Method and device for detecting and calibrating overall dimension of train wheel pair

Abstract

The invention discloses a train wheel pair contour dimension detection and calibration method and device, and belongs to the technical field of measuring instruments. In the device, a plurality of measuring positions and a synchronous triggering mechanism are introduced, so that the high-precision measurement of the geometric dimension of the wheel is ensured. The wheel detection module is used for sensing a signal that a wheel reaches a measurement position in real time, so that synchronous work of each vision measurement module and the image imaging module is ensured, and errors caused by inconsistent time sequences are effectively reduced. According to the design, the measurement stability can be kept under different environments and illumination conditions, error correction is carried out through the relation between the wheel rotation angle and the wheel diameter, the final wheel geometric dimension parameters are more accurate, and the reliability of wheel set dimension measurement is remarkably improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of rail transit safety detection and monitoring equipment, and in particular to a method and device for detecting and calibrating the outline dimensions of a train wheel pair. Background Art

[0002] The detection of train wheel set geometry is one of the important links in the safe operation of rail transit, and its detection results can directly reflect the degree of wear and operation status of train wheels. At present, most wheel set size detection equipment uses visual measurement technology. For example, in the related technology, a base is set on both sides of the track to install a beam, and a structured light vision sensor for measuring the geometric parameters of the train wheel set is installed on the beam. When the train passes, the structured light projector projects a light strip onto the train wheelset, and the camera captures the image of the light strip and calculates the cross-sectional contour data of the wheelset where the light strip is located based on the structured light vision model, and then calculates the geometric parameters of the wheelset.

[0003] However, the measurement performance of visual sensors is easily affected by the external environment, especially in the complex and harsh outdoor environment of railways. The high-speed movement of trains, vibration and impact, strong stray light, weak light, rain, snow, fog and other climatic conditions will interfere with the imaging quality of the sensor, resulting in a decrease in the measurement accuracy of the visual measurement system. In addition, stains or foreign matter attached to the wheel surface will also interfere with the imaging results, further affecting the stability and reliability of the measurement data. The combined effect of these factors may lead to increased measurement errors and have a certain impact on rail transit operation and maintenance.

[0004] When the measurement data obtained by the detection equipment has large errors, it may cause problems such as false alarms or missed alarms. False alarms will lead to unnecessary manual review and repeated inspections, resulting in a waste of resources, while missed alarms may fail to detect excessive wheel wear in time, posing a potential safety hazard. Therefore, how to improve the robustness and accuracy of the visual measurement system in complex environments and reduce the impact of environmental interference has become a key technical problem in the design and application of train wheel geometry detection equipment. Summary of the invention

[0005] The embodiment of the present application provides a method and device for detecting and calibrating the outline dimensions of a train wheel pair. The technical solution is as follows:

[0006] According to one aspect of the present application, a train wheel pair profile size detection device is provided, the device comprising: a visual measurement module for wheel pair size detection, an image imaging module for wheel angle detection, a wheel detection module, an electrical control module for wheel pair size measurement, and a data processing module;

[0007] The image imaging module corresponds to the visual measurement module and is used to detect the rotation angle of the measured wheel at the measurement position; the wheel detection module is installed at the upstream position of the visual measurement module in the travel direction of the train and is used to detect whether the wheel has arrived, wherein the measurement position is the position where the wheel triggers the photo taking during the actual measurement;

[0008] The electrical control module is connected to the visual measurement module, the image imaging module and the wheel detection module, and is used to start the visual measurement module and the image imaging module to perform image acquisition according to the train arrival signal sent by the wheel detection module, and is used to receive image acquisition information;

[0009] The data processing module is connected to the electrical control module and performs processing according to the image acquisition information.

[0010] Optionally, the image imaging module includes an imaging camera and a fill light, and the imaging camera supports triggered photography;

[0011] The imaging camera is facing the wheel at the triggering position of the visual measurement module, the imaging plane of the imaging camera is parallel to the wheel surface, the optical axis of the imaging camera is parallel to the wheel axle, and the field of view of the imaging camera includes the entire wheel and the upper edge of the track.

[0012] Optionally, the visual measurement module is provided with a structured light laser and a measurement camera, and in response to the visual measurement module being activated, the structured light laser is activated and used to determine whether the wheel reaches the measurement position and sends a measurement position arrival signal to the electrical control module;

[0013] The electrical control module is also used to control the measuring camera of the visual measurement module and the imaging camera of the image imaging module to collect images after receiving the measurement position arrival signal.

