An image processing method, a wheel size dynamic detection method, device and system

Abstract

The application discloses an image processing method, a wheel size dynamic detection method, device and system. The image processing method comprises the following steps: acquiring images of more than two inner light cut lines and images of more than two outer light cut lines on a measured wheel, wherein the inner light cut line and the outer light cut line are respectively formed by a line light source for detecting the inner and outer profiles of the measured wheel, and the inner profile line light source and the outer profile line light source are different wave bands; obtaining a wheel inner reference surface according to each inner light cut line; restoring the inner and outer light cut lines to standard cross-section inner and outer light cut lines respectively according to the wheel inner reference surface; respectively splicing the restored standard cross-section inner and outer light cut lines to form a complete profile; and comparing the complete profile with a standard profile to obtain each tread profile size information of the measured wheel. The application has the advantages of simple implementation method, low processing complexity, high detection efficiency and precision, etc.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rail transit vehicle technology, and in particular to an image processing method, a wheel size dynamic detection method, device and system. BACKGROUND

[0002] The wheel set of a rail transit vehicle is the direct contact part of the vehicle and the rail, and is thus an important component for ensuring the safety and comfort of train operation. During train operation, the wheel set shape size changes due to impact and wear between the wheel and the rail, which affects the contact stress and contact state between the wheel set and the rail, and further affects the safety and comfort of train operation. Therefore, it is urgently needed to automatically, intelligently and accurately detect the wheel shape size during intelligent operation and maintenance of the vehicle.

[0003] For wheel set size detection, at present, at least three groups of line light sources (such as lasers) of the same waveband are usually used to irradiate the wheel by a trackside wheel set size online detection system, to form at least three groups of complete light section lines, and the size of the wheel is determined after a camera collects the at least three groups of complete light section lines. For example, patent application CN105651168A discloses a method for measuring the shape size of a wheel, which uses at least three line light sources to irradiate the wheel, and calculates the wheel's axis line by shooting a complete light section line along the wheel width on the wheel, and then determines the shape size parameters of the wheel. However, the above scheme has the following problems:

[0004] 1. Multiple groups of line light sources of the same waveband are used at the same time, and each group of light sources interferes with each other, making the image processing of multiple light sources difficult, and easily causing interference due to factors such as reflection.

[0005] 2. Complete light section line profiles need to be obtained, but complete light section lines are actually difficult to obtain, and the inner and outer laser profiles are also difficult to adjust to completely coincide, and are also easily affected by vibration and other reasons, so that the detection accuracy is reduced due to the branching of the use. SUMMARY

[0006] The technical problem to be solved by the present application is that, in view of the technical problems existing in the prior art, the present application provides an image processing method and device which is simple to implement, can realize dynamic detection of wheel size, has low processing complexity, high efficiency and high accuracy, and a wheel size dynamic detection method and system which does not require complete light section lines, has high detection efficiency and detection accuracy, and is strong in anti-interference.

[0007] To solve the above technical problems, the technical solution provided by the present application is as follows:

[0008] An image processing method for dynamic detection of wheel size, comprising the following steps:

[0009] The images of more than two inner light sections and the images of more than two outer light sections on the measured wheel are acquired, the inner light sections and the outer light sections are respectively formed by line light sources of different wave bands for detecting the inner profile and the outer profile of the measured wheel.

[0010] The inner reference surface of the wheel is obtained according to each of the inner light sections;

[0011] The inner and outer light sections are respectively restored to standard cross-section inner and outer light sections according to the inner reference surface of the wheel;

[0012] The restored standard cross-section inner and outer light sections are respectively spliced to form a complete profile;

[0013] The complete profile is compared with a standard profile to obtain the profile size information of each tread of the measured wheel.

[0014] Further, the restoration of the inner and outer light sections to the standard cross-section inner and outer light sections according to the inner reference surface of the wheel comprises:

[0015] The required wheel parameters are obtained according to the inner reference surface of the wheel to calculate the relative positions of the inner and outer profile light sections and the measured wheel;

[0016] The restoration transformation parameters are calculated according to the relative positions of the inner and outer profile light sections and the measured wheel;

[0017] The inner and outer light sections are restored to the standard cross-section inner and outer light sections by using the restoration transformation parameters.

[0018] Further, the wheel parameters include the rolling circle diameter, the wheel center distance between the rolling circle axis and the inner and outer profile light sections, and the included angle between the inner reference surface of the wheel and the inner and outer profile light sections.

[0019] Further, the determination of the rolling circle axis comprises: acquiring a plurality of target points obtained by intersecting a rolling circle plane with each light section; fitting a target plane from each target point, and obtaining the rolling circle axis according to the intersection of the axis line and the target plane; and obtaining the rolling circle diameter according to the distance from the rolling circle axis to each target point.

[0020] Further, the determination of the axis line comprises:

[0021] The non-worn part of the detected profile is searched by translating the inner reference surface of the wheel to obtain the intersection with the profile;

[0022] The searched intersection is used to fit a plurality of axes in sequence;

[0023] The axis of the wheel set is fitted using the fitted plurality of axes.

[0024] Further, the axis of the wheel set is fitted using a robust straight line fitting algorithm.

