Method and device for detecting contact fatigue cracks in a rail

Abstract

The application discloses a rail contact fatigue crack detection method and device, and relates to the technical field of rail nondestructive testing, wherein the method comprises the following steps: acquiring rail detection data, wherein the rail detection data comprises a rail top surface image and magnetic flux leakage detection data; determining a rail contact fatigue crack area according to the rail top surface image; screening the magnetic flux leakage detection data of the rail according to the rail contact fatigue crack area, so as to obtain the magnetic flux leakage detection data of the rail contact fatigue crack area; determining a rail head transverse fatigue crack detection result of the rail contact fatigue crack area according to the magnetic flux leakage detection data of the rail contact fatigue crack area; and determining a rail maintenance grade of the rail contact fatigue crack area according to the rail head transverse fatigue crack detection result of the rail contact fatigue crack area. The application can improve the detection efficiency and accuracy of the rail contact fatigue crack.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology for rails, and in particular to a method and apparatus for detecting contact fatigue cracks in rails. Background Technology

[0002] This section is intended to provide background or context for the embodiments of the invention set forth in the claims. The description herein is not an admission that it is prior art simply because it is included in this section.

[0003] Rails are a major component of railway infrastructure, supporting trains and guiding their wheels. Contact fatigue cracks occur when the wheel-rail contact stress exceeds the rail's fatigue strength, and the rate of railhead wear is lower than the rate of fatigue crack initiation and propagation. If the rail continues to be subjected to repeated wheel loads, these contact fatigue cracks will further propagate in depth, developing into transverse fatigue cracks at the railhead, ultimately leading to transverse rail fracture. Therefore, timely and effective detection and assessment of contact fatigue damage are necessary.

[0004] Traditional rail contact fatigue crack detection relies on manual inspection, which is costly, slow, and lacks accuracy. Non-destructive testing (NDT) technology, with its advantages of being non-contact, fast, and efficient, is being applied to rail damage detection. Existing technologies include machine vision-based surface fatigue crack detection methods using CCD image features. These methods utilize image processing techniques to analyze crack images, obtaining the length and width features of fatigue cracks to detect them. However, this method requires manual cropping of the area containing the surface fatigue crack for detecting small cracks, followed by scaling to obtain the crack image. For long cracks, it requires manually selecting several points within the crack area, determining a segmentation threshold based on the grayscale values ​​of these points, binarizing the crack image, and extracting the crack region. Therefore, this method requires manual assistance to detect fatigue cracks from the acquired images, cannot achieve automatic identification of contact fatigue cracks, and suffers from high cost and low efficiency. Summary of the Invention

[0005] This invention provides a method for detecting rail contact fatigue cracks, which aims to achieve accurate detection of rail contact fatigue cracks, reduce the cost of rail contact fatigue crack detection, and improve the efficiency of rail contact fatigue crack detection. The method includes:

[0006] Acquire rail inspection data, which includes images of the top surface of the rail and magnetic flux leakage detection data;

[0007] Based on the image of the top surface of the rail, determine the area where the rail contact fatigue crack is located;

[0008] Based on the location of the rail contact fatigue crack, the magnetic flux leakage test data of the rail is filtered to obtain the magnetic flux leakage test data of the area where the rail contact fatigue crack is located.

[0009] Based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, the detection results of the transverse fatigue crack of the rail head in the area where the rail contact fatigue crack is located are determined.

[0010] Based on the detection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located, the rail maintenance level in the area where the rail contact fatigue cracks are located is determined.

[0011] This invention also provides a rail contact fatigue crack detection device to achieve accurate rail contact fatigue crack detection, reduce rail contact fatigue crack detection costs, and improve rail contact fatigue crack detection efficiency. The device includes:

[0012] The data acquisition module is used to acquire rail inspection data, which includes images of the top surface of the rail and magnetic flux leakage detection data.

[0013] The image processing module is used to determine the location of contact fatigue cracks in the rail based on the image of the rail top surface.

[0014] The electromagnetic processing module is used to filter the magnetic flux leakage detection data of the rail based on the location of the rail contact fatigue crack, and obtain the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located; based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, the detection result of the transverse fatigue crack of the rail head in the area where the rail contact fatigue crack is located is determined.

[0015] The results output module is used to determine the rail maintenance level in the area where the rail contact fatigue crack is located based on the detection results of transverse fatigue cracks in the rail head.

[0016] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for detecting rail contact fatigue cracks.

[0017] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for detecting rail contact fatigue cracks.

[0018] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for detecting rail contact fatigue cracks.