[0014] Optionally, the structured light laser adopts a line projection method.

[0015] Optionally, the wheel detection module is provided with a magnetic steel sensor, and the wheel detection module is used to sense the wheel through the magnetic steel sensor.

[0016] On the other hand, a method for detecting and calibrating the outline dimensions of a train wheel pair is provided, the method being used for the above-mentioned train wheel pair outline dimension detection device, the method comprising:

[0017] During offline deployment, determine the standard wheel diameter values ​​of different types of wheelsets included in the trains passing on the line where the device is located;

[0018] Calculate the position spacing of the visual measurement modules used for wheelset size detection according to the standard wheel diameter value and deploy the visual measurement modules according to the position spacing;

[0019] Place the wheel at the measurement position of N visual measurement modules, where N is a positive integer ≥ 2;

[0020] Mark a horizontal line through the wheel center on the side of the wheel;

[0021] Turn on the structured light laser of each visual measurement module;

[0022] Marking the center lines corresponding to the N laser lines according to the laser lines appearing on the side of the wheel, and extending them to intersect with the horizontal line passing through the wheel center;

[0023] The angle between the measured horizontal line and each laser line is recorded as the rotation angle correction value

[0024] Optionally, the method further includes:

[0025] During the online deployment process, in response to the wheel detection module continuously sensing at least two wheels, the train arrival signal is sent to the electrical control module;

[0026] Control each of the visual measurement modules to perform laser projection through the electrical control module;

[0027] The wheel detection module determines in real time whether the wheel reaches the detection position of the first group of visual measurement modules;

[0028] In response to the wheel reaching the detection position of the first group of visual measurement modules, the wheel detection module sends a train arrival signal to the electrical control module;

[0029] The electrical control module synchronously triggers the visual measurement module and the image imaging module of the first group of wheels to collect images;

[0030] In response to the wheel reaching the detection position of the second to N groups of visual measurement modules, the second to N groups of visual measurement modules and the image forming modules are synchronously triggered to collect images respectively;

[0031] The data processing module calculates the wheel rotation angle δ according to the image acquisition information corresponding to groups 1 to N i ;

[0032] According to the rotation angle correction value obtained in offline deployment, the wheel rotation angle at the current measurement position is corrected. After correction, the specific position of each wheel is obtained and described by the angle value as

[0033] The data processing module calculates the image acquisition information of 1 to N groups of visual measurement modules and obtains the position of each wheel θ. i The wheel set geometric dimension parameters include the wheel diameter value D i ;

[0034] The data processing module calculates the expected rotation angle when the wheel reaches the detection position of the second to Nth groups of visual measurement modules according to the wheel diameter value D1 corresponding to the first group of visual measurement modules and the distance between the first group of visual measurement modules and other groups of visual measurement modules, which is recorded as θ′2~θ′ N ,in,

[0035] Compare the wheel rotation angle θ at each measurement position j and the wheel rotation angle θ′ at each detection position j Get the deviation Δθ i =θ′ j -θ i , the deviation Δθ i =θ′ j -θ i Used to reflect the wheel diameter measurement error;

[0036] According to the deviation Δθ j =θ′ j -θ j Correct the wheel diameter value.

[0037] In an embodiment of the present application, a train wheel pair profile dimension detection device is provided that cooperates with a plurality of visual measurement modules and a wheel angle detection module. In this device, by introducing multiple measurement positions and a synchronous triggering mechanism, high-precision measurement of wheel geometry is ensured. The wheel detection module senses the signal of the wheel reaching the measurement position in real time, thereby ensuring the synchronization of each visual measurement module and the image imaging module, thereby effectively reducing the error caused by inconsistent timing. This design can maintain the stability of measurement under different environments and lighting conditions, and performs error correction through the relationship between the wheel rotation angle and the wheel diameter, so that the final wheel geometry parameters are more accurate, significantly improving the reliability of wheel pair dimension measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a structural schematic diagram of a train wheel pair outline dimension detection device provided by an illustrative embodiment of the present application;

[0039] Figure 2 is a schematic diagram of a structured light laser projected onto a wheel provided by an exemplary embodiment of the present application;

[0040] Figure 3It is a schematic diagram of the synchronous operation of a visual measurement module and an image forming module provided by an exemplary embodiment of the present application;

