Pantograph slip detection method, device and equipment and storage medium

Abstract

The invention discloses a pantograph slip detection method, device and equipment and a storage medium, and relates to the technical field of rail transit equipment detection. The pantograph slip detection method comprises the following steps: acquiring a pantograph image when a train runs normally; executing a preset normalization processing operation on the pantograph image; sampling the processed pantograph image for multiple times by using a residual network algorithm, and training a residual network model; and inputting a to-be-identified pantograph image into the residual network model, and identifying the slippage state of the pantograph. The problems that in actual subway operation work, due to the fact that a pantograph is a pure mechanical part, a large number of sensors such as ultraviolet sensors, infrared sensors and laser sensors are usually needed for detection, the early-stage investment is extremely high, the number of the sensors is too large, consequently, the roof space is limited, the sensors are troublesome to replace and maintain due to faults, and manpower resources are greatly wasted during regular cleaning are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of rail transit equipment detection, and in particular to a pantograph slip detection method, device, equipment and storage medium. Background Art

[0002] With the rapid development of rail transit, the safety of train operation has become increasingly prominent. The pantograph is an important electrical component of urban rail trains, used to collect current from the contact line to provide power for the train. At present, the pantograph will slip during the operation of the train, resulting in abnormal wear and arc sparks.

[0003] In actual subway operation, since the pantograph is a purely mechanical component, it usually requires a large number of external sensors for detection, such as ultraviolet, infrared, laser, etc. The initial investment is extremely high, and too many sensors cause tight roof space, sensor failure replacement and maintenance are troublesome, and regular cleaning results in a huge waste of human resources. No effective solution has been proposed so far. Summary of the invention

[0004] Purpose of the invention: To provide a pantograph slip detection method, device, equipment and storage medium to solve the above-mentioned problems existing in the prior art.

[0005] Technical solution: A pantograph slippage detection method, comprising: Obtain pantograph images during normal train operation; Performing a preset normalization processing operation on the pantograph image; The residual network algorithm is used to perform multiple sampling from the processed pantograph image to train the residual network model; The pantograph image to be identified is input into the residual network model to identify the sliding state of the pantograph.

[0006] Preferably, obtaining the pantograph image when the train is operating normally includes: The pantograph images taken when the train is in normal operation are collected. The pantograph images are in two states: a top-down view and an upward view, with 1,000 images each.

[0007] Preferably, the pantograph images taken from the normal operation of the train are collected, wherein the pantograph images are in two states, a top-down view and an upward view, and 1,000 images of each state are respectively taken, and the following further comprises: The obtained pantograph image is divided into a training set for model training and a test set for model testing according to a preset ratio.

[0008] Preferably, after performing a preset normalization operation on the pantograph image, multiple sampling is performed from the processed pantograph image using a residual network algorithm, and before training the residual network model, the method further includes: Perform preset data cleaning, target data extraction and data normalization operations on the normalized pantograph image.

[0009] Preferably, a residual network algorithm is used to perform multiple sampling from the processed pantograph image to train a residual network model, including: The divided training set is input into the residual backbone network to extract fault features, and the arcing and abnormal wear detection models are trained respectively; Five residual modules are preset, each of which consists of two convolutional layers. The first convolutional layer takes a 3x3 convolution kernel for convolution operation, passes through the ReLu activation function, and then passes through the second 3x3 convolutional layer for calculation. After the result is added to the residual value, a single training result is completed.

[0010] Preferably, the residual network algorithm is used to perform multiple sampling from the processed pantograph image to train a residual network model, and further includes: A dynamic learning rate hyperparameter model was used, with an initial learning rate of 0.01, a decay exponent of 0.1, a batch size of 5, and a total of 1500 iterations.

[0011] Preferably, a residual network algorithm is used to perform multiple sampling from the processed pantograph image, and before the residual network model is trained, the pantograph image to be identified is input into the residual network model, and after the pantograph slip state is identified, the method further includes: Confirm the calculation accuracy of the trained residual network model according to the preset accuracy formula; Wherein, the accuracy formula is: .

[0012] Preferably, the pantograph image to be identified is input into the residual network model to identify the sliding state of the pantograph, including: The divided training set is input into the residual network model, and the sliding state of the pantograph is output.

