Methods, devices, terminals, and storage media for measuring wheel tread and out-of-roundness.

Abstract

This invention provides a method, apparatus, terminal, and storage medium for measuring wheel tread and out-of-roundness. The method includes: acquiring real-time images from a first acquisition device and a second acquisition device; performing noise filtering on the real-time images to obtain denoised images; cropping the denoised images to obtain cropped images containing key areas; performing three-dimensional reconstruction on the cropped images to obtain a three-dimensional reconstruction model; obtaining the minimum horizontal distance between the first acquisition device and the tread of the wheel under test; and determining the tread dimensions and out-of-roundness of the wheel under test based on the minimum horizontal distance and the three-dimensional reconstruction model. This invention improves the efficiency of wheel tread and out-of-roundness measurement.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to a method, device, terminal, and storage medium for measuring wheel tread and out-of-roundness. Background Technology

[0002] Defects and out-of-roundness in train wheel treads can cause track vibrations, affecting train operation safety. Therefore, it is necessary to conduct regular inspections of train wheel treads and out-of-roundness to ensure safe train operation.

[0003] In existing technologies, laser spacing measurement is typically used to obtain the tread dimensions of train wheels when inspecting the tread surface. However, this method only acquires tread dimension data at a single location on the wheel; to obtain tread dimension data at other locations, multiple measurements are required. Similarly, when detecting the out-of-roundness of train wheels, sensors are typically used to collect displacement and circumferential distance data at a single location on the wheel tread in real time. This method only acquires the out-of-roundness at a fixed location on the wheel tread; to observe the out-of-roundness at multiple locations on the wheel tread, multiple measurements are necessary.

[0004] In summary, existing technologies for measuring out-of-roundness and tread surface are performed separately, requiring individual measurements. Furthermore, multiple measurements are needed to determine the overall out-of-roundness and tread surface dimensions of the train wheel. Therefore, existing out-of-roundness and tread surface measurement technologies are cumbersome and inefficient. Summary of the Invention

[0005] This invention provides a method, device, terminal, and storage medium for measuring wheel tread and out-of-roundness, in order to solve the problems of cumbersome operation and low measurement efficiency in existing out-of-roundness and tread measurement technologies.

[0006] In a first aspect, embodiments of the present invention provide a method for measuring wheel tread and out-of-roundness, wherein the wheel under test is in a rotating state; a first acquisition device is disposed directly in front of the current tread of the wheel under test; at least one second acquisition device is disposed on each of the left and right sides of the current tread of the wheel under test, the field of view of the first acquisition device and the field of view of the at least one second acquisition device together cover the current tread of the wheel under test, and the at least one second acquisition device and the first acquisition device are located on the same horizontal plane; the measurement method includes:

[0007] The real-time images acquired by the first acquisition device and the second acquisition device are acquired respectively;

[0008] The real-time image is subjected to noise filtering to obtain a denoised image;

[0009] The denoised image is cropped to obtain a cropped image containing the key regions;

[0010] The cropped image is reconstructed in three dimensions to obtain a three-dimensional reconstructed model;

[0011] The minimum horizontal distance between the first acquisition device and the tread of the wheel under test is obtained, and the tread size and out-of-roundness of the wheel under test are determined based on the minimum horizontal distance and the three-dimensional reconstruction model.

[0012] In one possible implementation, the step of performing noise filtering on the real-time image to obtain a denoised image includes:

[0013] The RGB values ​​of each pixel in the real-time image are obtained respectively, and the overexposure noise points in the real-time image are determined based on the RGB values;

[0014] Each overexposure noise point is centered on a filter window, and the corresponding RGB mean matrix is ​​calculated for each filter window.

[0015] The RGB values ​​of each overexposed noise point are updated according to the RGB mean matrix to obtain the denoised image.

[0016] In one possible implementation, the step of updating the RGB values ​​of each overexposed noise point according to the RGB mean matrix to obtain the denoised image includes:

[0017] For each overexposed noise point, the RGB matrix corresponding to each pixel in the filter window corresponding to the overexposed noise point is obtained, and the Euclidean distance between the RGB matrix corresponding to each pixel and the RGB mean matrix is ​​calculated.

[0018] Update the RGB value of the pixel corresponding to the minimum Euclidean distance to the RGB value of the overexposed noise point;

[0019] The RGB values ​​of all overexposed noise points are updated to obtain the denoised image.

[0020] In one possible implementation, the step of performing three-dimensional reconstruction on the cropped image to obtain a three-dimensional reconstruction model includes:

[0021] Feature points are extracted from each cropped image, and image matching is performed based on the feature points to obtain multiple sets of image connection maps;

[0022] From the multiple sets of image connection diagrams, determine an optimal set of image connection diagrams;

[0023] The optimal image connectivity graph is used to estimate the image pose.

[0024] Perform three-dimensional reconstruction on the optimal image connection map to obtain the three-dimensional coordinates of each matching point in the optimal image connection map;

[0025] The image pose and the three-dimensional coordinates of each matching point are optimized to obtain a three-dimensional reconstructed image, and the three-dimensional reconstructed image is added to the previous three-dimensional reconstructed image;

[0026] From the remaining multiple sets of image connection diagrams, a new set of optimal image connection diagrams is determined, and the process jumps to the step of "performing pose estimation on the optimal image connection diagram to obtain image pose" to continue execution until all 3D reconstructed images are added to the previous 3D reconstructed image to obtain a 3D reconstructed model.