[0041] Figure 4 Schematic diagram of the measurement position spacing and rotation angle changes under different wheel diameter values ​​(840mm and 915mm);

[0042] Figure 5 Schematic diagram of the effect of wheel diameter measurement error on rotation angle estimation error;

[0043] Figure 6 is a schematic diagram of the characteristic area of ​​the axle end cap in the wheel image;

[0044] Figure 7 A schematic diagram of the character box and character horizontal lines in the shaft end cover image. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0046] The term "multiple" as used herein refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0047] Example 1

[0048] Please refer to Figure 1 , which shows a schematic diagram of the structure of a train wheel set profile size detection device provided by an exemplary embodiment of the present application. The device includes: a visual measurement module for wheel set size detection, an image imaging module for wheel angle detection, a wheel detection module, an electrical control module for wheel set size measurement, and a data processing module.

[0049] Among them, in the actual design, in order to measure the contour dimensions at different positions of the wheel for comparative correction, at least 2 sets of visual measurement modules for measuring the size of train wheels are included.

[0050] The image imaging module corresponds to the visual measurement module and is used to detect the rotation angle of the measured wheel at the measurement position; the wheel detection module is installed at the upstream position of the visual measurement module in the direction of train travel and is used to detect whether the wheel has arrived, wherein the measurement position is the position where the wheel triggers the photo taking during actual measurement.

[0051] The electrical control module is connected to the visual measurement module, the image imaging module and the wheel detection module, and is used to start the visual measurement module and the image imaging module to collect images according to the train arrival signal sent by the wheel detection module, and to receive image collection information.

[0052] The data processing module is connected to the electrical control module and processes the image acquisition information. The data processing module for operation and processing obtains the geometric dimension parameters of the wheelset corresponding to the angle of the wheel, evaluates the accuracy of the measurement value based on the established relationship between the distance between each measurement unit, the wheel diameter value and the wheel rotation angle, and corrects the wheel diameter value with a large measurement error.

[0053] Specifically, the wheel size detection electrical control module obtains the magnetic steel sensor signal to turn on the lasers of each wheel size measurement module in advance, and directly triggers the response signal of the detected wheel or delays the triggering of each wheel size detection visual measurement module camera and wheel angle detection image imaging module camera to collect images at the same time according to the vehicle speed calculated by multiple sets of magnetic steel signals. In order to ensure the accuracy of the wheel size measurement value and its corresponding wheel rotation angle, the same set of wheel size measurement modules and their corresponding wheel angle detection modules use the same external trigger to synchronously trigger the visual camera and the angle detection camera.

[0054] In the above device, the data processing module simultaneously receives the image data collected by the camera of each wheel pair size detection visual measurement module as the input of the train wheel contour structured light three-dimensional reconstruction calculation, calculates the geometric size parameters of the wheelset respectively, and calculates the specific position of the wheel at each size measurement (described by the angle value relative to the axle end feature) according to the wheel axle end feature recognition result according to the image data collected by the camera of each wheel angle detection image imaging module corresponding to each wheel pair size detection position received. The specific method is described below; the data processing module stores the calculation method established in advance according to the relationship between the distance between each measurement unit, the wheel diameter value and the wheel rotation angle, and estimates the angle rotated by the wheel to reach the next wheel pair size measurement module position based on the wheel diameter value measured by the previous wheel pair size measurement module, and compares it with the actual angle of the next wheel pair size measurement position to evaluate the accuracy of the measurement value and correct the wheel diameter value with a large measurement error. The specific calculation method is described in detail in the method embodiment.

[0055] Furthermore, the imaging module includes an imaging camera and a fill light, and the imaging camera supports triggered photography. The imaging camera is facing the wheel at the trigger position of the visual measurement module, the imaging plane of the imaging camera is parallel to the wheel surface, the optical axis of the imaging camera is parallel to the wheel axle, and the field of view of the imaging camera includes the entire wheel and the upper edge of the track.

[0056] Furthermore, the visual measurement module is provided with a structured light laser and a measurement camera. In response to the visual measurement module being activated, the structured light laser is activated and used to determine whether the wheel reaches the measurement position and sends a measurement position arrival signal to the electrical control module.

[0057] The electrical control module is also used to control the measuring camera of the visual measurement module and the imaging camera of the image imaging module to collect images after receiving the measurement position arrival signal.

