Dynamic detection and intelligent early warning method, system and device for railway wagon brake shoe

Abstract

The invention discloses a rail wagon brake shoe dynamic detection and intelligent early warning method, system and device, and belongs to the technical field of rail traffic safety detection. The method comprises the following steps: extracting sub-pixel edge information of a brake shoe image from a wheel bogie high-definition image through a deep learning overall nested edge detection algorithm; extracting three-dimensional information of the wheel brake shoe from the sub-pixel edge information of the brake shoe image by using binocular stereoscopic vision; measuring edge contour information of upper and lower brake shoes according to the three-dimensional information of the wheel brake shoes, obtaining arcs through fitting, and calculating the distance between the arcs at the two ends as the surplus of the wheel brake shoes; comparing the surplus of the wheel brake shoe with a standard brake shoe to obtain a wear value of the wheel brake shoe; inputting the brake shoe wear value into a long-term memory network model, and evaluating a wheel brake shoe wear rule; brake shoe operation faults are early warned in real time according to a wheel brake shoe abrasion rule; according to the invention, the wear condition of each wheel brake shoe is detected through binocular stereo vision, and automatic early warning of the overrun or non-overrun brake shoe is realized.

Description

Technical Field

[0001] The invention relates to a method, system and device for dynamic detection and intelligent early warning of railway freight car brake shoes, belonging to the technical field of photoelectric detection in rail transit safety detection. Background Art

[0002] Promoting the precise maintenance of freight train brake shoes will bring great value to railway safety improvement and operation and maintenance cost reduction. The development of brake shoe online dynamic detection system is a prerequisite for precise brake shoe maintenance. At present, brake shoes are still the main braking method for freight trains in my country. In recent years, my country's railway freight mileage has continued to grow, showing characteristics such as heavy loads, high traffic density, and complex operating environment. The consumption and speed of vehicle brake shoes are accelerating.

[0003] At present, brake shoe inspection is mainly done by manual visual inspection. First, due to limited energy and efficiency of personnel, it is difficult to achieve 100% comprehensive inspection, and temporary spot inspections are prone to safety hazards. Second, simple mechanical work takes up a lot of train inspection personnel's time, reducing overall work efficiency. Third, qualitative observation by eyes relies heavily on skill experience and personnel fatigue status, making it difficult to achieve objective and quantitative accurate measurement. The application of brake shoe online dynamic detection system will promote the transformation from "manual inspection" to "intelligent inspection", free up more human resources for other high-value work, and help improve railway safety and reduce costs and increase efficiency.

[0004] Therefore, achieving high-precision dynamic visual inspection of train brake shoes is of great significance to ensuring train operation safety and realizing intelligent operation and maintenance. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method, system and device for dynamic detection and intelligent early warning of railway freight car brake shoes, so as to realize high-precision dynamic visual detection of train brake shoes, thereby ensuring the safety of train operation and realizing intelligent operation and maintenance; In order to achieve the above purpose / solve the above technical problems, the present invention is implemented by adopting the following technical solutions: In a first aspect, a method for dynamic detection and intelligent early warning of brake shoes of railway freight cars is provided, the method comprising: Get high-definition images of wheels and bogies; The wheel bogie high-definition image is used through a deep learning overall nested edge detection algorithm to extract the sub-pixel edge information of the brake shoe image; The three-dimensional information of wheel brake shoe is extracted by using binocular stereo vision to extract the sub-pixel edge information of brake shoe image; According to the three-dimensional information of the wheel brake shoe, the edge contour information of the upper and lower brake shoes is measured, and the arc is obtained by fitting, and the distance between the two ends of the arc is calculated as the remaining amount of the wheel brake shoe; Compare the remaining amount of wheel brake shoes with the standard brake shoes to obtain the wheel brake shoe wear value; Input the brake shoe wear value into the long-term memory network model and output the wheel brake shoe wear law; The wear level is determined based on the wheel brake shoe wear pattern and the preset wear threshold, and a brake shoe operation failure is warned in real time based on the wear level.

