A ternary diagnosis method for influence of basic deformation on ballastless track based on supervised learning

Abstract

The application provides a kind of ternary diagnostic method based on supervised learning of influence of foundation deformation on ballastless track, belongs to the technical field of ballastless track foundation deformation monitoring and diagnosis based on machine learning, obtains the video data of the motion posture of the ballastless track to be identified;The video data of the motion posture of the ballastless track to be identified is input into the trained ternary diagnostic model of the influence of foundation deformation on ballastless track for processing, and the identification result of the influence of foundation deformation on ballastless track is obtained and early warning is carried out.The application shoots the track motion posture video under driving conditions, extracts the key positions of steel rail and ballastless track slab from the motion posture video to establish posture function and oscillate function, thereby obtaining the service state and vibration law of track structure, and by comparing with the track motion posture video and function of normal section ballastless track structure under driving conditions, the service state of track structure is diagnosed, and part of the classical diagnostic results are put into the training set for subsequent diagnosis.

Description

Technical Field

[0001] This invention relates to the field of machine learning-based foundation deformation monitoring and diagnosis technology for ballastless tracks, specifically to a three-dimensional diagnostic method for the impact of foundation deformation on ballastless tracks based on supervised learning. Background Technology

[0002] In sections with complex geological conditions, the high stiffness of ballastless track makes it highly susceptible to adverse effects from foundation deformation. Therefore, predicting or diagnosing foundation deformation is crucial for railway safety and ride comfort. However, the foundation is buried beneath the track structure and not exposed, making monitoring difficult using traditional sensor deployment or non-contact measurement methods. Supervised learning, a type of artificial intelligence machine learning technology, has matured in its techniques for recognizing motion posture videos and functions, as well as feature extraction. Training on a sufficiently large training set ensures diagnostic accuracy, which increases with the expansion of the training set. Summary of the Invention

[0003] The purpose of this invention is to provide a three-dimensional diagnostic method and system for the influence of basic deformation on ballastless tracks based on supervised learning, so as to solve at least one of the technical problems existing in the background art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] In a first aspect, the present invention provides a three-dimensional diagnostic method for the influence of basic deformations on ballastless tracks based on supervised learning, comprising:

[0006] Acquire video data of the motion posture of the ballastless track to be identified;

[0007] The motion posture video data of the ballastless track to be identified is input into a pre-trained three-element diagnostic model for the impact of foundation deformation on the ballastless track. This model is then processed to obtain the results of identifying the impact of foundation deformation on the ballastless track and to issue an early warning. Subsequently, after receiving the early warning information, some classic samples labeled as early warning are added to the training set to retrain the three-element diagnostic model for the impact of foundation deformation on the ballastless track, so as to refine the graded identification label results in the future. The three-element diagnostic model for the impact of foundation deformation on the ballastless track is trained using motion posture video sub-models, mode shape posture function sub-models, and vibration oscillate function sub-models, which are obtained from multiple motion posture video samples, function samples, and safety labels. The three-element diagnostic model for the impact of foundation deformation on the ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe.

[0008] Optionally, the posture function is established as follows: The intersection points of the neutral axis and symmetry axis of each section of the rail and ballastless track slab are obtained from the motion posture video to form a curve. This curve is then fitted into the established Cartesian coordinate system to obtain a preliminary posture function, denoted as posture = f(x,y,z). After expanding the y-direction, the final posture function is obtained, denoted as posture = p(x,e). (y) The oscillate function is established by replacing the time parameter t in the motion posture video with posture = p(x, e). (y) In the function x, we get oscillate=g(t,e) (y) ,z).

[0009] Optionally, multiple sets of motion posture videos of the ballastless track structure under normal geological conditions and train operation conditions are captured and uploaded. The motion posture videos, along with the posture = p(x, e)... (y) ,z), oscillate=g(t,e (y) ,z) are used as benchmark samples in the training set, labeled as safe, to provide a ternary diagnostic model for learning the impact of foundation deformation on ballastless track.