[0058] Furthermore, the structured light laser adopts a line projection method. The wheelset size detection visual measurement module preferably adopts a structured light visual sensor, including a camera and a structured light laser. The structured light laser preferably projects a line projection method. When the train wheelset passes, the structured light laser projects at least one laser line light strip onto the wheel, forming a characteristic light strip on the outer surface of the wheel and the side of the wheel, such as Figure 2 The camera is controlled by a trigger pulse to capture the wheel image containing the laser light strip, which is used for structured light vision 3D measurement calculation to obtain the geometric size parameters of the wheels of the wheelset.

[0059] Further, the wheel detection module is provided with a magnetic steel sensor, and the wheel detection module is used to sense the wheel through the magnetic steel sensor. The wheel detection module is used to detect the arrival of the wheel and is set at a specified position near the visual measurement module. Preferably, a magnetic steel sensor is used to sense the wheel. Preferably, multiple groups of magnetic steel sensors can be used to determine the arrival of the train and detect the moment when each wheel pair reaches the sensing position. When the measurement position of the visual measurement module is not convenient for installing a magnetic steel sensor, the train speed can also be calculated by installing multiple magnetic steels with known distances between each other upstream of the detection position, and the triggering delay time is calculated according to the distance between the magnetic steel and the measurement position of the visual measurement module to trigger the camera. In addition to using magnetic steel sensors, other sensors such as photoelectric sensors can also be used to detect the arrival of the wheel. Preferably, an independent wheel detection module is installed near each set of wheel pair size measurement modules to ensure that each wheel pair size measurement visual camera is triggered accurately. If conditions do not allow, a set of wheel detection modules can also be installed to delay the triggering of the visual camera of each wheel pair size measurement module and the wheel angle detection image imaging module camera to take pictures according to the real-time measured vehicle speed.

[0060] The combination of wheel detection module and magnetic steel sensor can sense the arrival of the wheel at the measurement position in real time, ensuring that each measurement module is triggered synchronously at the same time, avoiding errors caused by asynchronous measurement time. Through this design, the system's measurement process is more accurate, reducing measurement deviations caused by asynchronous timing, thereby improving the reliability and accuracy of the overall measurement.

[0061] The embodiment of the present application provides a train wheel profile size detection device that works in coordination with a plurality of visual measurement modules and a wheel angle detection module. In this device, by introducing multiple measurement positions and a synchronous trigger mechanism, high-precision measurement of wheel geometry is ensured. The wheel detection module senses the signal of the wheel reaching the measurement position in real time, ensuring the synchronization of each visual measurement module and the image imaging module, thereby effectively reducing the error caused by inconsistent timing. The design can maintain the stability of measurement under different environments and lighting conditions, and performs error correction through the relationship between the wheel rotation angle and the wheel diameter, so that the final wheel geometry parameters are more accurate, significantly improving the reliability of wheelset size measurement.

[0062] The present application also provides a method for detecting and calibrating the outline dimensions of a train wheel pair, such as Figure 2 As shown, it includes two major parts: offline deployment and online deployment.

[0063] Example 2

[0064] Step 201, during offline deployment, determine the standard wheel diameter values ​​of different types of wheelsets included in the trains passing through the line where the device is located. For example, freight lines generally include two types of wheels: 840 mm and 915 mm.

[0065] Step 202 , calculating the position spacing of the visual measurement modules for wheelset size detection according to the standard wheel diameter value and deploying the visual measurement modules according to the position spacing.

[0066] It should be noted that, for simplification reasons, Figure 2 The wheelset size measurement module is the abbreviation of the visual measurement module used for wheelset size detection, and the wheel angle detection module is the abbreviation of the image imaging module used for wheel angle detection.

[0067] Preferably, a straight road section is selected as the device installation location.

[0068] The spacing calculation method is as follows:

[0069] Assuming that the distance between the two wheelset dimension measurement modules is L, the total angle that a wheel with a diameter of D rotates through this distance is:

[0070]

[0071] The modulo operation converts the relative angle within 360° to:

[0072]

[0073] In order to better distinguish the wheel rotation angle changes passing through the two wheel size measurement module positions, the α range should be avoided near 0° (360°). For different wheel diameter values, the α range is selected to be around 180°. Therefore, according to formula (1), D is set to 840mm and 915mm respectively, and L is used as the variable. The value range of α is as follows: Figure 4 As shown in the figure, it can be seen that the rotation angle values ​​α of wheel diameters of 840mm and 915mm are close when L is around 0 and 33000mm, that is, the rotation angle α of the same L value is not much different, while the difference is large when L is 15000mm in the middle, which easily leads to the situation that it is difficult to take both into account. Therefore, according to the requirement that different wheel diameters can simultaneously meet the rotation angle range of about 180°, the spacing L between modules is selected to be 33000mm, as shown in the figure. Figure 4 The square frame logo.