[0027] In one possible implementation, the tread dimensions and out-of-roundness of the wheel under test are determined based on the minimum horizontal distance and the three-dimensional reconstruction model, including:

[0028] Based on the minimum horizontal distance, the unit pixel distance in the 3D reconstruction model is calculated;

[0029] The number of pixels corresponding to the tread of the wheel under test in the three-dimensional reconstruction model is obtained, and the tread size and out-of-roundness of the wheel under test are determined based on the number of pixels and the unit pixel distance.

[0030] In one possible implementation, calculating the unit pixel distance in the 3D reconstruction model based on the minimum horizontal distance includes:

[0031] Calculate the true height of the 3D reconstructed model based on the minimum horizontal distance;

[0032] Based on the actual height and the corresponding number of pixels, the unit pixel distance in the 3D reconstruction model is determined.

[0033] In one possible implementation, determining the true height corresponding to the 3D reconstructed model based on the minimum horizontal distance includes:

[0034] according to Determine the true height corresponding to the three-dimensional reconstruction model;

[0035] Where f represents the focal length of the lens of the first acquisition device, L represents the minimum horizontal distance, h represents the height of the lens target surface of the first acquisition device, and H represents the actual height.

[0036] Determining the unit pixel distance in the 3D reconstruction model based on the actual height and its corresponding number of pixels includes:

[0037] according to Determine the unit pixel distance in the 3D reconstruction model;

[0038] Where p represents the unit pixel distance, and n represents the number of pixels corresponding to the actual height.

[0039] Secondly, embodiments of the present invention provide a measuring device for wheel tread and out-of-roundness, comprising:

[0040] The acquisition module is used to acquire real-time images acquired by the first acquisition device and the second acquisition device, respectively.

[0041] The reconstruction module is used to perform noise filtering on the real-time image to obtain a denoised image;

[0042] The reconstruction module is also used to crop the denoised image to obtain a cropped image containing the key regions;

[0043] The reconstruction module is also used to perform three-dimensional reconstruction on the cropped image to obtain a three-dimensional reconstruction model;

[0044] The measurement module is used to obtain the minimum horizontal distance between the first acquisition device and the tread of the wheel under test, and to determine the tread size and out-of-roundness of the wheel under test based on the minimum horizontal distance and the three-dimensional reconstruction model.

[0045] Thirdly, embodiments of the present invention provide a terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method as described in the first aspect or any possible implementation thereof.

[0046] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method as described in the first aspect or any possible implementation thereof.

[0047] This invention provides a method, apparatus, terminal, and storage medium for measuring wheel tread and out-of-roundness. By acquiring images of the wheel under test and performing 3D reconstruction, a 3D reconstructed model containing the actual tread dimensions of the wheel is obtained. Based on this 3D reconstructed model and the minimum horizontal distance between the first acquisition device and the tread of the wheel under test, the actual tread dimensions and out-of-roundness of the wheel can be determined. This eliminates the need for repeated measurements, effectively improving measurement efficiency and accuracy. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a left view of the positional relationship between the first data acquisition device and the wheel to be tested, provided in an embodiment of the present invention.

[0050] Figure 2 This is a top view of the positional relationship between the first data acquisition device and the wheel to be tested, provided in an embodiment of the present invention.

[0051] Figure 3 This is a top view showing the positional relationship between the first data acquisition device, the second data acquisition device, and the wheel to be measured, as provided in an embodiment of the present invention.

[0052] Figure 4 This is a schematic diagram showing the positions of the first acquisition device, the second acquisition device, and the third acquisition device provided in an embodiment of the present invention;

[0053] Figure 5 This is a top view showing the positional relationship between the first acquisition device, the second acquisition device, and the third acquisition device provided in an embodiment of the present invention;

[0054] Figure 6 This is a flowchart illustrating the implementation of the wheel tread and out-of-roundness measurement method provided in this embodiment of the invention.

[0055] Figure 7 This is a flowchart illustrating the implementation of noise filtering of real-time images to obtain denoised images, provided by an embodiment of the present invention.

[0056] Figure 8 This is a flowchart illustrating the implementation of three-dimensional reconstruction of a cropped image to obtain a three-dimensional reconstruction model, provided by an embodiment of the present invention.

[0057] Figure 9 This is a schematic diagram of the triangular relationship between the spatial point and the optical center of each acquisition device provided in the embodiments of the present invention;

[0058] Figure 10 This is a schematic diagram of the structure of the wheel tread and out-of-roundness measuring device provided in an embodiment of the present invention;

[0059] Figure 11 This is a schematic diagram of the terminal provided in an embodiment of the present invention. Detailed Implementation

[0060] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0061] To make the objectives, technical solutions, and advantages of the present invention clearer, specific embodiments will be described below in conjunction with the accompanying drawings.