[0019] In this embodiment of the invention, rail inspection data is acquired, including images of the rail top surface and magnetic flux leakage detection data. Based on the rail top surface image, the region where the rail contact fatigue crack is located is determined. Based on the region where the rail contact fatigue crack is located, the magnetic flux leakage detection data is filtered to obtain the magnetic flux leakage detection data for the region where the rail contact fatigue crack is located. Based on the magnetic flux leakage detection data for the region where the rail contact fatigue crack is located, the detection result of the transverse fatigue crack at the rail head in the region where the rail contact fatigue crack is located is determined. Based on the detection result of the transverse fatigue crack at the rail head in the region where the rail contact fatigue crack is located, the rail maintenance level for the region where the rail contact fatigue crack is located is determined. In this embodiment of the invention, images of the rail top surface and magnetic flux leakage detection data are acquired simultaneously. First, the rail top surface image is used to detect and identify the area where the rail contact fatigue crack is located. Then, based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, further detection is performed, and the detection result of the transverse fatigue crack at the rail head in the area where the rail contact fatigue crack is located is output. Finally, the maintenance level of the rail in the area where the rail contact fatigue crack is located is determined. This invention integrates machine vision technology and electromagnetic detection technology, realizing automatic and accurate detection of rail contact fatigue cracks. It effectively reduces the manual input of rail contact fatigue crack detection, lowers labor costs, and improves the efficiency and accuracy of rail contact fatigue crack detection. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, 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. In the drawings:

[0021] Figure 1 This is a schematic flowchart of the rail contact fatigue crack detection method in an embodiment of the present invention;

[0022] Figure 2 This is a schematic diagram of the rail top surface image and its corresponding Fourier spectrum in an embodiment of the present invention. Figure 1 ;

[0023] Figure 3 This is a schematic diagram of the rail top surface image and its corresponding Fourier spectrum in an embodiment of the present invention. Figure 2 ;

[0024] Figure 4 This is a schematic diagram of the area where the rail contact fatigue crack is located in an embodiment of the present invention;

[0025] Figure 5 This is a schematic diagram of the rail contact fatigue crack detection device in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0027] The applicant discovered that existing machine vision methods for detecting surface fatigue cracks based on CCD image features utilize image processing techniques to process crack images and obtain the length and width features of fatigue cracks to achieve surface fatigue crack detection. However, this method requires manual cropping of the region containing the surface fatigue crack for the detection of small cracks, followed by scaling to obtain the surface crack image; for long cracks, it requires manual selection of several points in the crack area, determining a segmentation threshold based on the gray values ​​of these points, binarizing the crack image, and extracting the crack region. Therefore, this method requires manual assistance to detect fatigue cracks from the acquired images, cannot achieve automatic identification of contact fatigue cracks, and has high detection costs and low efficiency. Based on this, the applicant proposes a method for detecting contact fatigue cracks in steel rails.

[0028] Figure 1 This is a schematic flowchart of the rail contact fatigue crack detection method in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:

[0029] Step 101: Obtain rail inspection data, which includes images of the top surface of the rail and magnetic flux leakage detection data;

[0030] Step 102: Based on the image of the top surface of the rail, determine the area where the rail contact fatigue crack is located;

[0031] Step 103: Based on the location of the rail contact fatigue crack, filter the magnetic flux leakage test data of the rail to obtain the magnetic flux leakage test data of the area where the rail contact fatigue crack is located.

[0032] Step 104: Based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, determine the detection results of the transverse fatigue crack of the rail head in the area where the rail contact fatigue crack is located.

[0033] Step 105: Based on the detection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located, determine the rail maintenance level in the area where the rail contact fatigue cracks are located.

[0034] from Figure 1As shown in the flowchart, in this embodiment of the invention, the rail top surface image and magnetic flux leakage detection data are collected simultaneously. First, the rail top surface image is used to detect and identify the area where the rail contact fatigue crack is located. Then, based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, the detection result of the transverse fatigue crack at the rail head in the area where the rail contact fatigue crack is located is output. Finally, the rail maintenance level in the area where the rail contact fatigue crack is located is determined. This invention integrates machine vision technology and electromagnetic detection technology, realizing automatic and accurate detection of rail contact fatigue cracks. It effectively reduces the manual input of rail contact fatigue crack detection, reduces labor costs, and improves the efficiency and accuracy of rail contact fatigue crack detection.

[0035] The method for detecting rail contact fatigue cracks in the embodiments of the present invention will be explained in detail below.