[0025] Further, the reduction transformation parameter is an affine transformation coefficient, the affine transformation coefficient of the inner and outer profiles is calculated according to the relative position of the inner and outer profile light sections and the measured wheel, the inner and outer light sections are corrected by affine transformation using the affine transformation coefficient, and the reduction of the standard section inner and outer light sections is realized.

[0026] Further, the reduced standard section inner light section and the reduced standard section outer light section are spliced to form the complete profile using a point cloud registration algorithm of the nearest point iteration.

[0027] Further, the comparison of the complete profile with the standard profile to obtain the profile size information of each tread of the measured wheel includes: the distance from the base point B1 to the highest point of the rim top to obtain the rim height, the distance from the rim thick point at the L2 point above the base point B1 to the inner side to obtain the rim thickness, the horizontal distance between the lowest qR value point at the L3 point above the base point B1 and the highest qR value point at the L4 point below the rim top to obtain the qR value, and the vertical distance between the standard base point B and the base point B1 to obtain the tread wear, wherein the base point B1 is the base point of the detected profile, L1-L4 are the plane positions of the four light source emitting devices arranged on both sides of the track, L1 and L3 are arranged on the outside of the track, and L2 and L4 are arranged on the inside of the track.

[0028] A dynamic wheel size detection method, the steps comprising:

[0029] At least two or more line light sources are emitted to the left and right sides of the measured wheel respectively for detecting the inner and outer profiles of the wheel respectively, wherein the line light sources for detecting the inner profile and the line light sources for detecting the outer profile are line light sources of different wavebands;

[0030] Images of two or more inner light sections and images of two or more outer light sections on the measured wheel are collected;

[0031] The collected images are processed according to the above image processing method, and the profile size information of each tread of the measured wheel is output.

[0032] An image processing device for dynamic wheel size detection, comprising:

[0033] The light section line acquisition module is configured to acquire images of two or more inner side light section lines and images of two or more outer side light section lines on the measured wheel, the inner side light section line and the outer side light section line being formed by a line light source of different wave bands for detecting the inner side profile and the outer side profile of the measured wheel.

[0034] The inner side reference surface determination module is configured to determine the inner side reference surface of the wheel according to each of the inner side light section lines.

[0035] The light section line restoration module is configured to correspondingly restore the inner and outer side light section lines to standard section inner and outer side light section lines according to the inner side reference surface of the wheel.

[0036] The light section line splicing module is configured to correspondingly splice the restored standard section inner and outer side light section lines to form a complete profile.

[0037] The profile size determination module is configured to compare the complete profile with a standard profile to obtain profile size information of each tread of the measured wheel.

[0038] An image processing device for dynamic detection of wheel size, comprising a processor and a memory, the memory being configured to store a computer program, and the processor being configured to execute the computer program to perform the above method.

[0039] A wheel size dynamic detection system, comprising:

[0040] The light source emission device is configured to emit at least two or more line light sources to the left and right sides of the measured wheel for detecting the inner and outer side profiles of the wheel, wherein the line light source for detecting the inner side profile and the line light source for detecting the outer side profile are line light sources of different wave bands.

[0041] The image acquisition device is configured to acquire images of two or more inner side light section lines and images of two or more outer side light section lines on the measured wheel.

[0042] The image processing device described above is configured to acquire detection images of the measured wheel for image processing and output profile size information of each tread of the measured wheel.

[0043] Further, the light source emission device comprises at least four groups of light source emission devices, at least two groups of light source emission devices are arranged on one side of the track for detecting the inner side profile of the measured wheel, and at least two groups of light source emission devices are arranged on the other side of the track for detecting the outer side profile of the measured wheel, the light source emission devices on one side of the track and the light source emission devices on the other side of the track are arranged in a staggered manner in the horizontal position to form a staggered arrangement.

[0044] Further, the image acquisition device comprises four or more imaging devices for acquiring the linear light source emitted by each group of light source emitting devices, and each imaging device is provided with a filter for filtering the linear light source of a specified wave band.

[0045] Further, the image acquisition device comprises four or more imaging devices for acquiring the linear light source emitted by each group of light source emitting devices, and each imaging device is provided with a filter for filtering the linear light source of a specified wave band.

[0046] A computer readable storage medium storing a computer program, the computer program being executed to implement the method as described above.

[0047] Compared with the prior art, the present application has the advantages that: the present application adopts different wave band linear light sources to irradiate the measured wheel to detect the inner side contour and the outer side contour of the measured wheel respectively, respectively images the different wave band linear light sources, acquires the images of the inner side light section line and the outer side light section line on the measured wheel, respectively and independently processes the images of the inner side light section line and the outer side light section line, can completely separate the inner side and the outer side laser contour, reduces the interference of the same wave band light source, can greatly reduce the difficulty of image processing, thereby greatly improving the image processing speed; at the same time, the detected inner side light section line is restored to a standard section light section line, and the complete contour is spliced, without the need of a complete light section line contour, without the need of adjusting the inner and outer sides of the laser to coincide, can greatly reduce the requirement for the completeness of the light section line, improve the application flexibility, can also avoid the influence of interference factors on the detection accuracy, and effectively improve the detection accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 is the implementation flow schematic diagram of the image processing method for the wheel size dynamic detection of the embodiment 1 of the present application.