[0010] Optionally, after outputting diagnostic warning information, the basic deformation status of the target section is manually investigated. If it is found that the basic deformation reaches the operational safety threshold or is of great significance, it is regarded as a classic motion posture video sample. The sample and the corresponding significance label are put into the training set to facilitate the subsequent clustering process. That is, the basic deformation status of the target section can be identified and diagnosed in a more refined and hierarchical manner. The significance includes, but is not limited to: the critical value of locomotive operation safety and comfort, cracks in the ballastless track structure or failure of the track structure, etc.

[0011] Optionally, the training method for the ternary diagnostic model of the influence of basic deformation on ballastless track is supervised learning, including but not limited to neural network algorithms, random forest algorithms, support vector machine algorithms, etc.

[0012] Secondly, the present invention provides a three-dimensional diagnostic system for the influence of basic deformation on ballastless tracks based on supervised learning, comprising:

[0013] The acquisition module is used to acquire video data of the motion posture of the ballastless track to be identified;

[0014] The processing module is used for preprocessing of motion posture video samples, such as enhancement and denoising. The preprocessed motion posture video sample data is then input into a trained ternary diagnostic model for the impact of foundation deformation on ballastless track for further processing. This model identifies the impact of foundation deformation on the ballastless track and issues warnings. After receiving warning information, classic samples are selected through manual investigation and assigned classic labels such as "affects driving comfort but not driving safety," "affects driving safety," or "reaches the limit state of the track structure." These classic motion posture video samples and their corresponding labels are added to the training set to retrain the ternary diagnostic model for the impact of foundation deformation on ballastless track, allowing for subsequent refinement of the hierarchical identification label results. The training of the ternary diagnostic model for the impact of foundation deformation on ballastless track includes a motion posture video sub-model, a mode shape posture function sub-model, and a vibration oscillate function sub-model, all trained from multiple motion posture video samples, function samples, and safety labels. The ternary diagnostic model for the impact of foundation deformation on ballastless track outputs a final "safety label" only when all three sub-models are deemed safe.

[0015] Furthermore, the acquisition module includes: a set of non-parallel high-definition cameras and a processing module pre-installed near the section being tested, capable of acquiring complete segment motion posture video. The high-definition cameras are powered independently; when their power level drops below 5%, a power alarm is triggered, prompting a power replacement reminder. The processing module performs pre-processing and uploading of the captured motion posture video. Recording begins when the high-definition camera detects track vibration and ends when it detects track cessation of vibration, thus acquiring a single motion posture video sample.

[0016] Thirdly, the present invention provides a non-transitory computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the ternary diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described in the first aspect.

[0017] Fourthly, the present invention provides a computer device including a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions executable by the processor, and the processor invokes the program instructions to execute the ternary diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described in the first aspect.

[0018] Fifthly, the present invention provides an electronic device, comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the ternary diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described in the first aspect.

[0019] The beneficial effects of this invention are as follows: By employing motion posture video shooting technology, spatial curve establishment, function fitting, and variable substitution, a three-dimensional real-time monitoring and early warning system for foundation deformation is achieved. This method has the advantages of being non-contact, highly efficient, highly accurate, and widely applicable. It can significantly reduce the cost and labor intensity of manual inspections, improve the safe operation level of railways, and provide a scientific basis for guiding engineering design, maintenance, and safety assessment.

[0020] The advantages of additional aspects of the invention will be set forth more clearly in the following description or will be learned by practice of the invention. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments 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.

[0022] Figure 1 This is a flowchart of a ternary diagnostic method for the influence of basic deformation on ballastless tracks based on supervised learning, as described in an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of the rail's axis of symmetry and neutral axis as described in an embodiment of the present invention.

[0024] Figure 3 This is a schematic diagram illustrating the establishment of a spatial Cartesian coordinate system according to an embodiment of the present invention.

[0025] Figure 4 This is a flowchart illustrating the identification process of the ternary diagnostic method for the influence of basic deformation on ballastless tracks based on supervised learning, as described in an embodiment of the present invention. Detailed Implementation

[0026] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0027] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0028] It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as here.