[0036] First, acquire rail inspection data, including images of the rail top surface and magnetic flux leakage detection data. This can be achieved by installing high-definition cameras and magnetic flux leakage detection equipment on a test train, and placing the equipment on the top surface of the rails to obtain the rail inspection data.

[0037] Then, based on the rail inspection data, rail contact fatigue crack detection and treatment are carried out.

[0038] In step 102, the area where the rail contact fatigue crack is located is determined based on the image of the rail top surface.

[0039] For example, image processing can be performed using an image of the top surface of the rail. The image can be divided into multiple regions based on grayscale values. The probability of rail contact fatigue cracks in each region can then be determined. By combining the probability results of all regions, the region where the rail contact fatigue cracks are located can be determined.

[0040] For example, a sample training set and a sample test set can be established using images of the top surface of historical rails and the areas where historical rail contact fatigue cracks are located. A machine learning model can be constructed, and the machine learning model can be trained using the sample training set and tested using the sample test set to obtain a trained rail contact fatigue crack recognition model. The rail top surface image can be input into the rail contact fatigue crack recognition model, and the area where the rail contact fatigue crack is located can be output.

[0041] In one embodiment, determining the region where the rail contact fatigue crack is located based on an image of the rail top surface may include the following steps 201 to 207:

[0042] Step 201: Perform Fourier transform processing on the rail top surface image to obtain the Fourier spectrum; the Fourier spectrum reflects the spectral characteristics of rail contact fatigue cracks.

[0043] In practice, the image of the top surface of the rail can be converted into a grayscale image first, and then the spectral characteristics of the rail contact fatigue crack can be identified from the image of the top surface of the rail using two-dimensional Fourier transform.

[0044] The formula for the two-dimensional Fourier transform is as follows:

[0045]

[0046] Explanation of parameters in the formula:

[0047] f(x,y) is the pixel value at (x,y) in the image;

[0048] M and N are the length and width of the image, respectively;

[0049] u and x range from 1 to M-1; v and y range from 1 to N-1.

[0050] Figure 2 This is a schematic diagram of the rail top surface image and its corresponding Fourier spectrum in an embodiment of the present invention. Figure 1 ,like Figure 2 The image shown is of a rail with contact fatigue cracks and the corresponding Fourier spectrum (the Fourier spectrum has the zero-frequency component moved to the center of the spectrum).

[0051] Figure 3 This is a schematic diagram of the rail top surface image and its corresponding Fourier spectrum in an embodiment of the present invention. Figure 2 ,like Figure 3 As shown, an image of a rail without contact fatigue cracks and its corresponding Fourier spectrum are displayed (the Fourier spectrum has the zero-frequency component moved to the center of the spectrum).

[0052] contrast Figure 2 and Figure 3 It can be seen that for images containing rail contact fatigue cracks, due to the obvious directionality of the cracks and the angle between the cracks and the rail, energy enhancement occurs at specific frequencies in the Fourier spectrum, except for the central axis (high brightness appears in the Fourier spectrum). Based on this, by suppressing this spectral feature and performing a two-dimensional inverse Fourier transform, a rail image with contact fatigue cracks removed can be obtained.

[0053] Step 202: Calculate the maximum and average energy values ​​of the Fourier spectrum based on the Fourier spectrum.

[0054] For example, on the Fourier spectrum, except for the central axis, calculate the maximum energy value a1 and the average energy value a2 of the Fourier spectrum.

[0055] Step 203: Determine a first threshold based on the maximum and average energy values ​​of the Fourier spectrum; the first threshold is used to replace the energy values ​​of the Fourier spectrum.

[0056] For example, based on the maximum energy value a1 and the average energy value a2, the first threshold t1 = a1 - a2 × 20% is determined.

[0057] Step 204: Replace the energy value of the Fourier spectrum with the first threshold to obtain the Fourier spectrum after replacing the energy value.

[0058] For example, replace the energy values ​​in all intervals [t1, a1] of the Fourier spectrum with a2.

[0059] Step 205: Perform an inverse Fourier transform on the Fourier spectrum after replacing the energy value to obtain the rail image with contact fatigue cracks removed.

[0060] The formula for the two-dimensional inverse Fourier transform is as follows:

[0061]

[0062] In the formula, F(u,v) is the energy value at (u,v), and u,v represent the spatial frequency of the Fourier transform;

[0063] M and N are the length and width of the image, respectively;

[0064] u and x range from 1 to M-1; v and y range from 1 to N-1.

[0065] Step 206: Normalize the pixel values ​​of the rail top surface image and the pixel values ​​of the rail image after removing contact fatigue cracks, and subtract them pixel by pixel to obtain the subtraction results of multiple pixels.