[0062] Before measuring the tread defects and out-of-roundness of the wheel under test, acquisition equipment can be set up around the wheel to capture real-time images of the wheel's tread, thereby enabling the measurement of wheel tread defects and out-of-roundness. See details. Figure 1 , Figure 2 and Figure 3 The wheel under test is positioned at a fixed location and is rotating. To prevent wheel vibration from affecting the acquisition results of each acquisition device, a support device can be used to suspend the wheel above the ground. This prevents the acquisition devices from shaking due to wheel vibration, which would affect the image clarity of each acquisition device.

[0063] A first acquisition device 11 is positioned directly in front of the current tread of the wheel under test. The first acquisition device 11 captures real-time images of the current tread of the wheel under test. These tread images are subsequently used for 3D reconstruction to generate a 3D reconstructed model of the wheel under test. To ensure the completeness of the subsequent 3D reconstruction, at least one second acquisition device 12 can be positioned on the left and right sides of the current tread of the wheel under test. The field of view of the first acquisition device 11 and the field of view of at least one second acquisition device 12 together cover the current tread of the wheel under test. The first acquisition device 11 and the second acquisition device 12 are used to capture the same tread of the wheel under test from different shooting angles, so as to facilitate the subsequent establishment of a complete wheel model and avoid missing wheel tread features. Specifically, the first acquisition device is positioned directly in front of the current tread of the wheel under test to maximize the acquisition of tread features, while the second acquisition devices are positioned on the left and right sides to acquire tread edge features. To ensure that each acquisition device can capture the same tread of the wheel under test from different angles, the first acquisition device 11 and the second acquisition device 12 can be positioned on the same horizontal plane.

[0064] See Figure 4 and Figure 5To ensure that the first and second acquisition devices are on the same horizontal plane and that their fields of view jointly cover the current tread surface of the wheel under test, they can be mounted on the same chassis. The chassis is equipped with adjustable brackets with adjustable angles and spacing. The second acquisition devices are mounted on these brackets. By adjusting the angles and spacing of the brackets, the positions of each second acquisition device can be adjusted, ensuring that each second acquisition device can cooperate with the first acquisition device to capture the current tread surface of the wheel under test. Furthermore, to facilitate subsequent image processing, the device parameters of the first and second acquisition devices can be kept consistent, thus ensuring consistent image parameters for each real-time image. Here, image parameters refer to image resolution and image size, etc.

[0065] It should be noted that, Figures 1-5 The accompanying drawings are provided as examples only. This is not intended to limit the number of second acquisition devices. Users can set the number of second acquisition devices according to their own needs. This embodiment of the invention does not impose specific limitations in this regard.

[0066] Based on the hardware setup described above, multiple images of the tread of the wheel under test can be acquired. Based on these multiple tread images, a wheel model can be reconstructed to measure the tread surface and out-of-roundness of the wheel under test.

[0067] Figure 6 The implementation flowchart of the wheel tread and out-of-roundness measurement method provided in the embodiments of the present invention is described in detail below:

[0068] Step 601: Acquire the real-time images acquired by the first acquisition device and the second acquisition device, respectively.

[0069] The wheel under test is rotating. The first and second acquisition devices capture real-time images of the wheel's tread surface from different angles, obtaining real-time images. These real-time images are used for subsequent 3D reconstruction to generate a 3D reconstructed model containing the actual tread surface dimensions of the wheel under test.

[0070] Step 602: Perform noise filtering on the real-time image to obtain a denoised image.

[0071] Because the tread of the wheel under test may have bright areas, the light reflected from these bright areas can impact the exposure of the acquisition device, causing overexposed pixels. These overexposed pixels appear as white spots in the real-time image, distorting the image and affecting the accuracy of subsequent measurements. Therefore, this embodiment of the invention pre-processes the real-time image with noise filtering to remove overexposed noise points and improve image clarity.

[0072] In some embodiments, see Figure 7Step 602 may include:

[0073] Step 621: Obtain the RGB values ​​of each pixel in the real-time image, and determine the overexposure noise points in the real-time image based on the RGB values.

[0074] Iterate through each pixel in the real-time image and obtain the RGB value of each pixel. Pixels whose RGB values ​​all exceed a first preset value are identified as overexposed pixels. For example, the first preset value can be any value between 245 and 255. For example, 250.

[0075] Step 622: Set up a filter window centered on each overexposure noise point, and calculate the RGB mean matrix corresponding to each filter window.

[0076] The size and shape of the filtering window can be set by the user, and this embodiment of the invention does not impose specific limitations on this. For example, this embodiment uses a square area as the filtering window. Meanwhile, to improve the noise filtering effect, the side length of the filtering window can be set to be larger. For example, the side length of the filtering window can be an odd number greater than 5. For example, the side length of the filtering window can be 7.

[0077] For all pixels within the filtering window, excluding overexposed noise points, calculate the average R-value, average G-value, and average B-value for all pixels. These three averages together constitute the RGB mean matrix corresponding to the filtering window. The RGB mean matrix can be represented as [R... J G J B J ]. Among them, R J G represents the average R value of all pixels. J G represents the average G value of all pixels. J This represents the average B value of all pixels.

[0078] Step 623: Update the RGB values ​​of each overexposed noise point according to the RGB mean matrix to obtain the denoised image.

