is improved, and the engineering cost is reduced; the system can timely find the potential safety hazard of the train and ensure the safe operation of the train, thereby improving the transport capacity and efficiency of the railway.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The utility model belongs to the technical field of wheel flaw detection, and particularly relates to a wheel tread flaw detection device for a running train. Background Art

[0002] As the construction of railway trains and tracks steadily develops towards high speed, intelligence and light weight, the operating speed, load capacity and transportation volume of trains continue to increase, the dynamic interaction between vehicle wheels and tracks gradually increases, and the problem of high-speed wheel tread defects becomes increasingly prominent. At the same time, during the running of trains, the unevenness of the track, excessive temperature and humidity, as well as long-term heavy load and high-speed operation will cause certain wear on the wheels. If these safety hazards are not discovered in time, they will have serious consequences. The wheel-rail impact caused by wheel tread defects may cause damage to the wheel-rail system components and abnormal vibration of the vehicle body, resulting in a series of impact and rolling noises, which seriously affect the service life of the vehicle track system components and the riding comfort of passengers, and even cause serious safety hazards to train operation.

[0003] The existing detection methods mainly include trackside detection methods (wheel-rail force detection method, laser detection method, acoustic emission detection method, etc.) and vehicle-mounted detection methods (axle box vibration acceleration detection method, fiber optic sensor detection method). At present, the detection accuracy of the wheel tread scratch detection method put into use is not high; traditional detection instruments perform inspection and maintenance work on a large number of trackside wheel treads in various environments, with low inspection efficiency and weak detection accuracy. They cannot promptly detect faults and abnormal problems that may be caused by defects in the trackside wheel treads; they can only perform qualitative analysis and cannot be visualized. The vibration acceleration method and the acoustic emission method have large adjacent pair noise interference and are suitable for low-speed train detection. Some methods cannot even achieve dynamic monitoring. Utility Model Content

[0004] In order to solve the above problems, the utility model provides a wheel tread flaw detection device for a running train.

[0005] In order to achieve the above purpose, the technical solution adopted by the utility model is as follows:

[0006] A wheel tread flaw detection device for a running train comprises four groups of flaw detection units arranged along two sides of the rails, respectively used for flaw detection on four wheels of a train bogie; the four groups of flaw detection units are symmetrically arranged in pairs along the two sides of the rails, and the two groups of flaw detection units on the same side are arranged in sequence front and back, respectively used for flaw detection on the front and rear wheels on both sides of the train bogie; each group of flaw detection units comprises a plurality of sets of image acquisition components, and the plurality of sets of image acquisition components of each group of flaw detection units are arranged at equal intervals along the same side of the rails, and are used for flaw detection on the entire circumference of the wheel; the image acquisition components of the four groups of flaw detection units are all connected to a main control computer.

[0007] Furthermore, the image acquisition component includes a sensor, a camera and a light source, the sensor is arranged at the inner bottom of the rail, the camera and the light source are both inclined toward the wheel and arranged on the outer side of the rail, the camera is used to capture images of different parts of the wheel, and the light source provides lighting for the camera; the sensor, camera and light source are all connected to a PLC controller, and the PLC controller is connected to a main control computer.

[0008] Furthermore, the sensor is arranged on the top of the bracket, and the lower end of the bracket is connected to the inner bottom edge of the rail through a fixing block. The bracket is an inverted L-shape, the sensor is threadedly matched with the upper horizontal short side of the bracket, and the lower end of the long side of the bracket is provided with a long hole for bolts to pass through, which is used to adjust the height of the sensor; the fixing block can be fixed to the bottom edge of the rail, and the bracket is connected to the fixing block by bolts.

[0009] Furthermore, the light source is tilted on the support, the camera is tilted on the top of the adjustment frame, the support and the adjustment frame are both fixed to the outside of the rail; the light source is set on one side of the camera, and the inclination angles of the two match, and the light sources and cameras of the front and rear groups of flaw detection units on the same side of the rail are tilted toward the front and rear wheels respectively.

[0010] Furthermore, the adjustment frame includes a base, a column, a swing arm and a clamping sleeve, the column is vertically fixed on the base, the swing arm is connected to the swing arm through a damping joint, the clamping sleeve is arranged on the top of the swing arm, and the camera can be fixed in the clamping sleeve.

[0011] Furthermore, the image acquisition components are in five sets, and the five cameras are respectively used to capture an image of one fifth of the wheel circumference.

[0012] Furthermore, the light source is an LED strobe light, and the light emission time of the LED strobe light is 2ms; the camera is an industrial area array camera, and the exposure time of the industrial area array camera is within 2ms.