[0074] Step 203: placing the wheel at the measurement positions of N visual measurement modules, where N is a positive integer ≥2.

[0075] Step 204: Mark a horizontal line passing through the wheel center on the side of the wheel.

[0076] Step 205 , turning on the structured light laser of each vision measurement module.

[0077] Step 206: mark the center lines corresponding to the N laser lines on the side of the wheel, and extend them to intersect with the horizontal line passing through the wheel center.

[0078] At each position, the structural laser of each visual measurement module is turned on to mark the center line of the laser line projected on the side of the wheel, and then extended to intersect with the horizontal line passing through the center of the wheel.

[0079] Step 207, measure the angle between the horizontal line and each laser line and record it as the rotation angle correction value

[0080] Specifically, a target point is attached to a horizontal line passing through the wheel center, and a laser tracker is used to track the target point, measure the three-dimensional coordinate values ​​of each point, and then perform plane and straight line fitting, and then calculate the angle between the fitted horizontal line and each laser line, which is the rotation angle correction value.

[0081] Example 2 further ensures the accuracy of measurement by obtaining the standard wheel diameter value during offline deployment and optimizing the calculation of the measurement position spacing. In this example, the installation position of the visual measurement module is accurately calculated to avoid large deviations in the wheel rotation angle during the measurement process, especially when the wheel rotation angle is close to 0° and 360°. In addition, combined with the wheel detection module and the magnetic steel sensor, the system can accurately sense the moment when the wheel reaches the measurement position, ensure the synchronous triggering of each visual measurement module, reduce the measurement error, and effectively improve the accuracy of the measurement results.

[0082] Example 3

[0083] Continue as Figure 2 As shown, first the arrival of a train is detected to trigger the online deployment mechanism.

[0084] Step 301, during the online deployment process, in response to the wheel detection module continuously sensing at least two wheels, a train arrival signal is sent to the electrical control module.

[0085] Step 302: Control each visual measurement module to perform laser projection through the electrical control module.

[0086] Step 303: The wheel detection module determines in real time whether the wheel reaches the detection position of the first group of visual measurement modules.

[0087] In a possible implementation, it is directly judged or delayed to judge whether the wheel has reached the first group of visual measurement module shooting positions according to the wheel sensor signal. The wheel sensor signal is sent by the wheel detection module, and the wheel detection module is equipped with a wheel sensor for sensing whether the wheel has arrived, and the shooting position is also the detection position of the visual measurement module.

[0088] Step 304 , in response to the wheel arriving at the detection position of the first group of visual measurement modules, the wheel detection module sends a train arrival signal to the electrical control module.

[0089] Figure 2 In order to simplify the description, whether the first set of wheelset dimension measurement shooting positions are reached is equal to whether the first set of visual measurement module detection positions are reached.

[0090] Step 305: The electrical control module synchronously triggers the visual measurement module and the image imaging module to collect images of the first group of wheels passing by.

[0091] Figure 2In order to simplify the description, the first group of wheelset size visual measurement cameras and wheel angle detection module 1 camera are synchronously triggered to collect images, that is, the electrical control module synchronously triggers the visual measurement module and image imaging module of the first group of wheels to collect images, wherein the camera is a component for collecting images of specific modules, and the number "1" of the wheel angle detection module 1 indicates that it is the visual measurement module of the first group.

[0092] Step 306 , in response to the wheel reaching the detection position of the second to N groups of visual measurement modules, synchronously triggering the second to N groups of visual measurement modules and image forming modules to perform image acquisition.

[0093] After completing the first group of graphics acquisition, the acquisition process is continuously circulated, and the wheel sensor signal is continuously directly judged or delayed to determine whether the wheel has reached the Nth group of visual measurement module shooting position. If yes, it is determined whether the wheel has reached the Nth group of wheelset size measurement shooting position (i.e., the measurement shooting position of the Nth group of visual detection module). If yes, the Nth group of wheelset size visual measurement camera and the wheel angle detection module N camera are synchronously triggered to acquire images. Otherwise, it continues to directly judge or delay to determine whether the wheel has reached the Nth group of visual measurement module shooting position according to the wheel sensor signal.

