Rail ultrasonic detection data processing method and device

Abstract

The one or more embodiments of the specification provide a steel rail ultrasonic detection data processing method and device, comprising: constructing super-layer image data with multiple layer images according to detection data of multiple ultrasonic detectors; one layer image corresponds to a target detected by one ultrasonic detector; inputting the super-layer image data into a preset detection model to obtain a detection result. The specification can automatically process ultrasonic detection data and obtain accurate detection results.

Description

Technical Field

[0001] This specification relates to the field of detection technology, and in particular to a method and apparatus for processing ultrasonic detection data of rails. Background Technology

[0002] Current ultrasonic flaw detection technology for rails generally converts ultrasonic detection data into either an A-scan (image obtained from ultrasonic A-scan) or a B-scan (image obtained from ultrasonic B-scan), and then manually observes these two images to analyze the internal structure or damage type and location of the rail. On the one hand, ultrasonic detection data is voluminous, resulting in a large workload and low efficiency; on the other hand, due to instrument limitations, data errors exist, and the two types of images may be distorted, leading to biased analysis results. Therefore, an automated method for processing ultrasonic detection data is needed to obtain accurate analysis results. Summary of the Invention

[0003] In view of the above, the purpose of one or more embodiments of this specification is to provide a method and apparatus for processing ultrasonic detection data of rails, which can automatically process data and obtain accurate detection results.

[0004] To achieve the above objectives, one or more embodiments of this specification provide a method for processing ultrasonic testing data of rails, including:

[0005] Based on the detection data from multiple ultrasonic detectors, a super-layer image data with multiple layers is constructed; one layer image corresponds to one target detected by an ultrasonic detector.

[0006] Input the superlayer image data into the preset detection model to obtain the detection results.

[0007] Optionally, based on the detection data from multiple ultrasonic detectors, a superlayer image data with multiple image layers is constructed, including:

[0008] Determine the corresponding image layer based on the detection location of the ultrasonic detector;

[0009] The superlayer image data is constructed based on the images of each layer.

[0010] Optionally, the layer image is a binary image;

[0011] The corresponding image layer is determined based on the detection position of the ultrasonic detector:

[0012] The target location is determined based on the deployment location of the ultrasonic detector, the incident angle, and the echo time.

[0013] The pixel value of the pixel corresponding to the target position in the layer image is set as the first pixel value, and the pixel values ​​of the other pixels in the layer image are set as the second pixel value.

[0014] Optionally, the input layer structure of the detection model includes: the number of superlayer image data, the height and width of the layer image, and the number of layer images contained in the superlayer image data.

[0015] Optionally, before constructing the super-layer image data with multiple layers based on the detection data from multiple ultrasonic detectors, the following steps are included:

[0016] Construct a training sample set; the training sample set includes images labeled with specific targets;

[0017] The detection model is obtained by training the deep neural network model using the training sample set.

[0018] This specification also provides an embodiment of a rail ultrasonic detection data processing device, comprising:

[0019] The construction module is used to construct super-layer image data with multiple layers based on the detection data of multiple ultrasonic detectors; one layer image corresponds to one target detected by an ultrasonic detector.

[0020] The detection module is used to input superlayer image data into a preset detection model to obtain detection results.

[0021] Optionally, the construction module is used to determine the corresponding layer image based on the detection position of the ultrasonic detector; and to construct the super-layer image data based on each layer image.

[0022] Optionally, the construction module is used to determine the target position of the target based on the deployment location, incident angle, and echo time of the ultrasonic detector; and to set the pixel value of the pixel corresponding to the target position in the layer image as the first pixel value, and the pixel value of the other pixels in the layer image as the second pixel value.

[0023] Optionally, the input layer structure of the detection model includes: the number of superlayer image data, the height and width of the layer image, and the number of layer images contained in the superlayer image data.

[0024] Optionally, the device further includes:

[0025] A training module is used to construct a training sample set, which includes images labeled with specific targets; and to train a deep neural network model using the training sample set to obtain the detection model.

[0026] As can be seen from the above description, the rail ultrasonic detection data processing method and apparatus provided in one or more embodiments of this specification construct a super-layer image data with multiple layers based on the detection data from multiple ultrasonic detectors; the super-layer image data is then input into a preset detection model to obtain the detection results. This specification can automatically process ultrasonic detection data and obtain accurate detection results. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in one or more embodiments of this specification 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 one or more embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a schematic diagram of one or more embodiments of the method described in this specification;

[0029] Figure 2 This is a schematic diagram of superlayer image data for one or more embodiments of this specification;

[0030] Figure 3A This is a schematic diagram illustrating the target location of a detector in one or more embodiments of this specification.