[0013] Furthermore, the sensor is a magnetic steel sensor.

[0014] Furthermore, five sets of image acquisition components are all arranged on a base plate, and the base plate is fixed to the outside of the rail; a foldable fan-shaped shield is provided on the inner edge of the base plate, and the fan-shaped shield is connected to a flipping mechanism for driving the fan-shaped shield to unfold and cover the camera and light source; the flipping mechanism is connected to a PLC controller.

[0015] Furthermore, one side of the fan-shaped shield is fixed to the bottom plate through a baffle, and the other side of the fan-shaped shield is connected to a movable rod, and both ends of the movable rod are respectively connected to a flip mechanism through a rocker rod; a flexible baffle cloth is provided between the baffle, both ends of the fan-shaped shield and the rocker rod.

[0016] Compared with the prior art, the technical progress achieved by the utility model is:

[0017] The utility model arranges four groups of flaw detection units on both sides of the rails, and the two groups of flaw detection units on the same side are arranged in sequence front and back, respectively, to perform flaw detection on the front and rear wheels on both sides of the train bogie; multiple sets of image acquisition components of each group of flaw detection units are arranged at equal intervals along the rails on the same side, and when a high-speed train passes by, different circumferential parts of the wheel can be continuously detected in sequence to achieve flaw detection of the entire circumference of the wheel; the image acquisition components of the four groups of flaw detection units transmit the detection results to the main control computer, which is convenient for timely detection of trackside wheel tread defects, solves the safety hazards of the train, promotes the continuous speeding up of the train, increases the load and transportation volume, and thus improves the railway transportation capacity and efficiency. The use of the utility model can improve the efficiency and safety of detection work, improve the accuracy of detection, meet the requirements of visualization and dynamic monitoring, and greatly reduce the engineering cost required to detect wheel tread defects in harsh environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention.

[0019] In the attached picture:

[0020] Figure 1 An application state diagram of a wheel tread flaw detection device for a traveling train provided by an embodiment of the utility model;

[0021] Figure 2 for Figure 1 A partial enlarged view of the middle A;

[0022] Figure 3 This is a diagram of the closed state of the fan-shaped shield in the embodiment of the utility model;

[0023] Figure 4 for Figure 3 The unfolded state diagram of the middle sector shield;

[0024] Figure 5 A specific application flow chart of an embodiment of the utility model;

[0025] Figure 6 This is a schematic diagram of the shooting ranges of different cameras during the wheel rolling process in the embodiment of the utility model;

[0026] Figure 7 It is the operation diagram of the multi-label classification algorithm;

[0027] In the figure:

[0028] 00-rail, 01-bogie, 02-wheel; 100-flaw detection unit;

[0029] 1-sensor; 2-camera; 3-light source; 4-bracket; 5-fixed block; 6-long hole; 7-support; 8-adjusting frame, 81-base, 82-column, 83-swing arm, 84-clamping sleeve, 85-damping joint; 9-bottom plate; 10-fan-shaped shield; 11-baffle; 12-movable rod; 13-swing rod; 14-flipping mechanism; 15-flexible baffle. DETAILED DESCRIPTION

[0030] The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. The following will describe the embodiments of the present utility model in conjunction with the accompanying drawings.

[0031] like Figure 1 , Figure 2 and Figure 6 As shown, a wheel tread flaw detection device for a running train comprises four groups of flaw detection units 100 arranged along the rails 00 on both sides, which are respectively used to perform flaw detection on four wheels 02 of a train bogie 01; the four groups of flaw detection units 100 are symmetrically arranged in pairs along the rails 00 on both sides, and the two groups of flaw detection units 100 on the same side are arranged in sequence front and back, which are respectively used to perform flaw detection on the front and rear wheels 02 on both sides of the train bogie 01; each group of flaw detection units 100 comprises a plurality of sets of image acquisition components, and the plurality of sets of image acquisition components of each group of flaw detection units 100 are arranged at equal intervals along the rails 00 on the same side, so as to realize flaw detection on the entire circumference of the running wheel 02; the image acquisition components of the four groups of flaw detection units 100 are all connected to a main control computer (not shown in the figure), so as to facilitate image processing and other operations on the main control computer, and to perform targeted analysis on the experimental results.