[0049] Figure 2 is the arrangement principle schematic diagram of the dynamic detection system in the embodiment 1 of the present application.

[0050] Figure 3 is the structure principle schematic diagram of the imager in the embodiment 1 of the present application.

[0051] Figure 4 is the principle schematic diagram of the implementation of the different wave band multi-light source dynamic detection in the embodiment 1 of the present application.

[0052] Figure 5 is the effect schematic diagram of the light section line and the inner side reference surface obtained in the embodiment 1 of the present application.

[0053] Figure 6 is the principle schematic diagram of the implementation of the rolling circle diameter calculation in the embodiment 1 of the present application.

[0054] Figure 7is a schematic diagram of the principle of obtaining the complete tread contour and size information in the embodiment of the present application.

[0055] Figure 8 is a schematic diagram of the implementation process of the dynamic detection of the wheel size in the embodiment 2 of the present application. DETAILED DESCRIPTION

[0056] The present application is further described below in conjunction with the accompanying drawings and specific preferred embodiments, but the protection scope of the present application is not limited by this.

[0057] Embodiment 1:

[0058] As shown in the figure, the steps of the image processing method for dynamic detection of the wheel size in the present embodiment include: Figure 1

[0059] S01. Obtain the images of two or more inner light cut lines and the images of two or more outer light cut lines on the measured wheel, the inner light cut line and the outer light cut line are respectively formed by the line light source for detecting the inner contour and the outer contour of the measured wheel, wherein the line light source for detecting the inner contour and the line light source for detecting the outer contour are line light sources of different wave bands;

[0060] S02. Obtain the inner reference surface of the wheel according to each inner light cut line;

[0061] S03. Correspondingly restore the inner and outer light cut lines to the standard cross section inner and outer light cut lines according to the inner reference surface of the wheel;

[0062] S04. Correspondingly splice the restored standard cross section inner and outer light cut lines to form a complete contour;

[0063] S05. Compare the complete contour with the standard contour to obtain the size information of each tread contour of the measured wheel.

[0064] ​The embodiment pre-irradiates the measured wheel with different waveband linear light sources (such as visible light + infrared light) to respectively detect the inner profile and the outer profile of the measured wheel. Since the wavebands of the inner and outer profiles are different, an image containing only the inner / outer light strip can be formed, that is, the different waveband linear light sources are respectively imaged alone. After the inner light cut line and the outer light cut line images on the measured wheel are acquired, the inner light cut line and the outer light cut line images are respectively processed independently, the inner and outer laser profiles can be completely separated, the interference of the same waveband light source can be reduced, the difficulty of image processing can be greatly reduced, and thus the image processing speed can be greatly improved. Meanwhile, the detected inner light cut line is restored to a standard cross-section light cut line, and a complete profile is formed by splicing, so that the complete light cut line profile is not required, the inner and outer laser light sources do not need to be adjusted to coincide, the requirement for the completeness of the light cut line can be greatly reduced, the application flexibility is improved, the influence of interference factors on the detection accuracy can be avoided, and the detection accuracy is effectively improved.

[0065] In the embodiment, a light source emitting device is arranged beside the track to emit at least two waveband linear light sources to the measured wheel, and an image acquisition device is arranged to acquire the inner and outer light cut line images formed by the linear light sources on the measured wheel. The acquired images are processed by using the above image processing method to realize dynamic detection of the wheel size.

[0066] The above light source emitting device includes at least four groups of light source emitting devices, at least two groups of light source emitting devices are arranged on one side of the track to detect the inner profile of the measured wheel, and at least two groups of light source emitting devices are arranged on the other side of the track to detect the outer profile of the measured wheel. The light source emitting devices on one side of the track and the light source emitting devices on the other side of the track are arranged in a staggered manner in the horizontal position to form a staggered position. The image acquisition device includes four or more imaging devices for respectively acquiring the linear light sources emitted by each group of light source emitting devices. Each imaging device is further provided with a filter to filter the specified waveband linear light source, so that each imaging device can independently acquire the image formed by one waveband linear light source, and thus the inner and outer profiles can be separated. Further, a trigger device connected with the image acquisition device is arranged to control the triggering of the imaging devices to simultaneously perform image acquisition when the wheels of the train pass.

[0067] In a specific application embodiment, as Figure 2As shown, four sets of lasers L1-L4, four imagers C1-C4, and a trigger T1 are respectively set on both sides of the track. L1 and L3 emit red light to detect the outer contour, while L2 and L3 emit infrared light to detect the inner contour. L1 and L3 are arranged on one side of the track, and L2 and L4 are arranged on the other side. L1 and L3, and L2 and L4 are symmetrical about trigger T1, while L1 and L2, and L3 and L4 are not symmetrical about the track, creating a misalignment by spacing them horizontally. The four imagers C1-C4 are correspondingly arranged near the four sets of lasers L1-L4 to acquire images of the light cross-sections formed by the corresponding lasers on the wheel. Using this configuration, bidirectional detection of vehicle wheel dimensions can be achieved. The specific positions of the four sets of lasers L1-L4 and the four imagers C1-C4 can be adjusted appropriately according to actual needs; for example, strict symmetry about the trigger is not necessary.