[0029] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, and / or groups thereof.

[0030] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. Furthermore, the described specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples. Without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0031] To facilitate understanding of the present invention, the present invention will be further explained and described below with reference to the accompanying drawings and specific embodiments. However, the specific embodiments do not constitute a limitation on the embodiments of the present invention.

[0032] Those skilled in the art should understand that the accompanying drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily essential for implementing the present invention.

[0033] Example 1

[0034] In this embodiment 1, a supervised learning-based diagnostic system for the impact of basic deformation on ballastless track is provided. The system includes: an acquisition module for acquiring video data of the motion posture of the ballastless track to be identified; and a processing module for inputting the video data of the motion posture of the ballastless track to be identified into a pre-trained ternary diagnostic model for the impact of basic deformation on ballastless track, processing the data to obtain the results of identifying the impact of basic deformation on ballastless track, and issuing an early warning. After obtaining the early warning information, some classic samples labeled as early warning are added to the training set to retrain the ternary diagnostic model for the impact of basic deformation on ballastless track, so as to refine the graded identification label results later. The training of the ternary diagnostic model for the impact of basic deformation on ballastless track includes a motion posture video sub-model, a mode shape posture function sub-model, and a vibration oscillate function sub-model, which are obtained from multiple motion posture video samples, function samples, and safety labels. The ternary diagnostic model for the impact of basic deformation on ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe.

[0035] The acquisition module includes a set of non-parallel high-definition cameras and a processing module pre-installed near the test section to acquire complete motion posture video of the section. Each high-definition camera is powered independently; when its power level drops below 5%, a power alarm is triggered, prompting a power replacement reminder. The processing module performs pre-processing and uploading of the captured motion posture video. Recording begins when the high-definition camera detects track vibration and ends when it detects the track stopping vibration, thus acquiring a single motion posture video sample.

[0036] In this embodiment, based on the above-described system, a method for diagnosing the impact of basic deformation on ballastless track based on supervised learning is implemented. This includes: acquiring video data of the motion posture of the ballastless track to be identified; inputting the video data of the motion posture of the ballastless track to be identified into a trained ternary diagnostic model of the impact of basic deformation on ballastless track for processing, obtaining the results of identifying the impact of basic deformation on ballastless track, and issuing an early warning; wherein, after obtaining the early warning information, some classic samples labeled as early warning are added to the training set to retrain the ternary diagnostic model of the impact of basic deformation on ballastless track, so as to refine the graded identification label results subsequently; wherein, the training of the ternary diagnostic model of the impact of basic deformation on ballastless track includes the motion posture video sub-model, the mode shape posture function sub-model, and the vibration oscillate function sub-model, which are obtained from multiple motion posture video samples, function samples, and safety labels. The ternary diagnostic model of the impact of basic deformation on ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe.

[0037] The process of establishing the posture function is as follows: The intersection points of the neutral axis and symmetry axis of each section of the rail and ballastless track slab are obtained from the motion posture video, forming a curve. This curve is then fitted into an established Cartesian coordinate system to obtain a preliminary posture function, denoted as posture = f(x,y,z). After expanding the y-direction, the final posture function is obtained, denoted as posture = p(x,e). (y) The oscillate function is established by replacing the time parameter t in the motion posture video with posture = p(x, e). (y) In the function x, we get oscillate=g(t,e) (y) ,z).

[0038] Multiple sets of videos showing the motion posture of the ballastless track structure under normal geological conditions and during train operation were filmed and uploaded. The motion posture videos, along with the posture=p(x,e) parameter, were then uploaded. (y) ,z), oscillate=g(t,e (y) ,z) are used as benchmark samples in the training set, labeled as safe, to provide a ternary diagnostic model for learning the impact of foundation deformation on ballastless track.

[0039] After the diagnostic warning information is output, the basic deformation status of the target section is manually investigated. If it is found that the basic deformation reaches the operational safety threshold or is of great significance, it is regarded as a classic sample and placed into the training set to facilitate the subsequent clustering process. This allows for more refined and hierarchical identification and diagnosis of the basic deformation status of the target section. The great significance includes, but is not limited to: the critical value for locomotive operation safety and comfort, cracks in the ballastless track structure, or damage and failure of the track structure.