[0066] The pixel values ​​of the rail top surface image and the rail image after removing contact fatigue cracks are normalized to the interval [0, 1]. The normalization formula is as follows:

[0067]

[0068] In the formula, I(x,y) is the pixel value at (x,y) in the normalized image, and f max =max(f(x,y)), f min =min(f(x,y)), where f(x,y) is the pixel value at (x,y) in the image.

[0069] After normalization, pixel-by-pixel subtraction is performed to obtain the pixel difference value after subtraction of each pixel.

[0070] Step 207: Determine the area where the rail contact fatigue crack is located based on the preset second threshold and the subtraction result of multiple pixels; the second threshold is used to filter the pixels in the area where the rail contact fatigue crack is located based on the subtraction result of multiple pixels.

[0071] For example, with a preset threshold of 0.2, pixels in the rail top surface image with a pixel difference greater than 0.2 after subtraction are selected. The area enclosed by the smallest bounding rectangle containing all the selected pixels is determined as the region where the rail contact fatigue crack is located. Figure 4 This is a schematic diagram of the area where the rail contact fatigue crack is located in an embodiment of the present invention, as shown below. Figure 4 As shown, the black rectangle indicates the area where the rail contact fatigue crack is located.

[0072] In step 103, the magnetic flux leakage test data of the rail is filtered according to the region where the rail contact fatigue crack is located, so as to obtain the magnetic flux leakage test data of the region where the rail contact fatigue crack is located.

[0073] After determining the area where the rail contact fatigue crack is located, the magnetic flux leakage test data of the area where the rail contact fatigue crack is located is obtained by screening the magnetic flux leakage test data of all rails.

[0074] In step 104, the detection results of transverse fatigue cracks at the rail head in the area where the rail contact fatigue cracks are located are determined based on the leakage magnetic field detection data.

[0075] For example, by using magnetic flux leakage detection data of the area where rail contact fatigue cracks are located, a spatial coordinate system is established to obtain the magnetic field distribution in multiple directions. The magnetic field distribution in multiple directions is divided into multiple small units, and damage signals are identified for each small unit. By combining the damage signal identification results of all small units, the detection result of transverse fatigue cracks at the rail head in the area where rail contact fatigue cracks are located is determined.

[0076] For example, by using historical magnetic flux leakage detection data of areas where rail contact fatigue cracks are located, and historical detection results of transverse fatigue cracks at the rail head in the same areas, training and testing sets are established. A vector machine classification machine learning model is constructed, trained using the training set, and tested using the testing set to obtain the final transverse fatigue crack detection model at the rail head. The magnetic flux leakage detection data of areas where rail contact fatigue cracks are located is input into the transverse fatigue crack detection model at the rail head, and the detection results of transverse fatigue cracks at the rail head in the same areas are output.

[0077] In one embodiment, determining the rail head transverse fatigue crack detection result in the area where the rail contact fatigue crack is located, based on the rail magnetic flux leakage detection data, may include the following steps 401 to 403:

[0078] Step 401: Based on the magnetic flux leakage detection data of the rail contact fatigue crack area, determine the characteristic attribute set of the rail contact fatigue crack area; the characteristic attribute set includes the peak factor and kurtosis factor of each magnetic flux leakage detection sequence signal in the magnetic flux leakage detection data of the rail contact fatigue crack area.

[0079] During implementation, for each rail leakage magnetic flux detection signal {x1,x2,…,x} in the region where the rail contact fatigue crack is located, n The peak factor is calculated using the following formula:

[0080]

[0081] In the formula, x peak The peak value of the signal;

[0082] This is the root mean square value.

[0083] For each rail magnetic flux leakage detection signal {x1, x2, ..., x} in the region where the rail contact fatigue crack is located, n The kurtosis factor is calculated using the following formula:

[0084]

[0085] In the formula, δ is the average value of the signal, and δ is the standard deviation.

[0086] The peak factor w of multiple rail leakage magnetic flux detection signals in the region where rail contact fatigue cracks are located is used. 1_t and kurtosis factor w 2_t Establish feature attribute set W t ={w 1_t ,w 2_t}, where t represents the number of rail magnetic flux leakage detection signals, and w 1_t Let w be the peak factor of the t-th rail leakage magnetic flux detection signal. 2_t Let be the kurtosis factor of the t-th rail magnetic flux leakage detection signal.