[0068] Each imager C1 to C4, as follows Figure 3 As shown, each imager, in addition to the industrial camera and lens, is equipped with filters for the corresponding wavelengths of the laser. Specifically, C1 and C3 filter infrared light, while C2 and C4 filter red light, so that C1 only captures the laser emitted by L1 (red light) without being interfered with by the infrared light emitted by L2, and similarly, C2 only captures the infrared light emitted by L2 without being interfered with by the red light emitted by L1.

[0069] When a train passes, each laser group (L1-L4) operates. Each wheel, upon passing trigger T1, triggers all four imagers (C1-C4) to simultaneously acquire an image. In other words, trigger T1 triggers all four imagers to simultaneously acquire an image. Each laser group then starts operating upon detecting the train's passage. It is understandable that other triggers can be configured to detect train passage and activate the lasers upon detection.

[0070] As Figure 2 Each of the laser groups L1 to L4 shown emits laser light towards the wheel being tested, thus forming four detection sections, as shown. Figure 4 As shown, each detection surface is unidirectional. Detection surface 1 (corresponding to L1) detects the outer side of the wheel profile, detection surface 2 (corresponding to L2) detects the inner side of the wheel profile, detection surface 3 (corresponding to L3) detects the outer side of the wheel profile, and detection surface 4 (corresponding to L4) detects the inner side of the wheel profile. Figure 5 As shown, the inner reference plane of the wheel can be obtained by performing plane fitting on the straight line of the light section on the inner side of the wheel in the inner light section of the inner light section of 2 and 4.

[0071] In this embodiment, step S03, which involves restoring the inner and outer light cross-sections to standard cross-sections based on the inner reference plane of the wheel, includes:

[0072] S301. Obtain the required wheel parameters based on the inner reference plane of the wheel, and calculate the relative positions of the inner and outer contour light sections and the wheel under test respectively.

[0073] S302. Calculate the restoration transformation parameters based on the relative positions of the inner and outer contour light sections and the wheel being measured;

[0074] S303. Use the restoration transformation parameters to restore the inner and outer light cross sections to the standard cross section inner and outer light cross sections.

[0075] The aforementioned wheel parameters specifically include the rolling circle diameter D, the wheel center distance from the rolling circle axis O to the inner and outer contour light sections, and the angle between the inner wheel reference surface and the inner and outer contour light sections. The rolling circle mentioned above is the circle on the tread surface corresponding to a distance of 70mm from the inner wheel reference surface. The specific distance from the inner wheel reference surface can be adjusted according to the standard definition.

[0076] In this embodiment, the steps for determining the center O of the rolling circular shaft are as follows:

[0077] S311. Obtain target points on multiple rolling circles obtained by intersecting the plane of the rolling circle (i.e., at a distance of 70mm from the inner reference surface of the wheel) with each light section line. In a specific application embodiment, such as Figure 6 As shown, points B1-B4 on the rolling circles of the four wheels are obtained.

[0078] S312. Fit each target point (B1-B4) to a target plane, and obtain the center of the rolling circle based on the intersection of the axis and the target plane.

[0079] like Figure 6 As shown, the distances from the center O of the rolling circle to the four points are calculated. Using the sum of the squares of the differences between each of the four distances and R as the objective function, the wheel radius R is obtained as the average of the four distances, and the wheel diameter D = 2R. That is, the diameter of the rolling circle can be obtained from the distances from the center of the rolling circle to each target point (B1-B4).

[0080] According to the above steps, the rolling circle axis O and the rolling circle diameter D are determined, and then the wheel center distance (the vertical distance from the rolling circle axis O to the detection surface) of the rolling circle axis O to the corresponding light section is calculated, and the angle between the inner reference surface and the corresponding light section is calculated. According to the rolling circle diameter D, the wheel center distance, and the angle, the relative position of the corresponding light section line and the wheel is calculated. According to the required light section line, the light section line is determined to be on the inner side or the outer side. For example, if the outer light section line needs to be restored, the relative position of the outer light section line and the wheel is calculated according to the rolling circle diameter D, the wheel center distance of the rolling circle axis O to the outer light section, and the angle between the inner surface and the outer light section, so as to restore the outer light section line to the standard section outer light section line. Correspondingly, if the inner light section line needs to be restored, the relative position of the inner light section line and the wheel is calculated according to the rolling circle diameter D, the wheel center distance of the rolling circle axis O to the inner light section, and the angle between the inner surface and the inner light section, so as to restore the inner light section line to the standard section inner light section line.

[0081] In this embodiment, the specific determination steps of the axis line are as follows:

[0082] The inner reference surface of the wheel is translated to search for the non-wearing part (i.e., the back of the rim and the top of the rim) in the detected profile, and the intersection with the profile is obtained. The range of the non-wearing part of the back of the rim and the top of the rim can be determined according to the standard profile.