[0040] The training method for the ternary diagnostic model of the influence of basic deformation on ballastless track is supervised learning, including but not limited to neural network algorithms, random forest algorithms, support vector machine algorithms, etc.

[0041] Example 2

[0042] This embodiment 2 provides a ternary diagnostic method for the impact of basic deformation on ballastless track based on supervised learning. It is a technique for diagnosing the service status of track by capturing the motion posture of the track under train conditions. The core of this technique is to capture the motion posture video of the track under train conditions, and extract the key positions of the rails and ballastless track slabs from the motion posture video to establish posture and oscillate functions, thereby obtaining the service status and vibration law of the track structure. By comparing the motion posture video and functions of the ballastless track structure under train conditions in normal sections, the service status of the track structure is diagnosed, and some classic diagnostic results are put into the training set for subsequent diagnosis.

[0043] The ternary diagnostic method for the impact of foundation deformation on ballastless track based on supervised learning includes the following steps: acquiring motion posture video samples in the training set, a ternary diagnostic model for the impact of foundation deformation on ballastless track, the motion posture video to be identified, the mode shape posture function, and the vibration oscillate function. The ternary diagnostic model for the impact of foundation deformation on ballastless track is trained from multiple motion posture video samples, function samples, and safety labels, and includes three sub-models: a motion posture video sub-model, a mode shape posture function sub-model, and a vibration oscillate function sub-model. The motion posture video to be identified is input into the trained ternary diagnostic model for the impact of foundation deformation on ballastless track to obtain identification labels and issue corresponding warnings. After obtaining warning information, some classic samples with warning labels are added to the training set to retrain the model, so as to refine the hierarchical identification label results in subsequent steps.

[0044] The process of establishing the posture function is as follows: The intersection points of the neutral axis and symmetry axis of each section of the rail and ballastless track slab are obtained from the motion posture video, forming a curve. This curve is then fitted into an established Cartesian coordinate system to obtain a preliminary posture function, denoted as posture = f(x,y,z). After expanding the y-direction, the final posture function is obtained, denoted as posture = p(x,e). (y) ,z).

[0045] The oscillate function is established as follows: The time parameter t in the motion pose video is replaced with posture = p(x, e). (y) In the function x, we get oscillate=g(t,e) (y) ,z).

[0046] Multiple sets of videos showing the motion posture of the ballastless track structure under normal geological conditions and during train operation were filmed and uploaded. The motion posture videos, along with the posture=p(x,e) parameter, were then uploaded. (y) ,z), oscillate=g(t,e (y) ,z) are used as benchmark samples in the training set, labeled as safe, to provide a ternary diagnostic model for learning the impact of foundation deformation on ballastless track.

[0047] Both the training set of motion posture video samples and the motion posture video to be identified are preprocessed after acquisition, namely denoising and enhancement, to improve the quality of motion posture video and feature extraction.

[0048] The training method for the ternary diagnostic model of the influence of basic deformation on ballastless track is supervised learning, including but not limited to neural network algorithms, random forest algorithms, support vector machine algorithms, etc.

[0049] The three-dimensional diagnostic model for the impact of foundation deformation on ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe.

[0050] The classic motion posture video samples are obtained as follows: After outputting diagnostic warning information, the basic deformation state of the target section is manually investigated. If the basic deformation reaches the operational safety threshold or has other significant implications, it is considered a classic sample and added to the training set along with the surplus critical labels obtained from the manual investigation. This facilitates the subsequent clustering process, enabling refined hierarchical identification and diagnosis of the basic deformation state of the target section. These significant implications include, but are not limited to: critical values ​​for locomotive operation safety and comfort, cracks in the ballastless track structure, or track structure failure.