[0013] In order to achieve the above objective, according to another aspect of the present application, a pantograph slip detection device is provided.

[0014] The pantograph slippage detection device according to the present application includes: An acquisition module is used to acquire the pantograph image when the train is operating normally; A normalization module, used for performing a preset normalization processing operation on the pantograph image; A model training module is used to train a residual network model by sampling multiple times from the processed pantograph image using a residual network algorithm; The pantograph slip state recognition module is used to input the pantograph image to be recognized into the residual network model to recognize the slip state of the pantograph.

[0015] In order to achieve the above-mentioned purpose, according to another aspect of the present application, an electronic device is provided, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the pantograph slip detection method described in any one of the present invention.

[0016] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a computer-readable storage medium is provided, in which computer instructions are stored, and the computer instructions are used to implement the pantograph slip detection method described in any one of the present invention when executed by a processor.

[0017] Beneficial effect: In the embodiment of the present application, a deep learning framework is established based on a residual network, by obtaining a pantograph image when the train is operating normally; performing a preset normalization operation on the pantograph image; using a residual network algorithm to perform multiple sampling from the processed pantograph image to train a residual network model; the pantograph image to be identified is input into the residual network model to identify the slip state of the pantograph, thereby achieving the purpose of automatically diagnosing and identifying the slip state of the pantograph, thereby realizing timely reporting of abnormal arcs in the pantograph network during operation, reporting abnormal wear of the carbon slide plate during maintenance, and helping users to accurately achieve the technical effect of status maintenance, thereby solving the problem in actual subway operation that, because the pantograph is a purely mechanical component, a large number of sensors are usually required for detection, such as ultraviolet, infrared, laser, etc., and the initial investment is extremely high. In addition, too many sensors cause tight roof space, troublesome sensor failure replacement and maintenance, and regular cleaning results in a huge waste of human resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of a pantograph slippage detection method according to an embodiment of the present application; Figure 2 is a structural schematic diagram of a pantograph slip detection device according to an embodiment of the present application; Figure 3 is a residual network schematic diagram of a pantograph slippage detection method according to an embodiment of the present application; Figure 4 is a normal wear diagram of a pantograph on one side according to an embodiment of the present application; Figure 5 is a diagram of abnormal wear of a pantograph according to an embodiment of the present application; Figure 6 is a pantograph arc spark detection diagram according to an embodiment of the present application; and Figure 7 It is a structural schematic diagram of an electronic device according to the pantograph slip detection method of an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0020] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0021] In addition, the terms "installed", "set", "provided with", "connected", "connected", and "socketed" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral structure; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be an internal connection between two devices, elements, or components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0023] According to an embodiment of the present invention, a pantograph slip detection method is provided, such as Figure 1 As shown, the method includes the following steps S101 to S104: Step S101, obtaining a pantograph image when the train is operating normally; By collecting pantograph images in real time during normal train operation, accurate image data collection can be achieved, thus providing a basis for subsequent processing.

[0024] It should be noted that the collection methods include but are not limited to: camera shooting. Install a high-resolution industrial camera on the train or along the track. The camera can be set on the roof of the train or in a fixed position near the pantograph to obtain dynamic images of the pantograph through high-speed photography.

[0025] Drones are deployed along the tracks and used to track and photograph high-speed trains, obtaining images of the train's pantograph at different angles.

[0026] According to an embodiment of the present invention, preferably, obtaining a pantograph image when a train is operating normally includes: The pantograph images taken when the train is in normal operation are collected. The pantograph images are in two states: a top-down view and an upward view, with 1,000 images each.

[0027] By taking images of the pantograph from multiple angles, we can avoid information collection errors and reduce the probability of misjudgment. At the same time, by collecting images separately, we can ensure a good quantitative foundation, thus providing a basis and guarantee for subsequent computer deep learning.

[0028] According to an embodiment of the present invention, preferably, the pantograph images taken from the normal operation of the train are collected, wherein the pantograph images are in two states, a top view and a bottom view, respectively, and after 1000 images of each, the following is further included: The obtained pantograph image is divided into a training set for model training and a test set for model testing according to a preset ratio.