[0087] Step 402: Using the Bayesian algorithm, calculate multiple likelihood probability values ​​for the feature attribute set and the pre-constructed magnetic flux leakage detection feature set; the pre-constructed magnetic flux leakage detection feature set includes multiple subsets, each subset including historical peak factor, kurtosis factor and category label of rail magnetic flux leakage detection sequence signal, the category label including a label for indicating whether the rail corresponding to the rail magnetic flux leakage detection sequence signal has or does not have transverse fatigue cracks at the rail head.

[0088] During implementation, multiple magnetic flux leakage (MFL) detection signals are collected in advance, including those from areas with and without transverse fatigue cracks in the railhead. The peak factor and kurtosis factor of each MFL detection signal are calculated. MFL detection samples are constructed using the peak factor and kurtosis factor of the multiple MFL detection signals, resulting in a MFL detection feature subset w. i ={w 1_i ,w 2_i}, where w 1_i w is the peak factor of the i-th magnetic flux leakage detection sample. 2_i Let w be the kurtosis factor of the i-th magnetic flux leakage detection sample. Based on whether each magnetic flux leakage detection sample originates from a region with existing transverse fatigue cracks in the railhead or a region without transverse fatigue cracks in the railhead, w is a subset of each magnetic flux leakage detection feature set. i Add a category tag Z j Where j = 1, 2, Z1 represents the presence of transverse fatigue cracks in the railhead, and Z2 represents the absence of transverse fatigue cracks in the railhead. The final leakage flux detection feature set W = {w1, w2, ..., w...} is obtained. l}, where l is the number of magnetic flux leakage detection samples.

[0089] Then, using the Bayesian algorithm, the feature attribute set W is calculated. t And multiple likelihood probability values ​​of the pre-constructed magnetic flux leakage detection feature set W, specifically, calculating the feature attribute set W. t and each subset w of the pre-constructed magnetic flux leakage detection feature set. i The likelihood probability value.

[0090] Step 403: Based on the category label corresponding to the maximum value among multiple likelihood probability values, determine the detection result of the transverse fatigue crack of the rail head in the area where the rail contact fatigue crack is located. The detection result of the transverse fatigue crack of the rail head in the area where the rail contact fatigue crack is located includes whether the transverse fatigue crack of the rail head exists or does not exist in the area where the rail contact fatigue crack is located.

[0091] In practice, the presence of transverse fatigue cracks at the rail head can be determined based on the category label corresponding to the maximum likelihood probability value of each subset of leakage magnetic field detection features in the pre-constructed feature set and the feature attribute set.

[0092] Finally, in step 105, the rail maintenance level in the area where the rail contact fatigue crack is located is determined based on the detection results of the transverse fatigue crack in the rail head.

[0093] In this step, the rail maintenance level is determined by combining the image processing results of the rail inspection data and the magnetic flux leakage detection data processing results.

[0094] For example, based on the specific magnitude of the likelihood probability value calculated by the Bayesian algorithm, the reliability score of the magnetic flux leakage detection data processing results is determined, and the rail maintenance level in the area where the rail contact fatigue crack is located is determined by combining the score results. For example, the maintenance level includes slight contact fatigue, moderate contact fatigue, and severe contact fatigue.

[0095] In one embodiment, determining the rail maintenance level in the area of ​​rail contact fatigue cracks based on the detection results of transverse fatigue cracks in the rail head may include:

[0096] Based on the inspection results of transverse fatigue cracks at the rail head in the area where the rail contact fatigue cracks are located: transverse fatigue cracks exist at the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level for the rail in the area where the rail contact fatigue cracks are located is determined to be: maintenance is recommended; and / or,

[0097] Based on the inspection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located: there are no transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level of the rail in the area where the rail contact fatigue cracks are located is determined to be: long-term monitoring is recommended.

[0098] During implementation, based on the combined results of magnetic flux leakage detection data processing and image data processing, the section of rail is considered to be in the initiation and development stage of rail contact fatigue cracks if image detection identifies rail contact fatigue cracks but magnetic flux leakage detection fails to identify rail head transverse fatigue cracks, requiring long-term monitoring. Conversely, the section of rail is considered to be in the deterioration stage if both image detection and magnetic flux leakage detection identify rail contact fatigue cracks, requiring maintenance and repair.

[0099] In summary, this invention provides a method for detecting rail contact fatigue cracks based on electromagnetic detection technology and machine vision technology. It simultaneously acquires rail magnetic flux leakage detection signals and images of the rail top surface, comprehensively utilizing image processing and signal processing techniques to automatically and accurately detect rail contact fatigue cracks. Furthermore, the rail magnetic flux leakage detection signals can be used to determine the degree of fatigue damage, effectively reducing the manual labor required for rail contact fatigue crack detection. This invention achieves automatic, efficient, and accurate detection and evaluation of rail contact fatigue cracks, providing a reliable basis for track maintenance and repair.