[0051] A set of non-parallel high-definition cameras and a processing module are pre-installed near the section being tested to acquire complete motion posture video of that section. Each high-definition camera is powered independently; when its power level drops below 5%, a power alarm is triggered, prompting a power replacement reminder. The processing module handles pre-processing and uploading of the captured motion posture video. Recording begins when the high-definition camera detects track vibration and ends when it detects the track stopping vibration, thus acquiring a single motion posture video sample.

[0052] Track motion attitude refers to the dynamic deformation of the track within the test section from the moment the track begins to vibrate until it stops vibrating, captured by a high-definition camera.

[0053] This method can be applied to ballastless track structures in any location prone to foundation deformation.

[0054] Example 3

[0055] Due to the characteristics of ballastless track foundations—hidden beneath the track structure, exhibiting irregular changes, and being difficult to predict—traditional monitoring methods struggle to diagnose foundation deformation. Figures 1 to 4 As shown, this embodiment provides a three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning. By capturing motion posture videos of the track to reflect its service status, the method ultimately achieves the purpose of inverting basic deformation.

[0056] During the installation phase of each module: High-definition cameras and processing modules are installed at the target location to ensure that the cameras can acquire clear and complete track motion attitudes. Preprocessing operations such as noise reduction and enhancement are performed in the processing module. The more reference samples, the better, allowing the three-dimensional diagnostic model of the impact of foundation deformation on the ballastless track to learn fully and improve diagnostic accuracy.

[0057] Upload the pre-processed motion posture video samples. In the initial stage after the ballastless track construction is completed, it is assumed that no deformation of the underlying foundation will occur. Therefore, all samples obtained at this stage are labeled "safe".

[0058] A sufficient number of motion posture video samples were captured to establish a spatial Cartesian coordinate system. The origin was set at the intersection of the neutral axis of the target section rail and the ballastless track slab with the axis of symmetry of the rail cross-section. The train's operating direction was the positive z-axis, and the vertically upward direction was the positive y-axis. The x-axis was established following the "right-hand rule." The neutral axis positions of the target section rail and ballastless track slab in the motion posture videos were obtained, generating their respective spatial position-mileage curves. These curves were then fitted to a mode shape posture space function, denoted as posture=f(x,y,z). Considering that the influence of foundation deformation on the ballastless track structure is mainly reflected in the y-direction, to amplify this feature, and because the natural exponential function has a rapid growth rate and its derivatives are consistent with the original function, the above function was modified to: posture=p(x,e (y) ,z).

[0059] The neutral axis was chosen as the primary feature representing the rail because the shear stress at the neutral axis is zero, reducing the impact of the rail and ballastless track slab's own torsion on the model. The axis of symmetry was chosen to represent the rail and ballastless track slab features because rail edges experience wear, cracks, spalling, and core damage during operation, while ballastless track slabs are prone to cracking and separation. Therefore, the axis of symmetry, located at the very center, provides clear features and is easy to obtain. The natural exponential function was chosen as the amplification function in the y-direction because its growth rate increases with the independent variable, making it a convex function with a significant amplification effect in the y-direction. Furthermore, its derivatives are consistent with the original function, meaning that differentiation does not change the features. The original function and its derivatives are important features of the posture function, which is extremely beneficial for the accuracy of subsequent feature extraction and recognition.

[0060] To account for dynamic factors, the above posture = p(x,e) is... (y) Replace x in the function with the time parameter t from the motion posture video, and establish the vibration oscillate = g(t, e) (y) The vibration function (x, y) reflects the change of rail vibration in the y-direction over time. The reasons for selecting the independent variable of the vibration function are as follows: foundation deformation mainly affects the vertical direction of the rail, so the y-direction must be retained; although the change in the z-direction is negligible relative to the rail length, it reflects the rail mileage and also needs to be retained; the x-direction reflects the rail lateral direction, and foundation deformation has a relatively small impact on this direction, and its change over time is not significant, so x is replaced by the time parameter t.