[0029] Specifically, the pantograph images taken are annotated and divided into training and test sets, and 700 training and 300 test set annotated data images are prepared respectively. Of course, the actual ratio of the test set and the training set can be set according to the actual use requirements, and is not limited in this application.

[0030] Step S102, performing a preset normalization operation on the pantograph image; Normalize the data samples, that is, map the data to (0, 1), improve the convergence speed of the next residual network, and reduce the impact of different ranges of data on the model. The normalization formula is as follows: .

[0031] According to an embodiment of the present invention, preferably, after performing a preset normalization operation on the pantograph image, multiple sampling is performed from the processed pantograph image using a residual network algorithm, and before training a residual network model, the method further includes: Perform preset data cleaning, target data extraction and data normalization operations on the normalized pantograph image.

[0032] Specifically, the purpose of data cleaning is to remove noise, redundant or incomplete data in image data and ensure the quality and consistency of image input data.

[0033] Noise Removal: Filtering, which removes noise from an image while retaining the edge details of the image by using Gaussian filtering, median filtering, or bilateral filtering.

[0034] Deblurring uses deblurring algorithms (such as deconvolution, Laplace sharpening, etc.) to improve image clarity, especially for blurred images obtained under high-speed motion conditions.

[0035] Image cropping: Automatic or manual cropping, based on the fixed area where the pantograph is located, remove irrelevant background information in the image and retain the pantograph area. The specific location of the pantograph can be framed and cropped through automatic target detection (such as YOLO, Faster R-CNN, etc.).

[0036] Image completion or repair: Image interpolation. For damaged image parts (such as partial loss due to signal loss or shooting angle problems), image interpolation algorithms can be used to repair them, such as bilinear interpolation, Lagrange interpolation, etc. Missing pixel filling: For small blank areas or missing pixels in the image, local mean or edge interpolation algorithms can be used to fill them.

[0037] The main purpose of target data extraction is to extract the feature data of the pantograph from the processed image to help subsequent fault detection or model training.

[0038] Edge detection: Use edge detection algorithms such as Canny, Sobel or Laplacian to extract pantograph contour features. Edge detection helps identify the wear or structural integrity of the pantograph.

[0039] Target segmentation: Use image segmentation techniques such as threshold segmentation, GrabCut or segmentation networks such as U-Net to separate the pantograph from the background. This will keep only the target area of ​​the pantograph for further analysis.

[0040] Feature point extraction: Extract key feature points, such as using Harris corner detection, SIFT or ORB algorithms to capture the feature points of the pantograph for subsequent image comparison or feature analysis. For pantograph detection, feature points can be used to analyze arcing positions or damage conditions.

[0041] Morphological operation: Morphological operations (dilation, erosion, opening, closing) are used to further enhance the contour features of the image target. This operation can remove small noise blocks in the image while retaining the main structure of the pantograph. Normalization is to ensure that the data format between different images is consistent and to normalize the data to a specific range for subsequent input processing of deep learning models; For example: normalization, standardization, data enhancement: In the normalization process, data enhancement techniques (such as random flipping, rotation, scaling, contrast adjustment, color jitter, etc.) can be introduced to increase the diversity of image data and improve the generalization ability of the model; Resizing: resize the image to the required input size (such as 448x448) to avoid input inconsistency caused by different image sizes. Common resizing methods include bilinear interpolation and bicubic interpolation.

[0042] Step S103, using a residual network algorithm to perform multiple sampling from the processed pantograph image to train a residual network model; It can achieve accurate model training results, thus providing a basis and guarantee for subsequent predictions.

[0043] According to an embodiment of the present invention, preferably, a residual network algorithm is used to perform multiple sampling from the processed pantograph image to train a residual network model, including: The divided training set is input into the residual backbone network to extract fault features, and the arcing and abnormal wear detection models are trained respectively; Five residual modules are preset, each of which consists of two convolutional layers. The first convolutional layer takes a 3x3 convolution kernel for convolution operation, passes through the ReLu activation function, and then passes through the second 3x3 convolutional layer for calculation. After the result is added to the residual value, a single training result is completed.