[0083] The obtained intersections are used in sequence to fit a plurality of axes: the first axis O1,..., the top of the rim On.

[0084] The axis line of the wheel set is fitted using the fitted plurality of axes O1- On.

[0085] The axis line of the wheel set is fitted using the robust straight line fitting algorithm. Since the light section line collected by the split light source will have different degrees of distortion, the robust positioning fitting method is further used to correct the collected profile, so as to reduce the distortion of the collected light section line.

[0086] In this embodiment, the restoration transformation parameter is specifically an affine transformation coefficient used for affine transformation. In step S302, the affine transformation coefficient of the inner and outer profile light section is calculated according to the relative position of the inner and outer profile light section and the measured wheel. In step S303, the affine transformation coefficient is used to perform affine transformation correction on the inner and outer light section line, so as to restore the standard section inner and outer light section line. By further combining the affine transformation restoration method, the inner and outer profile light section is restored to the standard section inner and outer light section line, which can further reduce the distortion of the collected light section line, correct the collected profile, and effectively improve the detection accuracy.

[0087] In a specific application embodiment, the affine transformation coefficients can be obtained in the following way: based on the obtained inner reference surface and the wheel center, a three-dimensional coordinate system containing the profile and the inner reference surface is established, a plane perpendicular to the inner reference surface and passing through the wheel center and parallel to the inner reference line of the profile is made, and the plane is the theoretical detection section; the coordinates of each profile point on the theoretical detection section are obtained by rotating the profile point around the axis, so that each detection profile forms a standard affine profile on the theoretical detection section, and the affine transformation coefficients are obtained by transforming the detection profile to the standard affine profile.

[0088] It can be understood that other transformation methods other than affine transformation can also be used to realize the restoration by using the corresponding relationship between the detected parameter information and the inner and outer light section lines of the standard section.

[0089] In this embodiment, the relative distance from the inner reference surface is searched to obtain the axis, the rolling circle axis, and the tread base point, and then the vertical distance between the detection light section and the wheel center, the angle between the detection light section and the inner reference surface of the wheel, and the rolling circle diameter are obtained. The position of the detected distorted profile is obtained by using the above parameters, affine transformation is performed to realize correction, and the corrected inner and outer profiles are spliced to form a complete profile, so that the split detection processing of the wheel tread profile is realized.

[0090] Taking the restoration of the outer light section line as an example, the detailed steps for restoring the outer light section line in this embodiment are as follows:

[0091] The inner reference surface is translated to the right to move, the non-worn part of the rim back and the rim top is searched, and the intersection points with the four profiles (the intersection points of the inner and outer profiles on both sides) are obtained. When the intersection points exceed three, the axis is fitted, the first obtained axis O1 is to the rim top On, and the robust straight line fitting algorithm is used to obtain the axis of the wheel set.

[0092] A plane 70 mm away from the inner reference surface 70 is made, and the intersection points of the plane and each light section line are obtained to obtain four points on the rolling circle of the wheel. A plane is fitted by the four points, and the intersection point of the axis and the plane is the rolling circle axis O. The distances from the rolling circle axis O to the four points are calculated, and the sum of the squares of the differences between the four distances and the wheel radius R is taken as the objective function to obtain the average of the four distances as the wheel radius R, and the wheel diameter D = 2R.

[0093] The wheel center distance (the vertical distance from O to the detection surface 1) of the rolling circle axis O to the outer light section 1 is calculated, and the angle between the fitted inner surface and the outer light section 1 is calculated. The relative position of the outer light section 1 and the wheel is calculated according to the diameter, the wheel center distance, and the angle, so that the affine transformation coefficients of the detection profile 1 can be calculated, and the outer light section 1 is restored to the standard section outer light section line.

[0094] According to the same principle, other light cut lines can also be reduced to standard cross-section light cut lines, in which the inner light cut lines (light cut lines 2, 4) are reduced to standard cross-section inner light cut lines, and the outer light cut lines (light cut lines 1, 3) are reduced to standard cross-section outer light cut lines.

[0095] In this embodiment, the reduced standard cross-section inner light cut lines and the reduced standard cross-section outer light cut lines are spliced to form a complete contour by using a point cloud registration algorithm of the closest point iteration. Taking the standard ICP algorithm as an example, since the coordinates after reduction are very close, coarse matching is not required, and the point cloud registration can be completed by directly taking the distance between the closest points of the two contour overlapping contours as the target loss function for iteration.

[0096] For example, as shown in Figure 5 , the inner and outer light cut lines 1 and 2 are spliced to obtain a complete contour 1, and the light cut lines 3 and 4 are spliced to obtain a complete contour 2, and the complete contour after splicing is shown in Figure 7 , in which the solid line contour is a standard contour. It can be understood that other point cloud registration algorithms can also be used to splice the light cut lines.