[0061] Using 80% of the aforementioned motion posture video samples and safety labels as the training set, features of track motion posture in the motion posture videos are extracted, including vibration frequency and mode shape. Using 80% of the aforementioned posture function samples as the training set, dynamic features are extracted, including the function's expression and the sign of its first derivative. Using 80% of the aforementioned oscillate function samples as the training set, dynamic features are extracted, including the function's expression and the sign of its first derivative. A supervised learning algorithm is used to train a ternary diagnostic model of the impact of foundation deformation on ballastless track. The feature vectors in the ternary diagnostic model of the impact of foundation deformation on ballastless track are trained using the aforementioned features and safety labels.

[0062] Twenty percent of the motion posture video samples are used as the test set and input into the pre-trained ternary diagnostic model encoder for the influence of basic deformation on ballastless track to obtain feature vectors, which are then output as "safety labels" by the decoder. Similarly, 20% of the posture function samples described in section 3 are used as the test set and input into the pre-trained ternary diagnostic model for the influence of basic deformation on ballastless track to obtain "safety labels". Twenty percent of the oscillate function samples described in section 3 are used as the test set and input into the pre-trained ternary diagnostic model for the influence of basic deformation on ballastless track to obtain "safety labels". The model is proven to have sufficient accuracy if and only if all three conditions are met simultaneously.

[0063] After training on the training set and validating on the test set, the basic deformation ternary diagnostic model can output corresponding "safety labels" or "warning labels" when the motion posture video to be tested is input into the model. The motion posture video to be identified is preprocessed to obtain the final posture = p(x, e) (y) ,z) followed by oscillate=g(t,e (y) The three-dimensional diagnostic model for the impact of foundation deformation on ballastless track is constructed by simultaneously inputting these three factors (safety label, warning label, and oscillation label) into the model. The model outputs either a "safety label" or a "warning label." When the model outputs a "warning label," a field survey is conducted to investigate the state of foundation deformation at the target location. When the foundation deformation reaches the threshold values ​​for each stage or has special significance, the corresponding motion posture video is assigned a corresponding foundation deformation label and considered a classic sample, which is then added to the training set to retrain the three-dimensional diagnostic model for the impact of foundation deformation on ballastless track. After a sufficient number of samples, the foundation deformation model iterates from initially only being able to output "safety labels" and "warning labels" to being able to output both "safety labels" and labels indicating that the foundation deformation has reached each threshold. Ultimately, the model achieves the function of inverting the state of the underlying foundation deformation by using motion posture videos and the service status of the track via the `posture` and `oscillate` functions.

[0064] Specifically, in this embodiment, after the ballastless track structure of the target section is completed, assuming no foundation deformation occurs, a set of non-parallel high-definition cameras and processing modules are installed nearby. Under operating conditions, the installed high-definition cameras acquire a sufficient number of preliminary track motion posture videos, which are then pre-processed with noise reduction and enhancement before being uploaded to obtain safety tags and corresponding motion posture videos. (See attached...) Figure 3 As shown, a spatial Cartesian coordinate system is established. The intersection points of the neutral axis and symmetry axis of each section of the rail and ballastless track slab in the aforementioned motion posture video are extracted to form curves. These curves are then fitted to obtain the preliminary rail vibration mode posture space function posture=f(x,y,z) corresponding to the aforementioned motion posture video. After improvement by increasing the y-direction, the final posture=p(x,e) is obtained. (y) The established posture = p(x,e) (y) oscillate = g(t, e) and the time parameters in the motion posture video to establish oscillate = g(t, e) (y) The model uses a ternary diagnostic model to assess the impact of foundation deformation on ballastless track. A neural network algorithm is trained using extracted motion posture video features. The model's diagnostic performance is optimized by continuously adjusting its parameters. After model validation and optimization, the ternary diagnostic model for the impact of foundation deformation on ballastless track is preliminarily trained and can diagnose whether foundation deformation has occurred, outputting a "safety label" or "warning label." High-definition cameras are used to acquire and preprocess motion posture videos to be identified. These videos are then input into the preliminarily trained ternary diagnostic model to diagnose foundation deformation, outputting the corresponding label for the tested sample. After the model issues a warning, a manual on-site investigation is conducted to assess the foundation deformation. If the deformation reaches various thresholds for safe operation or other significant levels, the preprocessed sample and its corresponding label are added to the training set as classic samples. At this point, the ternary diagnostic model for the impact of foundation deformation on ballastless track can identify the corresponding motion posture video samples under this foundation deformation condition, providing accurate warnings and outputting a "warning label" along with the foundation deformation status.