[0044] Specifically, Figure 3 As shown in the figure, the deep learning framework based on the residual network is established. The picture shows a two-layer residual block and a five-layer residual block network is set. The detailed residual design is shown in Table 1: It is important to know that the basic concept of residual network is the residual block, which is the core of ResNet. Its basic structure is as follows: The input X is transformed through several layers (usually convolutional layers, BatchNorm layers, ReLU activation functions, etc.) to obtain the output F(X), which is the feature after nonlinear transformation.

[0045] Through the skip connection, the input X is directly added to F(X) to obtain the final output: Y=F(X)+X; where F(X) represents the nonlinear transformation result of the input X, and X is the input data.

[0046] What you need to know from Table 1 is that residual module 1: uses a 7x7 convolution kernel, 64 channels, and a stride of 2. This layer is usually the input feature extraction layer, which is used to reduce the size of the input image and extract basic features.

[0047] Residual module 2: Use 3x3 convolution kernel, 64 channels, and stride 2. 3x3 convolution is a standard convolution operation that extracts more complex features.

[0048] Residual module 3: Use 3x3 convolution kernel, 64 channels, and stride 1. The convolution operation of this layer continues feature extraction.

[0049] Residual module 4: also a 3x3 convolution kernel, 64 channels, and a stride of 1. This layer continues feature extraction and maintains the size of the feature map.

[0050] Residual module 5: uses 1x1 convolution kernel and 64 channels. The 1x1 convolution of this layer is used to adjust the number of channels (feature map depth), mainly for dimensionality reduction or dimensionality increase operations without affecting the spatial dimension of the feature map.

[0051] Specifically, 1,000 pantograph images were first annotated, divided into 700 training sets and 300 test sets, and then input into the residual backbone network to extract fault features (arcing and abnormal wear were trained separately). The backbone network sets up 5 residual modules, each of which consists of 2 convolutional layers. The first convolutional layer takes a 3x3 convolution kernel for convolution operation, passes through the ReLu activation function, and then passes through a 3x3 convolutional layer for calculation. After the result is added to the residual value, a single training result is completed.

[0052] According to an embodiment of the present invention, preferably, the residual network algorithm is used to perform multiple sampling from the processed pantograph image to train a residual network model, and further includes: A dynamic learning rate hyperparameter model was used, with an initial learning rate of 0.01, a decay exponent of 0.1, a batch size of 5, and a total of 1500 iterations.

[0053] According to an embodiment of the present invention, preferably, a residual network algorithm is used to perform multiple sampling from the processed pantograph image, and before the residual network model is trained, the pantograph image to be identified is input into the residual network model, and after the pantograph slip state is identified, the method further includes: Confirm the calculation accuracy of the trained residual network model according to the preset accuracy formula; Wherein, the accuracy formula is: .

[0054] It is important to know that the meanings of the symbols are as follows: TP (1 Positive, true class): The number of positive classes correctly predicted by the model.

[0055] TN (1 Negative, true negative class): The number of negative classes correctly predicted by the model.

[0056] FP (0 ​​Positive, False Positive): The number of samples that the model incorrectly predicts as positive (they are actually negative but predicted as positive).

[0057] FN (0 Negative, false negative class): The number of negative classes incorrectly predicted by the model (actually positive classes, but predicted as negative classes).

[0058] The formula can be understood as: The numerator (TP+TN): represents the number of all correct predictions of the model, that is, the sum of the correctly predicted positive and negative classes.

[0059] Denominator (TP+FP+TN+FN): represents the total number of all predictions, including correct and incorrect predictions.

[0060] The higher the accuracy value, the better the model's overall classification effect. The multiplication by 100% in the formula is to express the result in the form of a percentage.

[0061] Step S104, inputting the pantograph image to be identified into the residual network model to identify the sliding state of the pantograph; It can realize automatic and accurate recognition of pantograph slip status.

[0062] According to an embodiment of the present invention, preferably, the pantograph image to be identified is input into the residual network model to identify the sliding state of the pantograph, including: The divided training set is input into the residual network model, and the sliding state of the pantograph is output.

[0063] This application can build a deep learning framework based on residual network only through the video collected by the roof camera. By analyzing the visual collection data and combining the self-learning algorithm of the residual network, the detection and diagnosis of the pantograph can be realized, and an alarm can be issued when the pantograph slips and the fault can be located. The pantograph slip detected in this application is divided into: abnormal arc (spark) and abnormal carbon slide wear.