[0079] When updating each overexposure noise point, the following steps can be followed:

[0080] For each overexposed noise point, obtain the RGB matrix corresponding to each pixel in the filter window corresponding to that overexposed noise point, and calculate the Euclidean distance between the RGB matrix corresponding to each pixel and the RGB mean matrix.

[0081] The RGB matrix corresponding to each pixel within the filtering window can be represented as [R ij G ij Bij ], where R ij G represents the R value of the pixel in the i-th row and j-th column. ij The G value and B value represent the pixel value in the i-th row and j-th column. ij This represents the B value of the pixel in the i-th row and j-th column. According to... The Euclidean distance between the RGB matrix and the RGB mean matrix corresponding to each pixel can be calculated separately. d represents the Euclidean distance.

[0082] Update the RGB value of the pixel corresponding to the minimum Euclidean distance to the RGB value of the overexposed noise point.

[0083] The RGB values ​​of all overexposed noise points are updated to obtain the denoised image.

[0084] By removing overexposed noise points, white spots in real-time images can be effectively removed, thereby improving the clarity of real-time images and preventing them from affecting the accuracy of subsequent measurements.

[0085] Step 603: Crop the denoised image to obtain a cropped image containing the key areas.

[0086] To reduce data volume and facilitate subsequent 3D reconstruction processing, the denoised image can be pre-cropped to obtain a cropped image containing the key regions. During 3D reconstruction, only the cropped image containing the key regions needs to be reconstructed.

[0087] When identifying key areas, the denoised image can be displayed in real time on a monitor, allowing users to select key areas within the denoised image themselves.

[0088] As described above, the denoised image includes a first image corresponding to the first acquisition device and a second image corresponding to the second acquisition device. The display shows either the first image or the second image in real time. The user selects a region within the displayed first or second image. In response to the user's selection, the first and second images are cropped to obtain a cropped image containing the key region. It should be noted that the region selected by the user is the key region.

[0089] The first and second images have the same size. After the user selects a region in either the first or second image using a selection box, the selected region can be assigned to all the noise-reduced images for cropping. For example, the display shows the first image, and the user inputs a rectangular selection box to select the center of the first image. The center of the first image can then be cropped based on this rectangular selection box to obtain a cropped image. Correspondingly, a rectangular selection box is also generated at the center of the second image, and the center of the second image is cropped based on this rectangular selection box to obtain a cropped image.

[0090] By cropping each denoised image, the amount of 3D reconstruction data can be effectively reduced, which facilitates the improvement of 3D reconstruction efficiency.

[0091] Step 604: Perform 3D reconstruction on the cropped image to obtain a 3D reconstruction model.

[0092] The wheel under test is rotating, which means the cropped images contain different tread surfaces of the wheel. By performing 3D reconstruction on all cropped images, multiple 2D wheel tread images can be reconstructed to generate a 3D reconstructed model of the wheel. This 3D reconstructed model corresponds to the actual dimensions of the wheel under test. Based on this 3D reconstructed model, the tread dimensions and out-of-roundness of the wheel under test can be measured.

[0093] In some embodiments, see Figure 8 When performing 3D reconstruction on a cropped image, the following steps can be followed:

[0094] Step 641: Extract feature points from each cropped image and perform image matching based on the feature points to obtain multiple sets of image connection maps.

[0095] When extracting feature points from each cropped image, algorithms such as Scale-invariant Feature Transform (SIFT) or Oriented FAST and Rotated BRIEF (ORB) can be used. The embodiments of this invention are not specifically limited. Exemplarily, the embodiments of this invention use the SIFT algorithm to extract feature points from each cropped image.

[0096] The process iterates through the feature points in all cropped images and calculates the distance between feature points in different cropped images to determine matching points and thus the matching relationships between them. This results in multiple sets of image connectivity maps. To avoid incorrect matches, the Random Sample Consensus (RANSAC) algorithm can be used to calculate the fundamental matrix between each set of image connectivity maps, thereby eliminating unsuitable connectivity maps and forming the final image connectivity map.

[0097] Step 642: Determine an optimal set of image connection graphs from multiple sets of image connection graphs.

[0098] When determining the optimal image connectivity map, an image connectivity map with a large number of matching points can be selected to obtain a larger sparse point cloud. However, since the disparity angle is proportional to the baseline length, a baseline length that is too small will introduce significant errors when determining the 3D coordinates of each feature point; conversely, a baseline length that is too large will result in less overlap and fewer matching points between image connectivity maps. Therefore, an image connectivity map with a large number of matching points and a moderate disparity angle can be selected as the optimal image connectivity map.

[0099] In 3D reconstruction, the selection of the first set of optimal image connectivity maps is crucial for reconstruction quality. For example, in this embodiment of the invention, image connectivity maps with more than 100 matching points and a parallax angle greater than 5° are selected as the first set of optimal image connectivity maps.

[0100] Step 643: Perform pose estimation on the optimal image connectivity graph to obtain the image pose.

[0101] In this embodiment of the invention, the Perspective-n-Point algorithm is used for pose estimation of the optimal image connectivity graph. Users may also choose other algorithms for pose estimation, and this embodiment of the invention does not specifically limit this choice.