[0031] Figure 3B for Figure 3A The diagram shown illustrates the target location corresponding to the layer image.

[0032] Figure 4 These are schematic diagrams of neural network structures in some embodiments;

[0033] Figure 5 This is a schematic diagram of the neural network structure of one or more embodiments of this specification;

[0034] Figure 6 This is a schematic diagram of the device structure of one or more embodiments of this specification;

[0035] Figure 7 This is a schematic diagram of the structure of an electronic device according to one or more embodiments of this specification. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0037] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of this specification should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar words used in one or more embodiments of this specification do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0038] As described in the background section, determining the rail structure through manual analysis of images is inefficient and prone to errors. In developing this application, the applicant discovered that a neural network model can obtain detection results from the input image; however, B-mode imaging... Figure 1 Generally, the image is an RGB image. Different RGB combinations and specific images are needed to correspond to the targets detected by different ultrasonic detectors. When the target positions detected by different detectors are close, the different RGB images will overlap and cause interference, resulting in information loss or even confusion, and reducing the accuracy of the detection results.

[0039] In view of this, this application provides a method for processing ultrasonic detection data of rails. A super-layer image data is constructed based on the detection data of each ultrasonic detector, with each layer corresponding to a target detected by an ultrasonic detector. Then, the super-layer image data is input into a pre-trained detection model, which outputs the detection results. The constructed super-layer image data does not produce interference and ensures data integrity, thereby improving the accuracy of the detection results.

[0040] like Figure 1 As shown in the embodiments of this specification, a method for processing ultrasonic detection data of rails is provided, including:

[0041] S101: Based on the detection data from multiple ultrasonic detectors, construct super-layer image data with multiple layers; one layer image corresponds to one target detected by an ultrasonic detector.

[0042] S102: Input the superlayer image data into the preset detection model to obtain the detection results.

[0043] In this embodiment, detection data from multiple ultrasonic detectors are acquired, and a super-layer image data is constructed based on the detection data from each ultrasonic detector. The number of ultrasonic detectors is consistent with the number of layer images, and each layer image corresponds to a target detected by one ultrasonic detector. The constructed super-layer image data can represent the set of targets detected by all ultrasonic detectors. After constructing the super-layer image data, it is input into the detection model, which processes the super-layer image data and outputs the detection results. Since there is no interference between the layers of images, data integrity is guaranteed, and the accuracy of the detection results is improved.

[0044] In some embodiments, a super-layer image data with multiple layers is constructed based on detection data from multiple ultrasonic detectors, including:

[0045] Determine the corresponding image layer based on the detection location of the ultrasonic detector;

[0046] Construct superlayer image data based on the images of each layer.

[0047] Combination Figure 2 As shown, in this embodiment, ultrasonic detectors are deployed at different locations on the rail, and different ultrasonic detectors detect the structural morphology of the rail at different locations or different angles at the same location. For example, several ultrasonic detectors are deployed at equal intervals along the rail's extension direction, and each ultrasonic detector detects the structural morphology of different locations on the rail. For the hyperlayer image data, one layer image corresponds to the target of one ultrasonic detector. For ease of processing, the order of the layer images can correspond to the deployment order of the ultrasonic detectors. For example, ultrasonic detectors numbered 1, 2, i, i+1...N are deployed along the rail's extension direction, and correspondingly, the hyperlayer image data consists of layers numbered 1, 2, i, i+1...N, and N layers are combined to form the hyperlayer image data.

[0048] In some embodiments, the layer image is a binary image; based on the detection position of the ultrasonic detector, the corresponding layer image is determined as follows:

[0049] The target location is determined based on the deployment location of the ultrasonic detector, the incident angle, and the echo time.

[0050] Set the pixel value of the pixel corresponding to the target position in the layer image as the first pixel value, and set the pixel values ​​of the other pixels in the layer image as the second pixel value.

[0051] Combination Figure 2 , 3AAs shown in Figure 3B, in this embodiment, the layer image is a binary image. When the ultrasonic detector detects an echo signal, that is, when a target is detected, the pixel value at the target location is set to the first pixel value in the corresponding layer image to indicate that a target exists at that point, and the pixel values ​​at other locations are set to the second pixel value to indicate that a target does not exist at that point. Optionally, the first pixel value can be 1 and the second pixel value can be 0 to simplify data processing; in other methods, other two pixel values ​​can also be selected, and the specific values ​​are not limited.