[0100] This invention also provides a rail contact fatigue crack detection device, as described in the following embodiments. Since the principle by which this device solves the problem is similar to the rail contact fatigue crack detection method, the implementation of this device can refer to the implementation of the rail contact fatigue crack detection method; repeated details will not be elaborated further.

[0101] Figure 5This is a schematic diagram of a rail contact fatigue crack detection device in an embodiment of the present invention, as shown below. Figure 5 As shown, the device includes:

[0102] The data acquisition module 501 is used to acquire rail inspection data, which includes rail top surface images and magnetic flux leakage detection data.

[0103] Image processing module 502 is used to determine the area where the rail contact fatigue crack is located based on the image of the rail top surface;

[0104] The electromagnetic processing module 503 is used to filter the magnetic flux leakage detection data of the rail according to the region where the rail contact fatigue crack is located, and obtain the magnetic flux leakage detection data of the region where the rail contact fatigue crack is located; and to determine the detection result of the transverse fatigue crack of the rail head in the region where the rail contact fatigue crack is located based on the magnetic flux leakage detection data of the region where the rail contact fatigue crack is located.

[0105] The result output module 504 is used to determine the rail maintenance level in the area where the rail contact fatigue crack is located based on the detection results of the transverse fatigue crack in the rail head.

[0106] In one embodiment, the image processing module 502 is specifically used for:

[0107] The Fourier transform of the rail top surface image is performed to obtain the Fourier spectrum.

[0108] Based on the Fourier spectrum, the maximum and average energy values ​​of the Fourier spectrum are calculated.

[0109] A first threshold is determined based on the maximum and average energy values ​​of the Fourier spectrum; the first threshold is used to replace the energy values ​​in the Fourier spectrum.

[0110] The energy values ​​of the Fourier spectrum are replaced using the first threshold to obtain the Fourier spectrum after energy replacement.

[0111] An inverse Fourier transform is performed on the Fourier spectrum after replacing the energy value to obtain a rail image with contact fatigue cracks removed.

[0112] The pixel values ​​of the rail top surface image and the pixel values ​​of the rail image after removing contact fatigue cracks are normalized and subtracted pixel by pixel to obtain the subtraction results of multiple pixels.

[0113] Based on a preset second threshold and the subtraction results of multiple pixels, the region where the rail contact fatigue crack is located is determined in the rail top surface image; the second threshold is used to filter the pixels in the region where the rail contact fatigue crack is located based on the subtraction results of multiple pixels.

[0114] In one embodiment, the electromagnetic processing module 503 is specifically used for:

[0115] Based on the magnetic flux leakage detection data of the rail contact fatigue crack area, the characteristic attribute set of the rail contact fatigue crack area is determined; the characteristic attribute set includes the peak factor and kurtosis factor of each magnetic flux leakage detection sequence signal in the magnetic flux leakage detection data of the rail contact fatigue crack area.

[0116] Using a Bayesian algorithm, multiple likelihood probability values ​​are calculated for the feature attribute set and the pre-constructed magnetic flux leakage detection feature set. The pre-constructed magnetic flux leakage detection feature set includes multiple subsets, each subset including historical peak factor, kurtosis factor and category label of the rail magnetic flux leakage detection sequence signal. The category label includes a label indicating whether the rail corresponding to the rail magnetic flux leakage detection sequence signal has or does not have a railhead transverse fatigue crack.

[0117] Based on the category label corresponding to the maximum value among multiple likelihood probability values, the detection result of transverse fatigue cracks in the rail head in the area where the rail contact fatigue crack is located is determined. The detection result of transverse fatigue cracks in the area where the rail contact fatigue crack is located includes whether or not transverse fatigue cracks in the rail head exist in the area where the rail contact fatigue crack is located.

[0118] In one embodiment, the result output module 504 is specifically used for:

[0119] Based on the inspection results of transverse fatigue cracks at the rail head in the area where the rail contact fatigue cracks are located: transverse fatigue cracks exist at the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level for the rail in the area where the rail contact fatigue cracks are located is determined to be: maintenance is recommended; and / or,

[0120] Based on the inspection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located: there are no transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level of the rail in the area where the rail contact fatigue cracks are located is determined to be: long-term monitoring is recommended.

[0121] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for detecting rail contact fatigue cracks.

[0122] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for detecting rail contact fatigue cracks.

[0123] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for detecting rail contact fatigue cracks.