[0094] Step 307: The data processing module calculates the wheel rotation angle δ according to the image acquisition information corresponding to groups 1 to N. i .

[0095] When it comes to step 307, the image collection of N groups has been completed. Further, in the data processing module, the wheel rotation angle δ at each wheel pair size measurement position 1 to N is calculated according to the camera images of each wheel angle detection module 1 to N. i .

[0096] Step 308: correct the wheel rotation angle at the current measurement position according to the rotation angle correction value obtained in the offline deployment. After correction, the specific position of each wheel is obtained and described by the angle value as

[0097] Furthermore, the wheel rotation angle δ at the measurement position of each wheel pair size from 1 to N is obtained. i After that, the wheel rotation angle at the current measurement position is corrected according to the rotation angle correction value obtained in offline deployment. After correction, the specific position of each wheel is obtained and described by the angle value as That is, the specific position of the wheel when measuring each dimension is obtained, which can also be simplified as θ1~θ N .

[0098] After step 308, it is the turn of the data processing module to analyze and correct the data.

[0099] Step 309: The data processing module calculates the image acquisition information of the 1 to N groups of visual measurement modules to obtain the image at each wheel position θ. i The geometric dimension parameters of the wheelset include the wheel diameter value D i .

[0100] like Figure 2 As shown in the figure, the data processing module calculates the wheelset geometric parameters at the corresponding positions of the wheels according to the visual camera images of the wheelset size measurement modules of groups 1 to N, where the wheel diameter values ​​are D1 to D N , that is, D i , i=1,2,...,N.

[0101] Step 310: The data processing module calculates the estimated rotation angle when the wheel reaches the detection position of the second to Nth groups of visual measurement modules according to the wheel diameter value D1 corresponding to the first group of visual measurement modules and the distance between the first group of visual measurement modules and other groups of visual measurement modules, which is recorded as θ′2 to θ′ N ,in,

[0102] Step 311, compare the wheel rotation angle θ at each measurement position j and the wheel rotation angle θ′ at each detection position j Get the deviation Δθ j =θ′ j -θ j .

[0103] like Figure 2 As shown, the wheel rotation angle θ at each measurement position is compared j and the wheel rotation angle θ′ at each detection position j That is, the comparison of θ2~θ N and θ′2~θ′ N The deviation between .

[0104] Step 312, based on the deviation Δθ j =θ′ j -θ j Correct the wheel diameter value.

[0105] In order to evaluate the error of the wheel diameter measurement value D1, the wheel rotation angle estimation value θ′ is theoretically calculated. j According to the error transfer calculation method, the above formula (1) can be used to obtain the influence of

[0106] like Figure 5 It indicates the error size of the estimated rotation angle caused by the different sizes of the wheel diameter measurement error ΔD1, where L is 33000mm, D1 = 840mm, 915mm, and Figure 5It can be seen that a small error in the wheel diameter measurement (e.g., 1-2 mm error) can cause a significant error of about 5°-10° in the estimated rotation angle. j When the angle is measured accurately, Δθ j To reflect whether the wheel diameter measurement of the wheelset is accurate. It should be noted that the deviation The numerical value is theoretically derived to prove that the angle deviation can better reflect the wheel diameter measurement error. It does not need to be calculated every time in specific applications.

[0107] According to the deviation Δθ j =θ′ j -θ j The specific method of correcting the wheel diameter value can be determined by judging Δθ j Whether it is greater than the set threshold value Trd1 (for example, 10°) is used to determine whether the wheel diameter measurement value of the wheelset has a large error. If it is greater than the threshold and less than the second threshold value Trd2 (for example, 15°), it is considered that the error is too large and the process goes to step 11 for correction; if it is greater than the second threshold value Trd2, it is considered that the error is too large and the entire wheelset dimension measurement parameter is set to invalid; if it is less than or equal to the set threshold value Trd1, it is considered that the wheelset dimension measurement value error is small and no correction is required. When the number of wheelset dimension measurement modules N is greater than 2, Δθ j The average value can be calculated and then judged.

[0108] When it is determined that the error of the first set of measurement values ​​is too large and D1 is set to invalid, if the number N of wheelset dimension measurement modules of the device is greater than 2, the same method as above can be used to evaluate the second set of wheelset dimension measurement values ​​D2, and so on.

[0109] Step 11: Rotate at an angle θ′ j For reference, the wheel diameter values ​​with large measurement errors are corrected.