[0097] In this embodiment, the complete contour is compared with the standard contour according to the tread contour size definition, and the size information of each tread contour of the measured wheel is obtained, as shown in Figure 7 , including: the distance from the base point B1 to the highest point of the rim top to obtain the rim height, the distance from the rim thick point at the L2 point above the base point B1 to the inner side to obtain the rim thickness, the horizontal distance between the lowest point of the qR value at the L3 point above the base point B1 and the highest point of the qR value at the L4 point below the rim top to obtain the qR value, and the vertical distance between the standard base point B (theoretically, the base point of a new wheel without wear) and the base point B1 to obtain the tread wear, wherein the base point B1 is the base point of the detected contour, L1-L4 are the plane positions of the four light source emitting devices arranged on the two sides of the track, L1 and L3 are arranged on the outside of the track, and L2 and L4 are arranged on the inside of the track. Specifically, L1 is arranged on the left side of the outside of the track, L2 is arranged on the left side of the inside of the track, L3 is arranged on the right side of the outside of the track, and L4 is arranged on the right side of the inside of the track. According to the above method, it can be applied not only to detection from the left side, but also to detection from the right side, that is, it can be applied to bidirectional detection.

[0098] It can be understood that more groups of line light sources can be provided by using more than four groups of lasers to further improve the accuracy, or other waveband lasers and imagers can be used to improve the acquisition frequency and accuracy. The specific configuration can be determined according to actual needs.

[0099] The image processing device for dynamic detection of wheel size in this embodiment comprises:

[0100] The light section line extraction module is configured to acquire images of two or more inner light section lines and images of two or more outer light section lines on the measured wheel, wherein the inner light section lines and the outer light section lines are respectively formed by line light sources of different wave bands for detecting inner profiles and outer profiles of the measured wheel.

[0101] The inner reference surface determination module is configured to obtain the inner reference surface of the wheel according to the inner light section lines.

[0102] The light section line restoration module is configured to restore the inner and outer light section lines to standard cross-section inner and outer light section lines according to the inner reference surface of the wheel.

[0103] The light section line splicing module is configured to splice the restored standard cross-section inner and outer light section lines to form complete profiles.

[0104] The profile size determination module is configured to compare the complete profiles with a standard profile to obtain profile size information of each tread of the measured wheel.

[0105] In the embodiment, the light section line restoration module specifically includes:

[0106] The relative position calculation unit is configured to obtain required wheel parameters according to the inner reference surface of the wheel to calculate relative positions of the inner and outer profile light sections and the measured wheel.

[0107] The transformation coefficient calculation unit is configured to calculate restoration transformation parameters according to the relative positions of the inner and outer profile light sections and the measured wheel.

[0108] The restoration transformation unit is configured to restore the inner and outer light section lines to the standard cross-section inner and outer light section lines using the restoration transformation parameters.

[0109] In the embodiment, the light section line restoration module calculates affine transformation coefficients of the inner and outer profiles according to the relative positions of the inner and outer profile light sections and the measured wheel, and uses the affine transformation coefficients to perform affine transformation correction on the inner and outer light section lines to realize restoration of the standard cross-section inner and outer light section lines.

[0110] In the embodiment, the light section line splicing module uses a nearest point iteration point cloud registration algorithm to splice the restored standard cross-section inner light section line and the restored standard cross-section outer light section line to form complete profiles.

[0111] The image processing device for dynamic detection of wheel sizes and the image processing method for dynamic detection of wheel sizes described above are one-to-one correspondence, and will not be described one by one here.

[0112] In another embodiment, the image processing device for dynamic detection of wheel size can further comprise a processor and a memory, the memory being configured to store a computer program, and the processor being configured to execute the computer program to perform the method as described above.

[0113] The embodiment also provides a computer readable storage medium storing a computer program, the computer program being configured to implement the method as described above when executed.

[0114] Embodiment 2:

[0115] As shown in the following table, the steps of the wheel size dynamic detection method of the embodiment include: Figure 8

[0116] At least two or more line light sources are respectively emitted to the left and right sides of the measured wheel for detecting the inner and outer profiles of the wheel, wherein the line light sources for detecting the inner and outer profiles are line light sources of different wavebands;

[0117] The detection image of the measured wheel is acquired;

[0118] The acquired image is processed according to the image processing method of embodiment 1, and the information of the profile size of each tread of the measured wheel is output.

[0119] In the embodiment, two or more line light sources are emitted to the measured wheel to detect the inner profile line, and line light sources of different wavebands are emitted to the measured wheel to detect the outer profile. After the images of the inner and outer profile lines on the measured wheel are acquired, the images of the inner and outer profile lines are independently processed according to the image processing method of embodiment 1. The inner and outer laser profiles can be completely separated, the interference of the line light sources of the same waveband is reduced, the difficulty of image processing is greatly reduced, and the image processing speed is greatly improved. In addition, the detected inner profile line is restored to a standard cross-sectional profile line, and the complete profile is formed by splicing. Therefore, the complete profile line is not required, the inner and outer laser light sources do not need to be adjusted to coincide, the requirement for the completeness of the profile line is greatly reduced, the application flexibility is improved, the influence of the interference factors on the detection accuracy is avoided, and the detection accuracy is effectively improved.