[0124] In this embodiment of the invention, rail inspection data is acquired, including images of the rail top surface and magnetic flux leakage detection data. Based on the rail top surface image, the region where the rail contact fatigue crack is located is determined. Based on the region where the rail contact fatigue crack is located, the magnetic flux leakage detection data is filtered to obtain the magnetic flux leakage detection data for the region where the rail contact fatigue crack is located. Based on the magnetic flux leakage detection data for the region where the rail contact fatigue crack is located, the detection result of the transverse fatigue crack at the rail head in the region where the rail contact fatigue crack is located is determined. Based on the detection result of the transverse fatigue crack at the rail head in the region where the rail contact fatigue crack is located, the rail maintenance level for the region where the rail contact fatigue crack is located is determined. In this embodiment of the invention, images of the rail top surface and magnetic flux leakage detection data are acquired simultaneously. First, the rail top surface image is used to detect and identify the area where the rail contact fatigue crack is located. Then, based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, further detection is performed, and the detection result of the transverse fatigue crack at the rail head in the area where the rail contact fatigue crack is located is output. Finally, the maintenance level of the rail in the area where the rail contact fatigue crack is located is determined. This invention integrates machine vision technology and electromagnetic detection technology, realizing automatic and accurate detection of rail contact fatigue cracks. It effectively reduces the manual input of rail contact fatigue crack detection, lowers labor costs, and improves the efficiency and accuracy of rail contact fatigue crack detection.

[0125] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0127] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0128] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0129] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting fatigue cracks in rail contact, characterized in that, include: Acquire rail inspection data, which includes images of the top surface of the rail and magnetic flux leakage detection data; Based on the image of the top surface of the rail, determine the area where the rail contact fatigue crack is located; Based on the location of the rail contact fatigue crack, the magnetic flux leakage test data of the rail is filtered to obtain the magnetic flux leakage test data of the area where the rail contact fatigue crack is located. Based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, the detection results of the transverse fatigue crack of the rail head in the area where the rail contact fatigue crack is located are determined. Based on the detection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located, the rail maintenance level in the area where the rail contact fatigue cracks are located is determined. Specifically, based on the image of the rail top surface, the region where the rail contact fatigue crack is located is determined, including: The Fourier transform of the rail top surface image is performed to obtain the Fourier spectrum. Based on the Fourier spectrum, the maximum energy value a1 and the average energy value a2 of the Fourier spectrum are calculated. A first threshold t1 is determined based on the maximum and average energy values ​​of the Fourier spectrum; the first threshold is used to replace the energy values ​​of the Fourier spectrum. The energy values ​​of the Fourier spectrum are replaced using the first threshold to obtain the Fourier spectrum after energy replacement; wherein, the energy values ​​in all intervals [t1, a1] of the Fourier spectrum are replaced with a2; An inverse Fourier transform is performed on the Fourier spectrum after replacing the energy value to obtain a rail image with contact fatigue cracks removed. The pixel values ​​of the rail top surface image and the pixel values ​​of the rail image after removing contact fatigue cracks are normalized and subtracted pixel by pixel to obtain the subtraction results of multiple pixels. Based on a preset second threshold and the subtraction result of multiple pixels, the region where the rail contact fatigue crack is located is determined in the rail top surface image; the second threshold is used to filter the pixels in the region where the rail contact fatigue crack is located based on the subtraction result of multiple pixels. Specifically, based on the rail magnetic flux leakage detection data in the area where the rail contact fatigue crack is located, the detection results of the transverse fatigue crack at the rail head in the area where the rail contact fatigue crack is located are determined, including: Based on the magnetic flux leakage detection data of the rail contact fatigue crack area, the characteristic attribute set of the rail contact fatigue crack area is determined; the characteristic attribute set includes the peak factor and kurtosis factor of each magnetic flux leakage detection sequence signal in the magnetic flux leakage detection data of the rail contact fatigue crack area. Using a Bayesian algorithm, multiple likelihood probability values ​​are calculated for the feature attribute set and the pre-constructed magnetic flux leakage detection feature set. The pre-constructed magnetic flux leakage detection feature set includes multiple subsets, each subset including historical peak factor, kurtosis factor and category label of the rail magnetic flux leakage detection sequence signal. The category label includes a label indicating whether the rail corresponding to the rail magnetic flux leakage detection sequence signal has or does not have a railhead transverse fatigue crack. Based on the category label corresponding to the maximum value among multiple likelihood probability values, the detection result of transverse fatigue cracks in the rail head in the area where the rail contact fatigue crack is located is determined. The detection result of transverse fatigue cracks in the area where the rail contact fatigue crack is located includes whether or not transverse fatigue cracks in the rail head exist in the area where the rail contact fatigue crack is located.