[0032] As a preferred structure, Figure 2As shown, the image acquisition component includes a sensor 1, a camera 2 and a light source 3. The sensor 1 is arranged at the inner bottom of the rail 00. The camera 2 and the light source 3 are both inclined toward the wheel 02 and arranged on the outer side of the rail 00. The camera 2 is used to collect images of different parts of the wheel 02, and the light source 3 provides lighting for the camera 2. The sensor 1, the camera 2 and the light source 3 are all connected to a PLC controller (not shown in the figure), and the PLC controller is connected to a main control machine. The camera in this solution uses an industrial area array camera, which is a digital camera with the advantages of excellent performance, compact design, affordable price, easy installation and use, etc. The camera can transmit image data through the GigE data interface and can work stably in various harsh environments. It is suitable for the detection of wheel tread defects in various complex railway environments. At the same time, the light source added on the side is a compensating light source, which effectively reduces the impact of reflection and vibration on the camera's acquisition of trackside wheel tread images, and takes into account abnormal detection in different lighting environments. It can promptly discover faults and abnormal problems that may be caused by trackside wheel tread defects, greatly reduce the serious consequences caused by further expansion due to untimely detection, and further ensure the stable and safe operation of railway vehicles.

[0033] During specific installation, the sensor 1 is arranged on the top of the bracket 4, and the lower end of the bracket 4 is connected to the inner bottom edge of the rail 00 through the fixing block 5. Among them, the bracket 4 is an inverted L-shape, and the sensor 1 is threadedly matched with the upper horizontal short side of the bracket 4. The lower end of the long side of the bracket 4 is provided with a long hole 6 for the bolt to pass through, so as to facilitate the adjustment of the height of the sensor 1; the fixing block 5 can be fixed to the bottom edge of the rail 00, and the bracket 4 is connected to the fixing block 5 by bolts. Among them, the sensor 1 is a magnetic steel sensor, which is a prior art, and the specific structure and working principle will not be repeated here. Under long-term use and different environmental conditions, the magnetic steel sensor can still maintain good performance and measurement accuracy, and can also measure a wide range from weak magnetic fields to strong magnetic fields, and can be applied to a variety of different environments and media. The specific working process of the magnetic steel sensor is as follows:

[0034] When the train reaches a specific position, the magnetic field changes, and the sensing element (such as Hall element or fluxgate element) in the magnetic steel sensor interacts with the measured magnetic field. The sensing element generates a corresponding electrical signal, and the circuit inside the sensor amplifies, conditions and digitizes the generated electrical signal, and then activates the camera to capture the wheel tread. In addition, the magnetic steel sensor adopts the above installation method to avoid direct contact with the measured wheel, reduce interference and damage to the measured object, and help ensure the accuracy of the measurement results.

[0035] In a specific embodiment of the present invention, Figure 2As shown, the light source 3 is tilted on the support 7, and the camera 2 is tilted on the top of the adjustment frame 8. The support 7 and the adjustment frame 8 are both fixed on the outside of the rail 00; the light source 3 is set on one side of the camera 2, and the tilt angles of the two are matched. The light sources 3 and cameras 2 of the two groups of flaw detection units 100 on the same side of the rail 00 are tilted toward the front and rear wheels 02 respectively. Among them, the adjustment frame 8 includes a base 81, a column 82, a swing arm 83 and a clamping sleeve 84. The column 82 is vertically fixed on the base 81, and the swing arm 83 is connected to the swing arm 83 through a damping joint 85. The clamping sleeve 84 is set on the top of the swing arm 83, and the camera 2 can be fixed in the clamping sleeve 84. The use of this structure can adjust the tilt angle of the camera according to the detection requirements.

[0036] The utility model has more advantages than the existing visual tread defect detection method. At present, cameras are used to shoot the wheels. If a few cameras are used to shoot, the images shot are not complete because the occlusion of the brake shoe of the bogie brake device is not considered; if a camera is used to shoot multiple photos, it will cause difficulty in later image splicing. If a linear array camera is used for shooting, it will cause large imaging distortion because it is suitable for scenes with unchanged working distance; and although the area array camera can meet the task requirements without reducing the speed of the vehicle, the image collection is still disturbed by reflective vibration, and it is expensive. The utility model uses five sets of image acquisition components to collect wheel tread images. The cameras are installed on the side of the rail at a certain pitch angle and horizontal offset angle. Each set of cameras needs to be equipped with an LED strobe light as a compensation light source. The magnetic steel sensor is fixed at a fixed distance from the camera and installed vertically on the side of the rail. When the train travels to a specific area, the magnetic steel sensor is triggered to generate a shooting trigger signal, and the magnetic steel sensor sends a level signal to the camera and the light source. The camera and the light source start working at the same time and capture a part of the wheel tread circumference image. When the train passes through five sets of image acquisition components in sequence, a complete tread image of one side of the train wheel is obtained to skip the reflective area.