[0102] Step 644: Perform 3D reconstruction on the optimal image connection graph to obtain the 3D coordinates of each matching point in the optimal image connection graph.

[0103] In this embodiment of the invention, a first acquisition device and two second acquisition devices are provided. In the world coordinate system, the three-dimensional coordinates of spatial point P are X, and corresponding to the three acquisition devices, the matching points projected onto the optimal image connection diagram of spatial point P are p1, p2, and p3, respectively. 2-1 and p 2-2 Where p1 corresponds to the first data acquisition device. 2-1 This corresponds to the first and second acquisition devices. 2-2 Corresponding to the second acquisition device. p1, p 2-1 and p 2-2The corresponding normalized coordinates in their respective acquisition device coordinate systems are x1 and x2, respectively. 2-1 and x 2-2 The optical centers of the first acquisition device and the two second acquisition devices are O1 and O2, respectively. 2-1 and O 2-2 The depths from point P to the imaging planes of each acquisition device are s1, s2, and s3, respectively. 2-1 and s 2-2 T1 is O1 and O 2-1 The translation matrix between O1 and O2, T2 is the translation matrix between O1 and O2. 2-2 The translation matrices between them. R1 is the rotation matrix of the first second acquisition device relative to the first acquisition device, and R2 is the rotation matrix of the second acquisition device relative to the first acquisition device. According to the definition of stacked geometry:

[0104] s1x1=s 2-1 R1x 2-1 +T1 (1)

[0105] s1x1=s 2-2 R2x 2-2 +T2 (2)

[0106] Multiply both sides of the above formulas (1) and (2) by an antisymmetric matrix. The following formula is obtained:

[0107]

[0108]

[0109] The depths s1 and s2 from point P to the imaging plane of each acquisition device can be calculated using formulas (3) and (4). 2-1 s 2-2 .

[0110] The coordinates of p1 in the world coordinate system are s1x1. 2-1 The coordinates in the world coordinate system are s 2-1 x 2-1 p 2-2 The coordinates in the world coordinate system are s 2-2 x 2-2 .

[0111] A fixed triangular relationship is formed between the optical centers of any two acquisition devices and the spatial point P. For example, the optical centers of the first acquisition device, the first and second acquisition devices, and the spatial point P form a triangle as follows: Figure 9 The triangular relationship is shown. Accordingly, matching points p1 and p2 are also shown. 2-1 This also satisfies the triangular relationship. Therefore, based on the coordinates of the optical center and the matching point, the coordinates of point P can be calculated accordingly.

[0112] According to p1, p 2-1 , O1 and O 2-1 From the coordinate values ​​of p1 and p2, the coordinates of the spatial point P determined by the first and second acquisition devices can be calculated. Similarly, based on p1 and p2, the coordinates of the spatial point P can be calculated. 2-2 , O1 and O 2-2 The coordinates of the spatial point P determined by the first acquisition device and the second acquisition device can be calculated from the coordinates of the first acquisition device and the second acquisition device.

[0113] The final coordinates of spatial point P are determined by calculating the average of the two coordinates. It should be noted that these final coordinates are two-dimensional coordinates, which, together with the depth value s1, constitute the three-dimensional coordinates X of spatial point P.

[0114] It is understandable that the coordinates of spatial point P determined solely by the first acquisition device and the first and second acquisition devices, or solely by the first acquisition device and the second acquisition device, may contain deviations. Therefore, this embodiment of the invention improves the accuracy of the coordinates of spatial point P by calculating the average value.

[0115] Step 645: Optimize the image pose and the 3D coordinates of each matching point to obtain a 3D reconstructed image, and add the 3D reconstructed image to the previous 3D reconstructed image.

[0116] The local bundle adjustment (BA) algorithm can be used to optimize the image pose and 3D coordinates, resulting in a set of 3D reconstructed images. These reconstructed images are then added to the previous set. In other words, each time a set of 3D reconstructed images is generated, it is added to the previous set, ultimately ensuring that all the reconstructed images together constitute the 3D reconstructed model.

[0117] Step 646: From the remaining multiple sets of image connection diagrams, determine a new set of optimal image connection diagrams and jump to step 643 to continue execution until all 3D reconstructed images are added to the previous 3D reconstructed image to obtain the 3D reconstructed model.

[0118] Based on the number of matching points and the disparity angle in each group of image connection maps, 3D reconstruction is performed on each group of image connection maps in sequence to generate corresponding 3D reconstructed images. All 3D reconstructed images together constitute a 3D reconstruction model.

[0119] It should be noted that if the previous 3D reconstructed images contain errors, each additional 3D reconstructed image will accumulate errors, potentially leading to drift in the reconstruction results. Therefore, after adding all the 3D reconstructed images, a global BA algorithm can be used to optimize the image pose and the 3D coordinates of the matching points. Furthermore, some outliers can be removed based on the reprojection error to ensure the reliability of the 3D reconstruction.

[0120] Step 605: Obtain the minimum horizontal distance between the first acquisition device and the tread of the wheel to be tested, and determine the tread size and out-of-roundness of the wheel to be tested based on the minimum horizontal distance and the three-dimensional reconstruction model.