[0006] Optionally, the method of extracting three-dimensional information of wheel brake shoes from sub-pixel edge information of brake shoe images using binocular stereo vision includes: setting is the camera coordinate system, and the image plane coordinate system is , is the camera coordinate system, and the image plane coordinate system is , the camera coordinate system is used as the global coordinate system of the binocular vision measurement system; The three-dimensional information of the wheel brake shoe, , They are Imaging on the image planes of camera 1 and camera 2; , are the rotation matrix and translation vector of the camera 1 to camera 2 coordinate system transformation, respectively. It can be expressed as: in, , For point The projection coordinates on the camera imaging plane are: , For point The projection coordinates on the second imaging plane of the camera, T represents transposition; Represents the rotation matrix represents the components of the rotation matrix, represents the displacement vector, , , represents the elements in the displacement vector; The focal length of camera one, The focal length of the second camera. Optionally, the formula for calculating the distance between the two end arcs as the remaining amount of the wheel brake shoe is: , where Contour point to Perpendicular intersection, calculate the distance between the two points, and calculate all Up to The vertical distance is recorded as the remaining thickness of the brake shoe.

[0007] Optionally, the step of inputting brake shoe wear values ​​into a long-term memory network model and outputting wheel brake shoe wear rules includes: The brake shoe wear values ​​are serialized or normalized and then input into the long-term memory network model.

[0008] In a second aspect, a railway freight car brake shoe dynamic detection and intelligent early warning system is provided, the system comprising: An image acquisition module, used to obtain high-definition images of the wheel bogie; A detection module is used to obtain the sub-pixel edge information of the brake shoe image by using a deep learning overall nested edge detection algorithm on the wheel bogie high-definition image; An extraction module, used for extracting three-dimensional information of wheel brake shoes from sub-pixel edge information of brake shoe images using binocular stereo vision; A calculation module is used to measure the edge profile information of the upper and lower brake shoes according to the three-dimensional information of the wheel brake shoes, obtain the arc by fitting, and calculate the distance between the two end arcs as the remaining amount of the wheel brake shoes; A comparison module is used to compare the remaining amount of the wheel brake shoe with the standard brake shoe to obtain the wheel brake shoe wear value; An evaluation module is used to input brake shoe wear values ​​into a long-term memory network model and output wheel brake shoe wear patterns; The early warning module is used to determine the wear level according to the wear law of the wheel brake shoe and the preset wear threshold, and to issue a real-time early warning of brake shoe operation failure based on the wear level.

[0009] In the third aspect, a dynamic detection and intelligent early warning device for brake shoes of railway freight cars includes trackside equipment, machine room equipment and a monitoring center; the trackside equipment includes a vehicle number antenna, an array magnetic steel and detection boxes symmetrically arranged on both sides of the track, and the detection box is provided with a binocular stereo vision sensor for high-definition imaging of wheel brake shoes; the vehicle number antenna is used to collect the model and number of passing trains; the array magnetic steel is used to trigger the binocular stereo vision sensor, and the monitoring center is used to detect brake shoe data for remote viewing and push alarm brake shoe information to user terminals in multiple locations; the machine room equipment provides power for the trackside equipment and the monitoring center.

[0010] Optionally, the binocular stereo vision sensor is composed of two cameras, and the two cameras and the object to be measured form a triangle in space.

[0011] Optionally, the camera is equipped with a light source device.

[0012] Optionally, a temperature sensor is provided in the detection box, and the temperature sensor is used to measure the internal temperature of the box.

[0013] Optionally, the signal antenna is fastened to the sleeper by clamping the sleeper.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention proposes an outdoor high-definition imaging technology for railway freight car brake shoes. It uses a high-power stroboscopic near-infrared light source to illuminate the industrial camera, which can effectively avoid the influence of outdoor visible light on the imaging effect. At the same time, it can achieve all-weather presentation, laying the foundation for accurate identification and detection of brake shoe status; 2. The present invention uses a deep learning overall nested edge detection algorithm to extract the sub-pixel edge information of the brake shoe image, ensuring that the visual system can accurately capture the brake shoe image of each wheel and achieve accurate positioning of the brake shoe; 3. The present invention adopts brake shoe edge image extraction technology based on deep learning to ensure that the brake shoe image edge can still be accurately extracted in an outdoor low signal-to-noise ratio environment. At the same time, based on the binocular stereo vision measurement principle, the brake shoe thickness can be accurately measured, laying the foundation for accurately evaluating the brake shoe status and wear. 4. Based on a large amount of data on the brake shoes corresponding to each wheel, the present invention models the wear of the brake shoes of each wheel based on the long-term memory network, accurately evaluates the wear patterns and trends, and more accurately evaluates the current brake shoe operating conditions and safety margins, providing precise maintenance data support to ensure brake shoe maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Shown is a work flow chart of Example 1 of the present invention; Figure 2 Shown is a device diagram of embodiment 3 of the present invention; Figure 3 The figure shows the binocular stereo vision measurement principle of the first embodiment of the present invention; Figure 4 Shown is a schematic diagram of brake shoe wear detection according to Example 1 of the present invention. DETAILED DESCRIPTION