[0064] like Figure 4-6 As shown, based on the analysis of actual fault data, during the operation of the train, a single roof camera observes the state of the pantograph from an upward perspective. When the pantograph-net state is abnormal, abnormal sparks and arcs appear in the image. When the train returns to the depot, a single roof camera observes the state of the pantograph carbon slide plate from a downward perspective. When the carbon slide plate wear is abnormal / non-abnormal, abnormal thickness and uneven distribution appear in the image.

[0065] When looking down at the pantograph carbon slide plate on one side, it is in normal condition, and the image detects no abnormal wear; When looking down at the abnormal state of one side of the pantograph carbon slide plate, the image detects abnormal wear and marks the corresponding area; When looking up at the pantograph carbon abnormal arc spark detection, the image detects the presence of abnormal arc sparks and an alarm is issued during operation.

[0066] From the above description, it can be seen that the present application achieves the following technical effects: In an embodiment of the present application, a deep learning framework is established based on a residual network, by obtaining a pantograph image when the train is operating normally; performing a preset normalization operation on the pantograph image; using a residual network algorithm to perform multiple sampling from the processed pantograph image to train a residual network model; inputting the pantograph image to be identified into the residual network model to identify the slip state of the pantograph, thereby achieving the purpose of automatically diagnosing and identifying the slip state of the pantograph, thereby realizing timely reporting of abnormal arcs in the pantograph network during operation, reporting abnormal wear of the carbon slide plate during maintenance, and helping users to accurately achieve the technical effect of status maintenance, thereby solving the problem in actual subway operation that, because the pantograph is a purely mechanical component, a large number of sensors are usually required for detection, such as ultraviolet, infrared, laser, etc., and the initial investment is extremely high. In addition, too many sensors cause tight roof space, troublesome sensor failure replacement and maintenance, and regular cleaning results in a huge waste of human resources.

[0067] The present invention also has the following beneficial effects: The present invention can be extended to all urban rail vehicles according to the application requirements of specific scenarios. It can safely and reliably diagnose the sliding status of the pantograph, can promptly report abnormal arcing of the pantograph and the catenary during operation, and can report abnormal wear of the carbon slide plate during maintenance, thereby helping users to accurately implement status maintenance.

[0068] The present application also relates to another aspect, providing a pantograph slippage detection device. Figure 2 As shown, the pantograph slippage detection device comprises: An acquisition module 201 is used to acquire a pantograph image when the train is operating normally; By collecting pantograph images in real time during normal train operation, accurate image data collection can be achieved, thus providing a basis for subsequent processing.

[0069] A normalization module 202, configured to perform a preset normalization operation on the pantograph image; The data samples are normalized, that is, the data is mapped to (0, 1), which improves the convergence speed of the next residual network and reduces the impact of data in different ranges on the model.

[0070] A model training module 203 is used to train a residual network model by sampling multiple times from the processed pantograph image using a residual network algorithm; It can achieve accurate model training results, thus providing a basis and guarantee for subsequent predictions.

[0071] The pantograph slip state recognition module 204 is used to input the pantograph image to be recognized into the residual network model to recognize the slip state of the pantograph.

[0072] It can realize automatic and accurate recognition of pantograph slip status.

[0073] From the above description, it can be seen that the present application achieves the following technical effects: In an embodiment of the present application, a deep learning framework is established based on a residual network, by obtaining a pantograph image when the train is operating normally; performing a preset normalization operation on the pantograph image; using a residual network algorithm to perform multiple sampling from the processed pantograph image to train a residual network model; the pantograph image to be identified is input into the residual network model to identify the slip state of the pantograph, thereby achieving the purpose of automatically diagnosing and identifying the slip state of the pantograph, thereby realizing timely reporting of abnormal arcs in the pantograph network during operation, reporting abnormal wear of the carbon slide plate during maintenance, and helping users to accurately achieve the technical effect of status maintenance, thereby solving the problem in actual subway operation that, because the pantograph is a purely mechanical component, a large number of sensors are usually required for detection, such as ultraviolet, infrared, laser, etc., and the initial investment is extremely high. In addition, too many sensors cause tight roof space, troublesome sensor failure replacement and maintenance, and regular cleaning results in a huge waste of human resources.