[0121] See Figure 1 The minimum horizontal distance L between the first data acquisition device and the tread of the wheel under test is used. Based on the minimum horizontal distance and the device parameters of the first data acquisition device, the dimensional ratio between the 3D reconstructed model and the wheel under test can be obtained. Therefore, the actual tread dimensions and out-of-roundness of the wheel under test can be determined based on the 3D reconstructed model.

[0122] In some embodiments, determining the tread size and out-of-roundness of the wheel under test based on the minimum horizontal distance and the three-dimensional reconstruction model may include:

[0123] Calculate the unit pixel distance in the 3D reconstruction model based on the minimum horizontal distance.

[0124] Obtain the number of pixels corresponding to the tread of the wheel under test in the 3D reconstruction model, and determine the tread size and out-of-roundness of the wheel under test based on the number of pixels and the unit pixel distance.

[0125] Based on the minimum horizontal distance and the device parameters of the first acquisition device, the unit pixel distance in the 3D reconstruction model can be calculated, that is, the actual length corresponding to one pixel in the 3D reconstruction model. By obtaining the number of pixels corresponding to the tread in the 3D reconstruction model, the tread size of the wheel under test can be obtained. The product of the unit pixel distance and the number of pixels mentioned above is the actual tread size of the wheel under test. Based on the actual tread size, the out-of-roundness of the wheel under test can be obtained.

[0126] In some embodiments, calculating the unit pixel distance in the 3D reconstructed model based on the minimum horizontal distance may include:

[0127] Calculate the true height of the 3D reconstructed model based on the minimum horizontal distance.

[0128] according to The true height corresponding to the 3D reconstructed model can be determined;

[0129] Where f represents the focal length of the lens of the first acquisition device, L represents the minimum horizontal distance, h represents the height of the lens target surface of the first acquisition device, and H represents the actual height.

[0130] The true height of the 3D reconstructed model can be calculated using the formula above. Based on this, the number of pixels along the height direction of the 3D reconstructed model can be obtained to calculate the unit pixel distance within the model.

[0131] Based on the actual height and the corresponding number of pixels, the unit pixel distance in the 3D reconstruction model is determined.

[0132] according to It is possible to determine the unit pixel distance in the 3D reconstruction model.

[0133] Where p represents the distance per unit pixel, and n represents the number of pixels corresponding to the actual height.

[0134] Compared with the prior art, the beneficial effects of the embodiments of the present invention are as follows:

[0135] This invention involves acquiring images of the wheel under test and performing 3D reconstruction to obtain a 3D reconstructed model containing the actual tread dimensions of the wheel. Based on this 3D reconstructed model and the minimum horizontal distance between the first acquisition device and the tread of the wheel under test, the actual tread dimensions and out-of-roundness of the wheel can be determined. This eliminates the need for repeated measurements, effectively improving measurement efficiency and accuracy.

[0136] Furthermore, this embodiment of the invention also performs noise filtering on the real-time images to improve their clarity and thus enhance measurement accuracy. Additionally, this embodiment of the invention performs image cropping before 3D reconstruction to reduce data volume and improve 3D reconstruction efficiency.

[0137] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0138] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0139] Figure 10 A schematic diagram of the wheel tread and out-of-roundness measuring device provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below:

[0140] like Figure 10As shown, the wheel tread and out-of-roundness measuring device 10 includes: an acquisition module 101, a reconstruction module 102, and a measurement module 103.

[0141] The acquisition module 101 is used to acquire real-time images acquired by the first acquisition device and the second acquisition device, respectively.

[0142] The reconstruction module 102 is used to perform noise filtering on the real-time image to obtain a denoised image.

[0143] The reconstruction module 102 is also used to crop the denoised image to obtain a cropped image containing the key regions.

[0144] The reconstruction module 102 is also used to perform three-dimensional reconstruction on the cropped image to obtain a three-dimensional reconstruction model.

[0145] The measurement module 103 is used to obtain the minimum horizontal distance between the first acquisition device and the tread of the wheel under test, and to determine the tread size and out-of-roundness of the wheel under test based on the minimum horizontal distance and the three-dimensional reconstruction model.

[0146] In one possible implementation, the reconstruction module 102 is used to acquire the RGB values ​​of each pixel in the real-time image and determine the overexposure noise points in the real-time image based on the RGB values.

[0147] The reconstruction module 102 is also used to set up filter windows centered on each overexposure noise point and to calculate the RGB mean matrix corresponding to each filter window.

[0148] The reconstruction module 102 is also used to update the RGB values ​​of each overexposed noise point according to the RGB mean matrix to obtain the denoised image.

[0149] In one possible implementation, the reconstruction module 102 is used to obtain the RGB matrix corresponding to each pixel in the filter window corresponding to each overexposed noise point, and calculate the Euclidean distance between the RGB matrix corresponding to each pixel and the RGB mean matrix.

[0150] The reconstruction module 102 is also used to update the RGB value of the pixel corresponding to the minimum Euclidean distance to the RGB value of the overexposed noise point.

[0151] The reconstruction module 102 is also used to update the RGB values ​​of all overexposed noise points to obtain a denoised image.