[0052] To determine the target location detected by the ultrasonic detector, the distance D from the ultrasonic detector to the target is calculated based on the detector's deployment location, the incident angle (the angle between the ultrasonic wave's incident path and the normal to the top surface of the rail), and the echo time. Then, the target location (Xi0, Yi0) is calculated based on the incident angle θ and the distance D, using the following formula:

[0053]

[0054] The incident position of the ultrasonic detector is (Xin, 0).

[0055] In some implementations, an XOY coordinate system is established with the top surface of the rail and the direction perpendicular to the rail's extension direction as the X-axis, and the direction perpendicular to the X-axis as the Y-axis. The XOY plane is the plane containing the rail's cross-section, and the target position detected by the ultrasonic detector is a point within the rail's cross-section. Correspondingly, the layer image is an image corresponding to the rail's cross-section detected by the ultrasonic detector, and the target position detected by the ultrasonic detector corresponds to a specific pixel in the image of the rail's cross-section. To ensure complete coverage of the rail structure, the height of the layer image must be greater than the height of the rail, and the width of the layer image must be greater than the width of the rail. Based on the detection accuracy of the ultrasonic detector and the accuracy of the layer image, the corresponding pixel positions of the target position detected by the ultrasonic detector and the target position in the layer image can be determined. For example, a detection distance of 1 mm corresponds to one pixel in the layer image, or a detection distance of 1 mm corresponds to several pixels in the layer image. The specific positional correspondence ratio can be determined according to the detection accuracy, and this embodiment does not impose a specific limitation.

[0056] like Figure 4 As shown, in some implementations, the detection model is a deep neural network model. Considering that existing deep neural network models can process input RGB images, the input layer structure of the model is a four-dimensional tensor (B,H,W,C), where B is the number of image data input to the model, H is the height of the image data, W is the width of the image data, and C is the number of color channels of the image data (for RGB images, the number of color channels is 3).

[0057] like Figure 5As shown, in this embodiment, to adapt to the data processing of superlayer data, the input layer structure of the deep neural network model is improved. The input layer structure of the resulting detection model is a four-dimensional tensor (B, H, W, L), where B is the number of superlayer image data in the input model, H is the height of the layer image, W is the width of the layer image, and L is the number of layer images contained in the superlayer image data. Since each layer image contains only the target information, and there is no interference or confusion among the L layer images, the detection model can ensure data integrity during the processing of the superlayer image data, thereby outputting accurate detection results.

[0058] In some embodiments, the method for training the detection model is as follows:

[0059] Construct a training sample set; the training sample set includes images labeled with specific targets;

[0060] The deep neural network model is trained using the training sample set to obtain the detection model.

[0061] In this embodiment, to train the detection model, a training sample set is first constructed. Then, the training sample set is divided into a training set, a validation set, and a test set according to a certain ratio. The deep neural network model is trained using the training set, the trained deep neural network model is validated and its parameters are tuned using the validation set, and the cross-trained and validated deep neural network model is tested using the test set. Finally, a detection model is obtained. This detection model has the function of rail flaw detection and can output the detection results of rail flaw detection based on the input super-layer image data.

[0062] In some methods, the construction of the training sample set involves: acquiring ultrasonic detection data; generating a B-scan image based on the ultrasonic detection data; analyzing the B-scan image to determine the annotation content, including the analyzed structure or damage type, center location, and minimum area that can cover the structure or damage; generating a single-channel grayscale image with the same height and width as the B-scan image; initializing the single-channel grayscale image (all pixel values ​​initialized to 0); and based on the initialized grayscale image, annotating specific targets in the grayscale image using different grayscale values ​​according to the analyzed annotation content. For example, using a first grayscale value to annotate one type of damage, and a second grayscale value to annotate another type of damage, etc., with no specific limitation on the annotation method. Finally, the grayscale images annotated with specific targets constitute the training sample set, used to train the detection model.

[0063] It should be noted that the methods of one or more embodiments of this specification can be executed by a single device, such as a computer or server. The methods of this embodiment can also be applied in a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the methods of one or more embodiments of this specification, and the multiple devices will interact with each other to complete the method described.

[0064] It should be noted that the above description describes specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims may be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0065] like Figure 6 As shown in the embodiments of this specification, a rail ultrasonic detection data processing device is also provided, comprising:

[0066] The construction module is used to construct super-layer image data with multiple layers based on the detection data of multiple ultrasonic detectors; one layer image corresponds to one target detected by an ultrasonic detector.

[0067] The detection module is used to input superlayer image data into a preset detection model to obtain detection results.