[0110] like Figure 2 As shown in the figure, for the wheel diameter value with large measurement error, that is, when the deviation is greater than the threshold, the rotation angle θ′ j (i.e. θ′2~θ′ N ) is used as a reference to correct the deviation of the wheel diameter value, and finally the wheelset size measurement is completed with a clear wheel position, otherwise it is completed.

[0111] Correct the wheel diameter value with large measurement error. The specific correction method is as follows:

[0112] Calculate the rotation angle deviation by taking the wheel diameter values ​​measured for each wheel pair except the first set as a reference If Δθ′ n If it is less than the set threshold value Trd3 (for example, 2°), the wheel diameter correction value is calculated as follows:

[0113]

[0114] If Δθ′ n If it is less than the set threshold value Trd3, then j=3 and recalculate Δθ′ n If it is less than the threshold value Trd3, D is calculated according to the above formula. If it is always greater than the threshold value, no correction is performed.

[0115] Embodiment 3 further enhances the measurement accuracy. Through the precise control of the electrical control module and the real-time feedback of the wheel detection module, the synchronous triggering of each visual measurement module and the image imaging module is ensured. The data processing module is based on the image acquisition information of multiple measurement positions, combined with the correction value of the wheel rotation angle and the calculation of the geometric dimension parameters, and can effectively correct the wheel diameter value deviation caused by the error. This method can correct the measurement data through error transmission and angle deviation analysis when the wheel diameter value measurement error is large, thereby significantly improving the stability and reliability of the overall wheel size measurement.

[0116] Example 4

[0117] In addition, a method for calculating the wheel angle by the data processing module is also provided.

[0118] Above Figure 2 The image taken by the camera facing the wheel is as follows Figure 6 As shown, the shaft end cover on the wheel axle is fixed, as shown in FIG. Figure 6 In the area shown in the red box, the rotation angle of the wheel is determined by detecting and identifying the features of the wheel axle.

[0119] In one possible implementation, the characters on the axle end cover sign are used as features, a method for detecting and identifying characters is adopted, the rotation angle of the characters is used as the wheel rotation angle, and the horizontal direction is defined as 0° and counterclockwise rotation is defined as a positive angle. The specific steps are as follows.

[0120] Step 1: In Figure 6 Detect the axle end cap in the wheel image and obtain the axle end cap image. This can be achieved using existing deep learning detection networks, such as the YOLO series network, SSD network, etc.

[0121] Step 2: Detect characters in the shaft end cover image and obtain the following Figure 7 The character box shown in the dotted line can be implemented using existing deep learning character detection networks, such as DBNet, EAST, TextBoxes++ and other networks.

[0122] Step 3: Get the center horizontal line of each character box as the character horizontal line.

[0123] Step 4: Calculate the angle between the character horizontal direction and the image horizontal direction as the character rotation angle, such as Figure 5As shown in . The character rotation angle is the wheel rotation angle. Counterclockwise is a positive angle.

[0124] Among them, in step one, in order to improve the angle calculation accuracy, the wheel image can be distorted and corrected to meet the conditions that the camera is facing the wheel and the rail is in a horizontal state in the image. Specifically, the Zhang Zhengyou calibration method can be adopted. After the camera is fixed, the track is used as the horizontal direction of the world coordinate system, and targets are placed multiple times at the position where the wheel appears to obtain multiple calibration images, and the internal and external parameters of the camera are obtained. The wheel image is corrected according to the internal and external parameters of the camera obtained by calibration. This method is well known and will not be described in detail.

[0125] Example 4 provides a new method for the accurate calculation of wheel angles, which determines the rotation angle of the wheel by detecting the character features on the axle end cap. This method uses a deep learning algorithm to recognize characters and calculate the rotation angle with high accuracy. By accurately measuring the wheel rotation angle and coordinating other measurement parameters, the accuracy of wheel size measurement can be effectively improved, and the overall measurement deviation caused by angle recognition errors can be reduced. The application of this method further enhances the system's ability to process wheel angles and provides reliable wheel angle data support.