[0152] In one possible implementation, the reconstruction module 102 is used to extract feature points from each cropped image and perform image matching based on the feature points to obtain multiple sets of image connection maps.

[0153] The reconstruction module 102 is also used to determine an optimal set of image connection maps from multiple sets of image connection maps.

[0154] The reconstruction module 102 is also used to estimate the pose of the optimal image connectivity graph to obtain the image pose.

[0155] The reconstruction module 102 is also used to perform three-dimensional reconstruction of the optimal image connection map to obtain the three-dimensional coordinates of each matching point in the optimal image connection map.

[0156] The reconstruction module 102 is also used to optimize the image pose and the three-dimensional coordinates of each matching point to obtain a three-dimensional reconstructed image, and add the three-dimensional reconstructed image to the previous three-dimensional reconstructed image.

[0157] The reconstruction module 102 is also used to re-determine a set of optimal image connection maps from the remaining multiple sets of image connection maps, and jump to the step of "performing pose estimation on the optimal image connection map to obtain image pose" to continue execution until all three-dimensional reconstructed images are added to the previous three-dimensional reconstructed image to obtain a three-dimensional reconstructed model.

[0158] In one possible implementation, the measurement module 103 is used to calculate the unit pixel distance in the 3D reconstruction model based on the minimum horizontal distance;

[0159] The measurement module 103 is also used to obtain the number of pixels corresponding to the tread of the wheel under test in the three-dimensional reconstruction model, and to determine the tread size and out-of-roundness of the wheel under test based on the number of pixels and the unit pixel distance.

[0160] In one possible implementation, the measurement module 103 is also used to calculate the true height corresponding to the three-dimensional reconstruction model based on the minimum horizontal distance.

[0161] The measurement module 103 is also used to determine the unit pixel distance in the 3D reconstruction model based on the true height and the number of pixels it corresponds to.

[0162] In one possible implementation, the measurement module 103 is further configured to, based on Determine the true height corresponding to the 3D reconstruction model;

[0163] Where f represents the focal length of the lens of the first acquisition device, L represents the minimum horizontal distance, h represents the height of the lens target surface of the first acquisition device, and H represents the actual height.

[0164] Measurement module 103 is also used for, according to Determine the unit pixel distance in the 3D reconstruction model;

[0165] Where p represents the distance per unit pixel, and n represents the number of pixels corresponding to the actual height.

[0166] Compared with the prior art, the beneficial effects of the embodiments of the present invention are as follows:

[0167] The acquisition module 101 acquires images of the wheel under test, and the reconstruction module 102 performs three-dimensional reconstruction on the acquired images to obtain a three-dimensional reconstruction model containing the actual tread dimensions of the wheel under test. Based on this three-dimensional reconstruction model and the minimum horizontal distance between the first acquisition device and the tread of the wheel under test, the measurement module 103 determines the actual tread dimensions and out-of-roundness of the wheel under test. This eliminates the need for repeated measurements, effectively improving measurement efficiency and accuracy.

[0168] In addition, the reconstruction module 102 performs noise filtering on the real-time image to improve its clarity and thus enhance measurement accuracy. Furthermore, before performing 3D reconstruction, the reconstruction module 102 pre-crops the image to reduce the amount of data and improve the efficiency of 3D reconstruction.

[0169] Figure 11 This is a schematic diagram of a terminal provided in an embodiment of the present invention. For example... Figure 11 As shown, the terminal in this embodiment includes: a processor 110, a memory 111, and a computer program 112 stored in the memory 111 and executable on the processor 110. When the processor 110 executes the computer program 112, it implements the steps in the various wheel tread and out-of-roundness measurement method embodiments described above, for example... Figure 6 Steps 601 to 605 are shown. Alternatively, when the processor 110 executes the computer program 112, it implements the functions of each module / unit in the above-described device embodiments, for example... Figure 10 The functions of modules 101 to 103 are shown.

[0170] For example, the computer program 112 can be divided into one or more modules / units, which are stored in the memory 111 and executed by the processor 110 to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 112 in the terminal 11. For example, the computer program 112 can be divided into... Figure 11 Modules 111 to 113 are shown.

[0171] The terminal can be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal may include, but is not limited to, a processor 110 and a memory 111. Those skilled in the art will understand that... Figure 11This is merely an example of a terminal and does not constitute a limitation on the terminal. It may include more or fewer components than shown, or combine certain components, or different components. For example, the terminal may also include input / output devices, network access devices, buses, etc.

[0172] The processor 110 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0173] The memory 111 can be an internal storage unit of the terminal, such as a hard drive or memory. The memory 111 can also be an external storage device of the terminal, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory 111 can include both internal and external storage units. The memory 111 is used to store the computer program and other programs and data required by the terminal. The memory 111 can also be used to temporarily store data that has been output or will be output.