[0068] In some embodiments, a construction module is used to determine the corresponding layer image based on the detection position of the ultrasonic detector; and to construct the super-layer image data based on each layer image.

[0069] In some embodiments, the construction module is used to determine the target position of the target based on the deployment location of the ultrasonic detector, the incident angle and the echo time; and to set the pixel value of the pixel corresponding to the target position in the layer image as the first pixel value, and the pixel value of other pixels in the layer image as the second pixel value.

[0070] In some embodiments, the input layer structure of the detection model includes: the number of superlayer image data, the height and width of the layer images, and the number of layer images contained in the superlayer image data.

[0071] In some embodiments, the apparatus further includes:

[0072] A training module is used to construct a training sample set, which includes images labeled with specific targets; and to train a deep neural network model using the training sample set to obtain the detection model.

[0073] For ease of description, the above apparatus is described in terms of function, divided into various modules. Of course, when implementing one or more embodiments of this specification, the functions of each module can be implemented in one or more software and / or hardware.

[0074] The apparatus described above is used to implement the corresponding methods in the foregoing embodiments and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0075] Figure 7 This embodiment illustrates a more specific hardware structure of an electronic device, which may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0076] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0077] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0078] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0079] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0080] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0081] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0082] The electronic devices described above are used to implement the corresponding methods in the foregoing embodiments and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0083] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0084] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this disclosure (including the claims) is limited to these examples; within the framework of this disclosure, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this specification as described above, which are not provided in detail for the sake of brevity.

[0085] Additionally, to simplify the description and discussion, and to avoid obscuring one or more embodiments of this specification, well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be illustrated in block diagram form to avoid obscuring one or more embodiments of this specification, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which one or more embodiments of this specification will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this disclosure, it will be apparent to those skilled in the art that one or more embodiments of this specification may be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0086] Although this disclosure has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0087] One or more embodiments of this specification are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments of this specification should be included within the scope of protection of this disclosure.

Claims

1. A method for processing ultrasonic testing data of steel rails, characterized in that, include: Based on the detection data from multiple ultrasonic detectors, a super-layer image data with multiple image layers is constructed; each image layer corresponds to a target detected by an ultrasonic detector; this includes: determining the target position of the target based on the deployment location, incident angle, and echo time of the ultrasonic detectors; setting the pixel value of the pixel corresponding to the target position in the image layer as the first pixel value, and setting the pixel values ​​of other pixels in the image layer as the second pixel value; constructing the super-layer image data based on each image layer; wherein, the method for calculating the target position (Xi0, Yi0) is as follows: Wherein, the incident position of the ultrasonic detector is (Xin,0), θ is the incident angle, and D is the distance from the ultrasonic detector to the target, which is calculated based on the deployment position of the ultrasonic detector, the incident angle, and the echo time. Input the superlayer image data into the preset detection model to obtain the detection results.

2. The method according to claim 1, characterized in that, The input layer structure of the detection model includes: the number of superlayer image data, the height and width of the layer image, and the number of layer images contained in the superlayer image data.

3. The method according to claim 1 or 2, characterized in that, Before constructing the super-layer image data with multiple layers based on the detection data from multiple ultrasonic detectors, the following steps are included: Construct a training sample set; the training sample set includes images labeled with specific targets; The detection model is obtained by training the deep neural network model using the training sample set.

4. A rail ultrasonic detection data processing device, characterized in that, include: The construction module is used to construct super-layer image data with multiple layers based on the detection data of multiple ultrasonic detectors. One layer image corresponds to one target detected by an ultrasonic detector; including: determining the target position of the target based on the deployment location, incident angle, and echo time of the ultrasonic detector; setting the pixel value of the pixel corresponding to the target position in the layer image as a first pixel value, and setting the pixel values ​​of other pixels in the layer image as second pixel values; constructing the super-layer image data based on each layer image; wherein, the method for calculating the target position (Xi0, Yi0) is as follows: Wherein, the incident position of the ultrasonic detector is (Xin,0), θ is the incident angle, and D is the distance from the ultrasonic detector to the target, which is calculated based on the deployment position of the ultrasonic detector, the incident angle, and the echo time. The detection module is used to input superlayer image data into a preset detection model to obtain detection results.

5. The apparatus according to claim 4, characterized in that, The input layer structure of the detection model includes: the number of superlayer image data, the height and width of the layer image, and the number of layer images contained in the superlayer image data.

6. The apparatus according to claim 4 or 5, characterized in that, Also includes: The training module is used to construct a training sample set, which includes images labeled with specific targets. The detection model is obtained by training the deep neural network model using the training sample set.

Images