[0016] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0017] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and the like are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0018] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0019] Example 1 The present invention Figure 1 As shown, a method for dynamic detection and intelligent early warning of railway freight car brake shoes is disclosed, which specifically includes: Step 1, obtaining a high-definition image of the wheel bogie; Step 2, the wheel bogie high-definition image is subjected to a deep learning overall nested edge detection algorithm to extract the sub-pixel edge information of the brake shoe image; Step 3, using binocular stereo vision to extract the three-dimensional information of the wheel brake shoe from the sub-pixel edge information of the brake shoe image; Step 4, based on the three-dimensional information of the wheel brake shoe, measure the edge contour information of the upper and lower brake shoes, obtain the arc by fitting, and calculate the distance between the two ends of the arc as the remaining amount of the wheel brake shoe; Step 5, comparing the remaining amount of the wheel brake shoe with the standard brake shoe to obtain the wheel brake shoe wear value; Step 6, input the brake shoe wear value into the long-term memory network model to evaluate the wheel brake shoe wear law; Step 7, judging the wear level according to the wheel brake shoe wear law and the preset wear threshold, and giving a real-time warning of brake shoe operation failure according to the wear level.

[0020] The present embodiment is as follows: Step 1: When a train arrives, start data collection and brake shoe image preprocessing.

[0021] The brake shoe of a truck is hidden inside the bogie, and the visible area is small. It is easily affected by self-occlusion, dust, light and other factors, resulting in poor recognition accuracy and edge extraction accuracy. In order to improve the anti-interference and accuracy of this system, a deep learning-based image edge accurate extraction method is adopted, which has the characteristics of anti-interference from outdoor environmental factors and high-precision extraction. The edge extraction algorithm of deep learning using Holistically-Nested Edge Detection (HED) has two major characteristics: (1) overall image training and prediction; (2) multi-scale and multi-layer feature learning. The deep model uses a fully convolutional network to automatically learn rich hierarchical representations (based on deep supervision guidance of side responses).

[0022] Step 2-3: Three-dimensional measurement of brake shoe wear.

[0023] The measurement principle based on binocular stereo vision is adopted. Binocular stereo vision imitates human eyes to obtain three-dimensional information. Figure 3 The two cameras and the object to be measured form a triangle in space, and the three-dimensional coordinates of the space point are obtained by using the coordinates of the imaging points of the two cameras.

[0024] set up is the camera coordinate system, and the image plane coordinate system is , is the camera coordinate system, and the image plane coordinate system is , the camera coordinate system is used as the global coordinate system of the binocular vision measurement system; The three-dimensional information of the wheel brake shoe, , They are Imaging on the image planes of camera 1 and camera 2; , are the rotation matrix and translation vector of the camera 1 to camera 2 coordinate system transformation, respectively. It can be expressed as: in, , For point Projection coordinates on the imaging surface of camera 1, , For point Projection coordinates on the imaging surface of camera 2, T represents transposition; Represents the rotation matrix represents the components of the rotation matrix, represents the displacement vector, , , represents the elements in the displacement vector; The focal length of camera 1, The focal length of camera 2.

[0025] , where Contour point to Perpendicular intersection, calculate the distance between the two points, and calculate all Up to The vertical distance is recorded as the remaining thickness of the brake shoe.

[0026] For step 7: Intelligent early warning of brake shoe operation status.