[0074] like Figure 7As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, ROM 12 and RAM 13 are connected to each other through a bus 14. The input / output (I / 0) interface 15 is also connected to the bus 14.

[0075] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0076] The processor 11 may be a variety of general and / or dedicated processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a pantograph slippage detection method.

[0077] In some embodiments, the pantograph slip detection method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the pantograph slip detection method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the pantograph slip detection method in any other appropriate manner (e.g., by means of firmware).

[0078] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0079] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0080] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0081] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0082] The systems and techniques described herein may be implemented in a computing system that includes a backend component (e.g., as a data server), or a computing system that includes a middleware component (e.g., an application server), or a computing system that includes a frontend component (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include: a local area network (LA), a wide area network (WA), a blockchain network, and the Internet.

[0083] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0084] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0085] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A pantograph slippage detection method, characterized in that: include: Obtain pantograph images during normal train operation; Performing a preset normalization processing operation on the pantograph image; The residual network algorithm is used to perform multiple sampling from the processed pantograph image to train the residual network model; The pantograph image to be identified is input into the residual network model to identify the sliding state of the pantograph.

2. The pantograph slip detection method according to claim 1, characterized in that: Obtain pantograph images during normal train operation, including: The pantograph images taken when the train is in normal operation are collected. The pantograph images are in two states: a top-down view and an upward view, with 1,000 images each.

3. The pantograph slip detection method according to claim 2, characterized in that: The pantograph images taken by the train from the beginning of normal operation are collected. The pantograph images are in two states: a top-down view and an upward view. After 1,000 images of each state, the following are also included: The obtained pantograph image is divided into a training set for model training and a test set for model testing according to a preset ratio.

4. The pantograph slip detection method according to claim 1, characterized in that: After performing a preset normalization operation on the pantograph image, multiple sampling is performed from the processed pantograph image using a residual network algorithm, and before training a residual network model, the method further includes: Perform preset data cleaning, target data extraction and data normalization operations on the normalized pantograph image.

5. The pantograph slip detection method according to claim 3, characterized in that: The residual network algorithm is used to perform multiple sampling from the processed pantograph image to train the residual network model, including: The divided training set is input into the residual backbone network to extract fault features, and the arcing and abnormal wear detection models are trained respectively; Five residual modules are preset, each of which consists of two convolutional layers. The first convolutional layer takes a 3x3 convolution kernel for convolution operation, passes through the ReLu activation function, and then passes through the second 3x3 convolutional layer for calculation. After the result is added to the residual value, a single training result is completed.

6. The pantograph slip detection method according to claim 5, characterized in that: The residual network algorithm is used to perform multiple sampling from the processed pantograph image to train the residual network model, which also includes: A dynamic learning rate hyperparameter model was used, with an initial learning rate of 0.01, a decay exponent of 0.1, a batch size of 5, and a total of 1500 iterations.

7. The pantograph slip detection method according to claim 6, characterized in that: The residual network algorithm is used to perform multiple sampling from the processed pantograph image. Before training the residual network model, the pantograph image to be identified is input into the residual network model. After the pantograph sliding state is identified, the method further includes: Confirm the calculation accuracy of the trained residual network model according to the preset accuracy formula; Wherein, the accuracy formula is: 。 8. The pantograph slip detection method according to claim 7, characterized in that: Inputting the pantograph image to be identified into the residual network model to identify the sliding state of the pantograph includes: The divided training set is input into the residual network model, and the sliding state of the pantograph is output.

9. A pantograph slippage detection device, characterized in that: include: An acquisition module is used to acquire the pantograph image when the train is operating normally; A normalization module, used for performing a preset normalization processing operation on the pantograph image; A model training module is used to train a residual network model by sampling multiple times from the processed pantograph image using a residual network algorithm; The pantograph slip state recognition module is used to input the pantograph image to be recognized into the residual network model to recognize the slip state of the pantograph.

10. An electronic device, characterized in that The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the pantograph slippage detection method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the pantograph slip detection method according to any one of claims 1 to 8 when executed.

Images