[0126] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A train wheel pair profile size detection device, characterized in that: The device comprises: a visual measurement module for wheel set size detection, an image imaging module for wheel angle detection, a wheel detection module, an electrical control module for wheel set size measurement, and a data processing module; The image imaging module corresponds to the visual measurement module and is used to detect the rotation angle of the measured wheel at the measurement position; the wheel detection module is installed at the upstream position of the visual measurement module in the travel direction of the train and is used to detect whether the wheel has arrived, wherein the measurement position is the position where the wheel triggers the photo taking during the actual measurement; The electrical control module is connected to the visual measurement module, the image imaging module and the wheel detection module, and is used to start the visual measurement module and the image imaging module to perform image acquisition according to the train arrival signal sent by the wheel detection module, and is used to receive image acquisition information; The data processing module is connected to the electrical control module and performs processing according to the image acquisition information.

2. The device according to claim 1, characterized in that The image imaging module includes an imaging camera and a fill light, and the imaging camera supports trigger-type photography; The imaging camera is facing the wheel at the triggering position of the visual measurement module, the imaging plane of the imaging camera is parallel to the wheel surface, the optical axis of the imaging camera is parallel to the wheel axle, and the field of view of the imaging camera includes the entire wheel and the upper edge of the track.

3. The device according to claim 2, characterized in that The visual measurement module is provided with a structured light laser and a measurement camera. In response to the visual measurement module being activated, the structured light laser is activated and used to determine whether the wheel reaches the measurement position and send a measurement position arrival signal to the electrical control module; The electrical control module is also used to control the measuring camera of the visual measurement module and the imaging camera of the image imaging module to collect images after receiving the measurement position arrival signal.

4. The device according to claim 3, characterized in that The structured light laser adopts a line projection method.

5. The device according to claim 1, characterized in that The wheel detection module is provided with a magnetic steel sensor, and the wheel detection module is used to sense the wheel through the magnetic steel sensor.

6. A method for detecting and calibrating the outline dimensions of a train wheel pair, characterized in that: The method is used for the train wheel pair profile size detection device according to any one of claims 1 to 5, and the method comprises: During offline deployment, determine the standard wheel diameter values ​​of different types of wheelsets included in the trains passing on the line where the device is located; Calculate the position spacing of the visual measurement modules used for wheelset size detection according to the standard wheel diameter value and deploy the visual measurement modules according to the position spacing; Place the wheel at the measurement position of N visual measurement modules, where N is a positive integer ≥ 2; Mark a horizontal line through the wheel center on the side of the wheel; Turn on the structured light laser of each visual measurement module; Marking the center lines corresponding to the N laser lines according to the laser lines appearing on the side of the wheel, and extending them to intersect with the horizontal line passing through the wheel center; The angle between the measured horizontal line and each laser line is recorded as the rotation angle correction value 7. The method according to claim 6, characterized in that The method further comprises: During the online deployment process, in response to the wheel detection module continuously sensing at least two wheels, the train arrival signal is sent to the electrical control module; Control each of the visual measurement modules to perform laser projection through the electrical control module; The wheel detection module determines in real time whether the wheel reaches the detection position of the first group of visual measurement modules; In response to the wheel reaching the detection position of the first group of visual measurement modules, the wheel detection module sends a train arrival signal to the electrical control module; The electrical control module synchronously triggers the visual measurement module and the image imaging module of the first group of wheels to collect images; In response to the wheel reaching the detection position of the second to N groups of visual measurement modules, the second to N groups of visual measurement modules and the image forming modules are synchronously triggered to collect images respectively; The data processing module calculates the wheel rotation angle δ according to the image acquisition information corresponding to groups 1 to N i ; According to the rotation angle correction value obtained in offline deployment, the wheel rotation angle at the current measurement position is corrected. After correction, the specific position of each wheel is obtained and described by the angle value as i=1, 2, ..., N.

8. The method according to claim 7, characterized in that The method further comprises: The data processing module calculates the image acquisition information of 1 to N groups of visual measurement modules and obtains the position of each wheel θ. i The wheel set geometric dimension parameters include the wheel diameter value D i ; The data processing module calculates the expected rotation angle when the wheel reaches the detection position of the second to Nth groups of visual measurement modules according to the wheel diameter value D1 corresponding to the first group of visual measurement modules and the distance between the first group of visual measurement modules and other groups of visual measurement modules, which is recorded as θ′2~θ′ N ,in, j=2, ..., N; Compare the wheel rotation angle θ at each measurement position j and the wheel rotation angle θ′ at each detection position j Get the deviation Δθ j =θ′ j -θ j , the deviation Δθ j =θ′ j -θ j Used to reflect the wheel diameter measurement error; According to the deviation Δθ j =θ′ j -θ j Correct the wheel diameter value.

Images