[0120] The wheel size dynamic detection system of the embodiment includes:

[0121] The light source emitting device is configured to emit at least two or more line light sources to the left and right sides of the measured wheel for respectively detecting the inner and outer profiles of the wheel, wherein the line light sources for detecting the inner and outer profiles are line light sources of different wavebands;

[0122] ​The image acquisition device is used for acquiring images of two or more inner light cut lines and images of two or more outer light cut lines on the measured wheel, wherein the inner light cut line and the outer light cut line are respectively emitted to the measured wheel for detecting the inner profile and the outer profile of the measured wheel.

[0123] The image processing device in the embodiment 1 is used for acquiring the detection image of the measured wheel to perform image processing and output the profile size information of each tread of the measured wheel.

[0124] In the embodiment, the light source emitting device includes at least four groups of light source emitting equipment, wherein at least two groups of light source emitting equipment are arranged on one side of the track to detect the inner profile of the measured wheel, and at least two groups of light source emitting equipment are arranged on the other side of the track to detect the outer profile of the measured wheel, and the light source emitting equipment on one side of the track and the light source emitting equipment on the other side of the track are arranged in a staggered manner in the horizontal position to form a staggered position.

[0125] In the embodiment, the image acquisition device includes four or more imaging equipment to respectively acquire the linear light source emitted by each group of light source emitting equipment, and each imaging equipment is further provided with a filter to filter the linear light source of a specified wave band. The filter can form an independent processing of the inner / outer light strip image, and the inner profile and the outer profile of the laser profile can be completely separated.

[0126] In the embodiment, a trigger device connected with the light source emitting device and the image acquisition device is further included to control the trigger device to control the light source emitting device and the image acquisition device to start working when the train passes.

[0127] In the specific application embodiment, the light source emitting device, the image acquisition device and the trigger device are arranged as shown in Figure 2As shown. That is, four groups of lasers L1-L4, four imagers C1-C4 and a trigger T1 are arranged on both sides of the track, wherein L1 and L3 emit red light for detecting the outer profile, L2 and L3 emit infrared light for detecting the inner profile, L1 and L3 are arranged on one side of the track, L2 and L4 are arranged on the other side, and L1 and L3, L2 and L4 are symmetrical about the trigger T1, respectively, L1 and L2, L3 and L4 are not symmetrical about the track, respectively, and are staggered by being spaced apart in the horizontal direction; the four imagers C1-C4 are arranged near the four groups of lasers L1-L4 to collect images of the light cut lines formed by the corresponding lasers on the wheels. Each imager C1-C4 is also provided with a corresponding waveband filter according to the laser, specifically C1 and C3 filter infrared light, C2 and C4 filter red light, and when the train passes, each group of lasers L1-L4 works, and each wheel triggers the four groups of imagers C1-C4 once when passing through the trigger T1, that is, the four groups of imagers are triggered by the trigger T1 to collect images simultaneously, and each group of lasers starts to work when detecting the passing of the train.

[0128] The above is only the preferred embodiment of the present application, and does not limit the present application in any form. Although the present application has been disclosed as above with the preferred embodiment, it is not intended to limit the present application. Therefore, any simple modification, equivalent change and modification of the above embodiment without departing from the technical solution of the present application, according to the technical essence of the present application, should fall within the scope of protection of the technical solution of the present application.

Claims

1. An image processing method for dynamic detection of wheel dimensions, characterized by the steps of The method comprises the following steps: acquiring images of two or more inner light sections and images of two or more outer light sections of a measured wheel, the inner light sections and the outer light sections being formed by line light sources of different wave bands for detecting inner and outer profiles of the measured wheel; acquiring a wheel inner reference surface according to the inner light sections; restoring the inner and outer light sections to standard cross-section inner and outer light sections according to the wheel inner reference surface; splicing the restored standard cross-section inner and outer light sections to form a complete profile; comparing the complete profile with a standard profile to obtain profile size information of each tread of the measured wheel.

2. The image processing method for dynamic detection of wheel size according to claim 1, characterized in that, The step of restoring the inner and outer light sections to standard cross-section inner and outer light sections according to the wheel inner reference surface comprises the following steps: acquiring wheel parameters according to the wheel inner reference surface to calculate relative positions of the inner and outer profile cross sections and the measured wheel; calculating restoration transformation parameters according to the relative positions of the inner and outer profile cross sections and the measured wheel; restoring the inner and outer light sections to standard cross-section inner and outer light sections using the restoration transformation parameters.

3. The image processing method for dynamic detection of wheel size according to claim 2, characterized in that, The wheel parameters include a rolling circle diameter, a wheel center distance between a rolling circle axis and the inner and outer profile cross sections, and an included angle between the wheel inner reference surface and the inner and outer profile cross sections.

4. The image processing method for dynamic detection of wheel size according to claim 3, characterized in that, The determination of the rolling circle axis comprises the following steps: acquiring a plurality of target points obtained by intersecting a rolling circle plane with the light sections; fitting a target plane from the target points; and obtaining the rolling circle axis according to an intersection of an axis line and the target plane; and the rolling circle diameter is obtained according to distances from the rolling circle axis to the target points.

5. The image processing method for dynamic detection of wheel size according to claim 4, characterized in that, The determination of the axis line comprises the following steps: translating the wheel inner reference surface to search for non-worn parts of the detected profile to obtain intersections with the profile; fitting a plurality of axes using the searched intersections in sequence; and fitting the axis line of the wheel set using the fitted plurality of axes.