[0037] exist Figure 1 In the specific embodiment shown, there are five sets of image acquisition components, and five cameras 2 are used to capture one-fifth of the circumference of the wheel 02. The light source 3 is an LED strobe light, and the light emission time of the LED strobe light is 2ms. The corresponding exposure time of the industrial array camera is within 2ms. The LED strobe light is matched with the exposure time of the industrial array camera to capture a portion of the tread circumference image. When the train passes through the five sets of image acquisition components in sequence, a complete tread image of a single wheel of the train is obtained, such as Figure 6 The flaw detection unit composed of five sets of image acquisition components can save a lot of resources and greatly improve the detection efficiency and quality.

[0038] In specific applications, the MER2-503-23GM / CP camera driver software is used to support simultaneous acquisition of multiple cameras. The IP addresses of the five cameras are configured and read in turn. The acquisition control settings, basic property settings, and basic image processing settings can be performed in the parameter setting interface on the right. When the trigger mode is Off, the acquisition interface is a real-time interface. When the acquisition control is set to frame trigger mode, a signal from the sensor is received and an image is captured. The video storage plug-in in the camera driver software is used to continuously store images of multiple cameras. Through the PLC controller, each of them is responsible for the corresponding wheel pair and output signal, thereby avoiding the problem of later image screening and realizing the synchronous operation of the camera and light source. Finally, a multi-label classification algorithm is used for the image to effectively identify the various defects in the tread image. The specific process is as follows: Figure 5 As shown in the figure. Through intelligent label classification, the rail defect situation is presented, which greatly improves the detection efficiency and meets the needs of flaw detection.

[0039] The multi-label classification algorithm can be used to import tread images with marked tread defect types into the algorithm for training. Users only need to import the test image into the trained model for recognition to obtain the predicted score of the wheel tread defect in the corresponding image and its defect type, which greatly improves the efficiency of wheel tread defect detection and effectively presents the collected image information in a concise label form, greatly improving the detection efficiency of wheel inspectors. The specific flow chart is as follows Figure 7 shown.

[0040] The above MER2-503-23GM / CP camera driver software also supports real-time preview and recording functions, which is convenient for users to view the shooting effects in real time. After installing the driver, users can download and start the image viewing software. When starting the image viewing software, users need to connect the 5 cameras to the computer for power supply to ensure that the cameras are in normal working condition. After double-clicking the detected camera device, the image viewing software will display the real-time preview screen of the camera. In the preview screen, users can click the capture button to capture the image. The captured image will be saved in the specified folder, and users can view and process it in the image viewing software. It is also equipped with a PLC controller to control its corresponding wheel pair, and the output signal is completed through the PLC controller to achieve synchronous operation of the camera and the light source.

[0041] At the same time, the above-mentioned MER2-503-23GM / CP camera driver software also supports SDK development, allowing users to customize camera functions and image processing algorithms according to their own needs, and can realize more advanced functions such as autofocus, image recognition, etc.

[0042] Further optimize the above technical solution, such as Figure 3 , 4As shown, five sets of image acquisition components are all arranged on the bottom plate 9, and the bottom plate 9 is fixed on the outside of the rail 00; a foldable fan-shaped shield 10 is provided on the inner edge of the bottom plate 9, and the fan-shaped shield 10 is connected to a flip mechanism 14, which is used to drive the fan-shaped shield 10 to unfold and cover the camera 2 and the light source 3; the flip mechanism 14 is connected to a PLC controller. Among them, one side of the fan-shaped shield 10 is fixed to the bottom plate 9 through a baffle 11, and the other side of the fan-shaped shield 10 is connected to a movable rod 12, and the two ends of the movable rod 12 are respectively connected to the flip mechanism 14 through a swing rod 13; a flexible baffle 15 is provided between the baffle 11, the two ends of the fan-shaped shield 10 and the swing rod 13, and the baffle, the fan-shaped shield and the flexible baffles at both ends can be used to enclose a semi-enclosed space to protect the LED strobe light and the camera. The flip mechanism includes a flip motor and a transmission gear set, and the output shaft of the flip motor drives the swing rod to swing through the transmission gear set to realize the opening and closing of the fan-shaped shield. The specific application process is as follows:

[0043] During the inspection of the moving wheels, the fan-shaped shield closes backwards without affecting the cooperation of the LED strobe light and the camera to capture images of the wheels; when the acquisition is completed, the fan-shaped shield unfolds to cover the LED strobe light and the camera to prevent the lampshade and lens from being contaminated by debris carried by passing trains and affecting subsequent image acquisition.