[0027] Based on the large amount of data of the brake shoe corresponding to each wheel, the wear of each wheel brake shoe is modeled based on the long-term memory network (LSTM) to accurately evaluate its wear pattern and trend. The Long Short-Term Memory Network (LSTM) model is a specific form of Recurrent Neural Network (RNN). The LSTM model solves the problem of RNN short-term memory by adding gates on the basis of the RNN model, so that the recurrent neural network can truly and effectively utilize long-distance time series information. LSTM adds three logical control units, namely input gate, output gate, and forget gate, to the basic structure of RNN, and each is connected to a multiplication element. The weights at the edge where the memory unit of the neural network is connected to other parts control the input and output of the information flow and the state of the cell unit (Memory cell).

[0028] The system uses technologies such as machine vision and artificial intelligence to provide real-time warning of brake shoe operation failures, meeting the train inspection's needs for online dynamic detection of brake shoe operation status, eliminating safety accidents, and improving train driving safety; it can effectively free up the workload of manual inspection and greatly promote inspection efficiency and quality. The inspection time for a train can be reduced from the original 2 hours of manual inspection to 10 minutes of automatic inspection, effectively reducing labor costs while also avoiding manual operation risks; through modular design and rapid on-site installation and maintenance operations, the subsequent disassembly and assembly operations for line maintenance are reduced, which can reduce operating costs; multi-point inspection data can be networked for big data analysis and early warning, effectively promoting the railway train inspection work to leap to the intelligent operation and maintenance stage.

[0029] The brake shoe wear values ​​are input into the long-term memory network model (LSTM) to evaluate the wheel brake shoe wear law. The data processing and training are as follows: (1) Convert the input data into a format suitable for LSTM processing, including serialization, normalization and other steps; set the structure of the LSTM network, including the number of hidden layers, the number of units in each layer, the activation function, etc., and initialize the weights and bias items in the network.

[0030] (2) Input the sequence data into the LSTM network one by one according to the time step; Cell state update: At each time step, LSTM controls the flow of information through the forget gate, input gate, and output gate to update the cell state and hidden state.

[0031] (3) Select an appropriate loss function based on the specific task, such as the cross entropy loss function in classification tasks, the mean square error loss function in regression tasks, etc. Calculate the loss value: Compare the actual output of the network with the expected output and calculate the loss value.

[0032] (4) Starting from the last time step, reversely calculate the error term for each time step. The calculation of the error term involves backpropagating the gradient of the loss function to the LSTM unit at each time step. The error term is propagated from the current time step to the past time steps.

[0033] (5) Use the gradient descent algorithm or its variants (such as Adam, RMSprop, etc.) to update the weights and bias terms in the network. The weight update direction is the opposite direction of the gradient of the loss function to minimize the loss function. During the training process, it may be necessary to adjust the learning rate to control the step size of the weight update to speed up the training speed or improve the training effect. Repeat the process of forward propagation, loss calculation, backpropagation and weight update until a certain stopping condition is met (such as reaching the maximum number of iterations, the loss value is less than a certain threshold, etc.). Example 2

[0034] Disclosed is a railway freight car brake shoe dynamic detection and intelligent early warning system, the system comprising: An image acquisition module, used to obtain high-definition images of the wheel bogie; A detection module is used to obtain the sub-pixel edge information of the brake shoe image by using a deep learning overall nested edge detection algorithm on the wheel bogie high-definition image; An extraction module, used for extracting three-dimensional information of wheel brake shoes from sub-pixel edge information of brake shoe images using binocular stereo vision; A calculation module is used to measure the edge profile information of the upper and lower brake shoes according to the three-dimensional information of the wheel brake shoes, obtain the arc by fitting, and calculate the distance between the two end arcs as the remaining amount of the wheel brake shoes; A comparison module is used to compare the remaining amount of the wheel brake shoe with the standard brake shoe to obtain the wheel brake shoe wear value; An evaluation module is used to input brake shoe wear values ​​into a long-term memory network model to evaluate the wear pattern of wheel brake shoes; The early warning module is used to determine the wear level according to the wear law of the wheel brake shoe and the preset wear threshold, and to issue a real-time early warning of brake shoe operation failure based on the wear level. Example 3