2. The method as described in claim 1, characterized in that, Based on the inspection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located, the rail maintenance level in the area where the rail contact fatigue cracks are located is determined, including: Based on the inspection results of transverse fatigue cracks at the rail head in the area where the rail contact fatigue cracks are located: transverse fatigue cracks exist at the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level for the rail in the area where the rail contact fatigue cracks are located is determined to be: maintenance is recommended; and / or, Based on the inspection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located: there are no transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level of the rail in the area where the rail contact fatigue cracks are located is determined to be: long-term monitoring is recommended.

3. A rail contact fatigue crack detection device, characterized in that, include: The data acquisition module is used to acquire rail inspection data, which includes images of the top surface of the rail and magnetic flux leakage detection data. The image processing module is used to determine the location of contact fatigue cracks in the rail based on the image of the rail top surface. The electromagnetic processing module is used to filter the magnetic flux leakage detection data of the rail based on the location of the rail contact fatigue crack, and obtain the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located; based on the magnetic flux leakage detection data of the area where the rail contact fatigue crack is located, the detection result of the transverse fatigue crack of the rail head in the area where the rail contact fatigue crack is located is determined. The result output module is used to determine the rail maintenance level in the area where the rail contact fatigue crack is located based on the detection results of the transverse fatigue crack in the rail head. Specifically, the image processing module is used for: The Fourier transform of the rail top surface image is performed to obtain the Fourier spectrum. Based on the Fourier spectrum, the maximum energy value a1 and the average energy value a2 of the Fourier spectrum are calculated. A first threshold t1 is determined based on the maximum and average energy values ​​of the Fourier spectrum; the first threshold is used to replace the energy values ​​of the Fourier spectrum. The energy values ​​of the Fourier spectrum are replaced using the first threshold to obtain the Fourier spectrum after energy replacement; wherein, the energy values ​​in all intervals [t1, a1] of the Fourier spectrum are replaced with a2; An inverse Fourier transform is performed on the Fourier spectrum after replacing the energy value to obtain a rail image with contact fatigue cracks removed. The pixel values ​​of the rail top surface image and the pixel values ​​of the rail image after removing contact fatigue cracks are normalized and subtracted pixel by pixel to obtain the subtraction results of multiple pixels. Based on a preset second threshold and the subtraction result of multiple pixels, the region where the rail contact fatigue crack is located is determined in the rail top surface image; the second threshold is used to filter the pixels in the region where the rail contact fatigue crack is located based on the subtraction result of multiple pixels. Specifically, the electromagnetic processing module is used for: Based on the magnetic flux leakage detection data of the rail contact fatigue crack area, the characteristic attribute set of the rail contact fatigue crack area is determined; the characteristic attribute set includes the peak factor and kurtosis factor of each magnetic flux leakage detection sequence signal in the magnetic flux leakage detection data of the rail contact fatigue crack area. Using a Bayesian algorithm, multiple likelihood probability values ​​are calculated for the feature attribute set and the pre-constructed magnetic flux leakage detection feature set. The pre-constructed magnetic flux leakage detection feature set includes multiple subsets, each subset including historical peak factor, kurtosis factor and category label of the rail magnetic flux leakage detection sequence signal. The category label includes a label indicating whether the rail corresponding to the rail magnetic flux leakage detection sequence signal has or does not have a railhead transverse fatigue crack. Based on the category label corresponding to the maximum value among multiple likelihood probability values, the detection result of transverse fatigue cracks in the rail head in the area where the rail contact fatigue crack is located is determined. The detection result of transverse fatigue cracks in the area where the rail contact fatigue crack is located includes whether or not transverse fatigue cracks in the rail head exist in the area where the rail contact fatigue crack is located.

4. The apparatus as described in claim 3, characterized in that, The results output module is specifically used for: Based on the inspection results of transverse fatigue cracks at the rail head in the area where the rail contact fatigue cracks are located: transverse fatigue cracks exist at the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level for the rail in the area where the rail contact fatigue cracks are located is determined to be: maintenance is recommended; and / or, Based on the inspection results of transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located: there are no transverse fatigue cracks in the rail head in the area where the rail contact fatigue cracks are located. Therefore, the maintenance level of the rail in the area where the rail contact fatigue cracks are located is determined to be: long-term monitoring is recommended.

5. A computer device, 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 method of any one of claims 1 to 2.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 2.

7. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 2.

Images