6. The image processing method for dynamic detection of wheel size according to claim 5, characterized in that, The axis line of the wheel set is fitted using a robust straight line fitting algorithm.

7. The image processing method for dynamic detection of wheel size according to claim 2, characterized in that, The restoration transformation parameters are affine transformation coefficients, the affine transformation coefficients of the inner and outer profiles are calculated according to the relative positions of the inner and outer profile cross sections and the measured wheel, the inner and outer light sections are corrected using the affine transformation coefficients to realize the restoration of the standard cross-section inner and outer light sections.

8. The image processing method for dynamic detection of wheel dimensions according to any one of claims 1 to 7, characterized in that, The restored standard cross-section inner light sections and the restored standard cross-section outer light sections are spliced to form the complete profile using a point cloud registration algorithm of a nearest point iteration.

9. The image processing method for dynamic detection of wheel dimensions according to any one of claims 1 to 7, characterized in that, The comparison of the complete profile with the standard profile obtains each tread profile size information of the measured wheel, including: the distance from the base point B1 to the highest point of the rim top obtains the rim height, the distance from the thick point of the rim at the point L2 above the base point B1 to the inner side surface obtains the rim thickness, the horizontal distance between the lowest point of the qR value at the point L3 above the base point B1 and the highest point of the qR value at the point L4 below the rim top obtains the qR value, and the vertical distance between the standard base point B and the base point B1 obtains the tread wear, wherein the base point B1 is the base point of the detected profile, L1-L4 are the plane positions of the four light source emitting devices arranged on the two sides of the track, L1 and L3 are arranged on the outside of the track, and L2 and L4 are arranged on the inside of the track.

10. A method of dynamically detecting the size of a vehicle wheel, characterized by the steps of It comprises: At least two linear light sources are respectively emitted to the left and right sides of the measured wheel for detecting the inner and outer profiles of the wheel, wherein the linear light sources for detecting the inner profile and the linear light sources for detecting the outer profile are linear light sources of different wave bands; Images of two or more inner light cut lines and images of two or more outer light cut lines on the measured wheel are collected; The collected images are processed by the image processing method according to any one of claims 1-9, and each tread profile size information of the measured wheel is output.

11. An image processing apparatus for dynamic detection of wheel dimensions, characterized in that It comprises: A light cut line acquisition module is configured to acquire images of two or more inner light cut lines and images of two or more outer light cut lines on the measured wheel, wherein the inner light cut lines and the outer light cut lines are respectively formed by linear light sources for detecting the inner profile and the outer profile of the measured wheel, and the linear light sources for detecting the inner profile and the linear light sources for detecting the outer profile are linear light sources of different wave bands; An inner reference surface determination module is configured to obtain a wheel inner reference surface according to each inner light cut line; A light cut line restoration module is configured to correspondingly restore the inner and outer light cut lines to standard cross-sectional inner and outer light cut lines according to the wheel inner reference surface; A light cut line splicing module is configured to correspondingly splice the restored standard cross-sectional inner and outer light cut lines to form a complete profile; A profile size determination module is configured to compare the complete profile with a standard profile to obtain each tread profile size information of the measured wheel.

12. An image processing apparatus for dynamic detection of wheel dimensions, comprising a processor and a memory for storing a computer program, characterized in that, The processor is configured to execute the computer program to execute the method according to any one of claims 1-9.

13. A dynamic wheel sizing system, comprising: It comprises: A light source emitting device is configured to emit at least two linear light sources to the left and right sides of the measured wheel for respectively detecting the inner and outer profiles of the wheel, wherein the linear light sources for detecting the inner profile and the linear light sources for detecting the outer profile are linear light sources of different wave bands; An image acquisition device is configured to acquire images of two or more inner light cut lines and images of two or more outer light cut lines on the measured wheel; The image processing device according to claim 11 or 12 is configured to acquire the detected images of the measured wheel for image processing and output each tread profile size information of the measured wheel.

14. The dynamic wheel size detection system of claim 13, wherein, The light source emitting device comprises at least four groups of light source emitting apparatuses, at least two groups of light source emitting apparatuses are arranged on one side of the track for detecting the inner profile of the measured wheel, and at least two groups of light source emitting apparatuses are arranged on the other side of the track for detecting the outer profile of the measured wheel, the light source emitting apparatuses on one side of the track and the light source emitting apparatuses on the other side of the track are arranged in a staggered manner in the horizontal position to form a staggered position.

15. The dynamic wheel size detection system of claim 13 or 14, wherein The image acquisition device comprises four or more imaging apparatuses for respectively acquiring the linear light source emitted by each group of light source emitting apparatuses, and each imaging apparatus is further provided with a filter for filtering the linear light source of a specified wave band.

16. The dynamic wheel size detection system of claim 15, wherein The trigger device connected with the image acquisition device is further included for controlling the trigger of the image acquisition of each imaging apparatus when the wheel of the train passes.

17. A computer readable storage medium storing a computer program, characterized in that, The computer program is executed to realize the method of any one of claims 1-9.

Images