[0044] In summary, the utility model has the advantages of simple structure, low cost and high detection accuracy. The image acquisition components of the flaw detection units on both sides of the rails can continuously detect different circumferential parts of the wheels of high-speed trains to achieve flaw detection of the entire circumference of the wheels, and transmit the detection results to the main control computer for image processing and analysis, thereby improving the efficiency of wheel tread defect detection. The utility model can timely detect wheel tread defects on the trackside, solve the safety hazards of the train, and ensure the safe operation of the train. The utility model meets the requirements of visualization and dynamic monitoring, can improve the accuracy of detection, and greatly reduce the engineering cost required to detect wheel tread defects in harsh environments.

[0045] Finally, it should be noted that the above description is only a preferred embodiment of the utility model and is not intended to limit the utility model. Although the utility model is described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein with equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the utility model shall be included in the scope of protection of the claims of the utility model.

Claims

1. A wheel tread flaw detection device for a moving train, characterized in that: It includes four groups of flaw detection units arranged along the rails on both sides, which are respectively used to perform flaw detection on the four wheels of the train bogie; the four groups of flaw detection units are symmetrically arranged in pairs along the rails on both sides, and the two groups of flaw detection units on the same side are arranged in sequence front and back, which are respectively used to perform flaw detection on the front and rear wheels on both sides of the train bogie; each group of flaw detection units includes multiple sets of image acquisition components, and the multiple sets of image acquisition components of each group of flaw detection units are arranged at equal intervals along the rails on the same side, which are used to perform flaw detection on the entire circumference of the wheel; the image acquisition components of the four groups of flaw detection units are all connected to the main control computer.

2. A wheel tread flaw detection device for a moving train according to claim 1, characterized in that: The image acquisition component includes a sensor, a camera and a light source. The sensor is arranged at the inner bottom of the rail. The camera and the light source are both inclined toward the wheel and arranged on the outer side of the rail. The camera is used to collect images of different parts of the wheel. The light source provides lighting for the camera. The sensor, camera and light source are all connected to a PLC controller, and the PLC controller is connected to a main control computer.

3. A wheel tread flaw detection device for a moving train according to claim 2, characterized in that: The sensor is arranged on the top of the bracket, and the lower end of the bracket is connected to the inner bottom edge of the rail through a fixing block.

4. The wheel tread flaw detection device for a moving train according to claim 2, characterized in that: The light source is tilted on the support, the camera is tilted on the top of the adjustment frame, and the support and the adjustment frame are both fixed to the outside of the rail; the light source is set on one side of the camera, and the tilt angles of the two match, and the light sources and cameras of the front and rear flaw detection units on the same side of the rail are tilted toward the front and rear wheels respectively.

5. The wheel tread flaw detection device for a moving train according to claim 4, characterized in that: The adjustment frame includes a base, a column, a swing arm and a clamping sleeve, the column is vertically fixed on the base, the swing arm is connected to the swing arm through a damping joint, the clamping sleeve is arranged on the top of the swing arm, and the camera can be fixed in the clamping sleeve.

6. The wheel tread flaw detection device for a moving train according to claim 2, characterized in that: There are five sets of image acquisition components, and the five cameras are respectively used to capture an image of one fifth of the wheel circumference.

7. The wheel tread flaw detection device for a moving train according to claim 6, characterized in that: The light source is an LED strobe light, and the light emission time of the LED strobe light is 2ms; the camera is an industrial area array camera, and the exposure time of the industrial area array camera is within 2ms.

8. The wheel tread flaw detection device for a moving train according to claim 2, characterized in that: The sensor is a magnetic steel sensor.

9. A wheel tread flaw detection device for a moving train according to any one of claims 2 to 8, characterized in that: Five sets of image acquisition components are all arranged on a base plate, and the base plate is fixed to the outside of the rail; a foldable fan-shaped shield is provided on the inner edge of the base plate, and the fan-shaped shield is connected to a flip mechanism for driving the fan-shaped shield to unfold and cover the camera and light source; the flip mechanism is connected to a PLC controller.

10. The wheel tread flaw detection device for a moving train according to claim 9, characterized in that: One side of the fan-shaped shield is fixed to the bottom plate through a baffle, and the other side of the fan-shaped shield is connected to a movable rod, and both ends of the movable rod are respectively connected to a flip mechanism through a swing rod; a flexible baffle cloth is provided between the baffle, both ends of the fan-shaped shield and the swing rod.

Citation Information

Cited By

  • Railway wheel detection system and method based on prefabricated ballast bed

    CN120628649A