[0035] Disclosed is a dynamic detection and intelligent early warning device for brake shoes of railway freight cars, comprising trackside equipment, machine room equipment and a monitoring center; the trackside equipment comprises a vehicle number antenna, an array magnetic steel and detection boxes symmetrically arranged on both sides of the track, wherein a binocular stereo vision sensor for high-definition imaging of wheel brake shoes is arranged in the detection box; the vehicle number antenna is used to collect the model and train number of passing trains; the array magnetic steel is used to trigger the binocular stereo vision sensor, and the monitoring center is used to detect brake shoe data for remote viewing and to push alarm brake shoe information to user terminals in multiple locations; the machine room equipment provides power for the trackside equipment and the monitoring center.

[0036] The railway freight car brake shoe dynamic detection and intelligent early warning system consists of three parts: trackside equipment, machine room equipment, and monitoring center. The trackside equipment consists of detection boxes, car number antennas, and array magnets. The detection boxes are symmetrically arranged on the prefabricated cement foundations on both sides of the track. They have the functions of high-definition imaging of the brake shoes on both sides of the wheels, imaging the upper and lower brake shoes of the wheels respectively, and measuring each brake shoe through binocular stereo vision sensors. The specific layout is as follows: Figure 2 As shown in the figure; the AEI car number antenna is fastened to the sleeper by means of a clamp, which is used to collect the model and number of the passing train, meeting the needs of the detection system for vehicle templates, data analysis and status tracking. The array magnet is used for axle metering and triggering functions such as cameras. The outer protective box of the detection box adopts anti-corrosion processing technology and is designed in accordance with the TB / T2846-2015 standard. It is used to install light sources, cameras and other devices; it is also equipped with a protective door and a temperature sensor. When the train is detected, the protective door is automatically opened, and the protective door is automatically closed after the train passes, which plays a dust and rainproof function, ensuring that the maintenance cycle is not less than 3 months. The temperature sensor is used to measure the internal temperature of the box, and timely take cooling and heating operations to ensure the stability of the sensor working environment. The equipment in the trackside machine room is placed in a 36U standard cabinet, mainly including electrical control boxes, industrial computers, car number hosts, remote power management and lightning protection units. The monitoring center is mainly composed of repeater terminals, network access equipment, servers, etc., which are terminal devices for remote viewing of system detection data. The application platform can push alarm brake shoe information to user terminals in multiple locations, and at the same time support users in multiple locations to perform queries and other operations through the application platform in parallel through terminal devices.

[0037] The present invention uses a high-definition industrial camera, supplemented by a near-infrared light source, to obtain high-definition images of wheel bogies in outdoor environments around the clock. Secondly, the sub-pixel edge information of the brake shoe image is obtained through a deep learning target detection and edge extraction algorithm, and the wear condition of each wheel brake shoe is obtained through the binocular stereo vision measurement principle. Finally, by identifying information such as the train number, the wear condition of each wheel brake shoe is analyzed, and an early warning model is established to achieve automatic early warning of over-limit or under-limit brake shoes. At the same time, it provides simultaneous access to multiple users within the railway bureau, realizes accurate maintenance of the wheel brake shoe device, greatly improves the inspection and maintenance efficiency of truck brake shoes, and achieves cost reduction and efficiency improvement.

[0038] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A railway freight car brake shoe dynamic detection and intelligent early warning method, characterized in that: The method comprises: Obtain high-definition images of wheels and bogies; The wheel bogie high-definition image is used through a deep learning overall nested edge detection algorithm to extract the sub-pixel edge information of the brake shoe image; The three-dimensional information of wheel brake shoe is extracted by using binocular stereo vision to extract the sub-pixel edge information of brake shoe image; According to the three-dimensional information of the wheel brake shoe, the edge contour information of the upper and lower brake shoes is measured, and the arc is obtained by fitting, and the distance between the two ends of the arc is calculated as the remaining amount of the wheel brake shoe; Compare the remaining amount of wheel brake shoes with the standard brake shoes to obtain the wheel brake shoe wear value; Input the brake shoe wear value into the long-term memory network model and output the wheel brake shoe wear law; The wear level is determined based on the wheel brake shoe wear pattern and the preset wear threshold, and a brake shoe operation failure is warned in real time based on the wear level.