[0174] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0175] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0176] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0177] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0178] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0179] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0180] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above embodiments of the wheel tread and out-of-roundness measurement methods. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for measuring wheel tread and out-of-roundness, characterized in that, The wheel under test is rotating; a first acquisition device is positioned directly in front of the current tread of the wheel under test; at least one second acquisition device is positioned on each of the left and right sides of the current tread of the wheel under test, the field of view of the first acquisition device and the field of view of the at least one second acquisition device together cover the current tread of the wheel under test, and the at least one second acquisition device and the first acquisition device are located on the same horizontal plane; the measurement method includes: The real-time images acquired by the first acquisition device and the second acquisition device are acquired respectively; The real-time image is subjected to noise filtering to obtain a denoised image; The denoised image is cropped to obtain a cropped image containing the key regions; The cropped image is reconstructed in three dimensions to obtain a three-dimensional reconstructed model; The minimum horizontal distance between the first acquisition device and the tread of the wheel under test is obtained, and the tread size and out-of-roundness of the wheel under test are determined based on the minimum horizontal distance and the three-dimensional reconstruction model. The step of performing three-dimensional reconstruction on the cropped image to obtain a three-dimensional reconstruction model includes: Feature points are extracted from each cropped image, and image matching is performed based on the feature points to obtain multiple sets of image connection maps; From the multiple sets of image connection diagrams, determine an optimal set of image connection diagrams; The optimal image connectivity graph is used to estimate the image pose. Perform three-dimensional reconstruction on the optimal image connection map to obtain the three-dimensional coordinates of each matching point in the optimal image connection map; The image pose and the three-dimensional coordinates of each matching point are optimized to obtain a three-dimensional reconstructed image, and the three-dimensional reconstructed image is added to the previous three-dimensional reconstructed image; From the remaining multiple sets of image connection diagrams, a new set of optimal image connection diagrams is determined, and the process jumps to the step of "performing pose estimation on the optimal image connection diagram to obtain image pose" to continue execution until all 3D reconstructed images are added to the previous 3D reconstructed image to obtain a 3D reconstructed model. Based on the minimum horizontal distance and the three-dimensional reconstruction model, the tread dimensions and out-of-roundness of the wheel under test are determined, including: Based on the minimum horizontal distance, the unit pixel distance in the 3D reconstruction model is calculated; The number of pixels corresponding to the tread of the wheel under test in the three-dimensional reconstruction model is obtained, and the tread size and out-of-roundness of the wheel under test are determined based on the number of pixels and the unit pixel distance. The calculation of the unit pixel distance in the 3D reconstruction model based on the minimum horizontal distance includes: Calculate the true height of the 3D reconstructed model based on the minimum horizontal distance; Based on the actual height and the number of pixels it corresponds to, the unit pixel distance in the 3D reconstruction model is determined; Determining the true height of the 3D reconstructed model based on the minimum horizontal distance includes: according to Determine the true height corresponding to the three-dimensional reconstruction model; in, This indicates the focal length of the lens of the first acquisition device. This represents the minimum horizontal distance. This indicates the height of the lens target surface of the first acquisition device. This indicates the actual height; Determining the unit pixel distance in the 3D reconstruction model based on the actual height and its corresponding number of pixels includes: according to Determine the unit pixel distance in the 3D reconstruction model; in, This represents the unit pixel distance. This indicates the number of pixels corresponding to the actual height.

2. The method for measuring wheel tread and out-of-roundness according to claim 1, characterized in that, The step of performing noise filtering on the real-time image to obtain a denoised image includes: The RGB values ​​of each pixel in the real-time image are obtained respectively, and the overexposure noise points in the real-time image are determined based on the RGB values; Each overexposure noise point is centered on a filter window, and the corresponding RGB mean matrix is ​​calculated for each filter window. The RGB values ​​of each overexposed noise point are updated according to the RGB mean matrix to obtain the denoised image.

3. The method for measuring wheel tread and out-of-roundness according to claim 2, characterized in that, The step of updating the RGB values ​​of each overexposed noise point according to the RGB mean matrix to obtain the denoised image includes: For each overexposed noise point, the RGB matrix corresponding to each pixel in the filter window corresponding to the overexposed noise point is obtained, and the Euclidean distance between the RGB matrix corresponding to each pixel and the RGB mean matrix is ​​calculated. Update the RGB value of the pixel corresponding to the minimum Euclidean distance to the RGB value of the overexposed noise point; The RGB values ​​of all overexposed noise points are updated to obtain the denoised image.

4. A device for measuring wheel tread and out-of-roundness, characterized in that, The apparatus used in the method for measuring wheel tread and out-of-roundness according to any one of claims 1-3, the apparatus comprising: The acquisition module is used to acquire real-time images acquired by the first acquisition device and the second acquisition device, respectively. The reconstruction module is used to perform noise filtering on the real-time image to obtain a denoised image; The reconstruction module is also used to crop the denoised image to obtain a cropped image containing the key regions; The reconstruction module is also used to perform three-dimensional reconstruction on the cropped image to obtain a three-dimensional reconstruction model; The measurement module is used to obtain the minimum horizontal distance between the first acquisition device and the tread of the wheel under test, and to determine the tread size and out-of-roundness of the wheel under test based on the minimum horizontal distance and the three-dimensional reconstruction model.

5. A terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the wheel tread and out-of-roundness measurement method as described in any one of claims 1 to 3.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for measuring wheel tread and out-of-roundness as described in any one of claims 1 to 3.

Images