2. The railway freight car brake shoe dynamic detection and intelligent early warning method according to claim 1 is characterized in that: The method of extracting three-dimensional information of wheel brake shoes from sub-pixel edge information of brake shoe images using binocular stereo vision includes: set up is the camera coordinate system, and the image plane coordinate system is , is the camera coordinate system, and the image plane coordinate system is , the camera coordinate system is used as the global coordinate system of the binocular vision measurement system; The three-dimensional information of the wheel brake shoe, , They are Imaging on the image planes of camera 1 and camera 2; , are the rotation matrix and translation vector of the camera 1 to camera 2 coordinate system transformation, respectively. It can be expressed as: in, , For point The projection coordinates on the camera imaging plane are: , For point The projection coordinates on the second imaging plane of the camera, T represents transposition; Represents the rotation matrix represents the components of the rotation matrix, represents the displacement vector, , , represents the elements in the displacement vector; The focal length of camera one, The focal length of camera two.

3. The railway freight car brake shoe dynamic detection and intelligent early warning method according to claim 1 is characterized in that: The formula for calculating the distance between the two end arcs as the remaining amount of the wheel brake shoe is: Among them, respectively Contour point to Perpendicular intersection, calculate the distance between the two points, and calculate all Up to The vertical distance is recorded as the remaining thickness of the brake shoe.

4. The railway freight car brake shoe dynamic detection and intelligent early warning method according to claim 1 is characterized in that: The step of inputting the brake shoe wear value into the long-term memory network model and outputting the wheel brake shoe wear law includes: The brake shoe wear values ​​are serialized or normalized and then input into the long-term memory network model.

5. A railway freight car brake shoe dynamic detection and intelligent early warning system, characterized in that: The system comprises: An image acquisition module, used to obtain high-definition images of the wheel bogie; A detection module is used to obtain the sub-pixel edge information of the brake shoe image by using a deep learning overall nested edge detection algorithm on the wheel bogie high-definition image; An extraction module, used for extracting three-dimensional information of wheel brake shoes from sub-pixel edge information of brake shoe images using binocular stereo vision; A calculation module is used to measure the edge profile information of the upper and lower brake shoes according to the three-dimensional information of the wheel brake shoes, obtain the arc by fitting, and calculate the distance between the two end arcs as the remaining amount of the wheel brake shoes; A comparison module is used to compare the remaining amount of the wheel brake shoe with the standard brake shoe to obtain the wheel brake shoe wear value; An evaluation module is used to input brake shoe wear values ​​into a long-term memory network model and output wheel brake shoe wear patterns; The early warning module is used to determine the wear level according to the wheel brake shoe wear law and the preset wear threshold, and to warn of brake shoe operation failure in real time according to the wear level.

6. A railway freight car brake shoe dynamic detection and intelligent early warning device, characterized in that: It includes trackside equipment, machine room equipment and a monitoring center; the trackside equipment includes a vehicle number antenna, array magnetic steel and detection boxes symmetrically arranged on both sides of the track, and the detection box is equipped with a binocular stereo vision sensor for high-definition imaging of wheel brake shoes; the vehicle number antenna is used to collect the model and number of passing trains; the array magnetic steel is used to trigger the binocular stereo vision sensor, and the monitoring center is used to detect brake shoe data for remote viewing and push alarm brake shoe information to user terminals in multiple locations; the machine room equipment provides power for the trackside equipment and the monitoring center.

7. The railway freight car brake shoe dynamic detection and intelligent early warning method according to claim 6 is characterized in that: The binocular stereo vision sensor is composed of two cameras, and the two cameras and the object to be measured form a triangle in space.

8. The railway freight car brake shoe dynamic detection and intelligent early warning device according to claim 7 is characterized in that: The camera is equipped with a light source device.

9. The railway freight car brake shoe dynamic detection and intelligent early warning device according to claim 6, characterized in that: A temperature sensor is arranged in the detection box, and the temperature sensor is used to measure the internal temperature of the box.

10. The railway freight car brake shoe dynamic detection and intelligent early warning device according to claim 6, characterized in that: The antenna is fastened to the sleeper by clamping the sleeper.

Images