[0065] In summary, this embodiment provides a three-dimensional diagnostic method for the impact of foundation deformation on ballastless track based on supervised learning. By employing motion posture video recording technology and function fitting, it achieves real-time three-dimensional monitoring and early warning of foundation deformation. This method has advantages such as being non-contact, highly efficient, highly accurate, and widely applicable. It can significantly reduce the cost and labor intensity of manual inspections, improve the safe operation level of railways, and provide a scientific basis for guiding engineering design, maintenance, and safety assessment.

[0066] Example 4

[0067] This embodiment 4 provides a non-transitory computer-readable storage medium for storing computer instructions. When executed by a processor, the computer instructions implement the three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning, as described above. The method includes:

[0068] Acquire video data of the motion posture of the ballastless track to be identified;

[0069] The motion posture video data of the ballastless track to be identified is input into a pre-trained three-element diagnostic model for the impact of foundation deformation on the ballastless track. This model is then processed to obtain the results of identifying the impact of foundation deformation on the ballastless track and to issue an early warning. Subsequently, after receiving the early warning information, some classic samples labeled as early warning are added to the training set to retrain the three-element diagnostic model for the impact of foundation deformation on the ballastless track, so as to refine the graded identification label results in the future. The three-element diagnostic model for the impact of foundation deformation on the ballastless track is trained using motion posture video sub-models, mode shape posture function sub-models, and vibration oscillate function sub-models, which are obtained from multiple motion posture video samples, function samples, and safety labels. The three-element diagnostic model for the impact of foundation deformation on the ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe.

[0070] Example 5

[0071] This embodiment 5 provides a computer device, including a memory and a processor, wherein the processor and the memory communicate with each other, and the memory stores program instructions that can be executed by the processor. The processor calls the program instructions to execute the three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described above. The method includes:

[0072] Acquire video data of the motion posture of the ballastless track to be identified;

[0073] The motion posture video data of the ballastless track to be identified is input into a pre-trained three-element diagnostic model for the impact of foundation deformation on the ballastless track. This model is then processed to obtain the results of identifying the impact of foundation deformation on the ballastless track and to issue an early warning. Subsequently, after receiving the early warning information, some classic samples labeled as early warning are added to the training set to retrain the three-element diagnostic model for the impact of foundation deformation on the ballastless track, so as to refine the graded identification label results in the future. The three-element diagnostic model for the impact of foundation deformation on the ballastless track is trained using motion posture video sub-models, mode shape posture function sub-models, and vibration oscillate function sub-models, which are obtained from multiple motion posture video samples, function samples, and safety labels. The three-element diagnostic model for the impact of foundation deformation on the ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe.

[0074] Example 6

[0075] This embodiment 6 provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described above. The method includes:

[0076] Acquire video data of the motion posture of the ballastless track to be identified;

[0077] The motion posture video data of the ballastless track to be identified is input into a pre-trained three-element diagnostic model for the impact of foundation deformation on the ballastless track. This model is then processed to obtain the results of identifying the impact of foundation deformation on the ballastless track and to issue an early warning. Subsequently, after receiving the early warning information, some classic samples labeled as early warning are added to the training set to retrain the three-element diagnostic model for the impact of foundation deformation on the ballastless track, so as to refine the graded identification label results in the future. The three-element diagnostic model for the impact of foundation deformation on the ballastless track is trained using motion posture video sub-models, mode shape posture function sub-models, and vibration oscillate function sub-models, which are obtained from multiple motion posture video samples, function samples, and safety labels. The three-element diagnostic model for the impact of foundation deformation on the ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe.

[0078] 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.

[0079] 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 1 A device that provides the functions specified in one or more boxes.

[0080] 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.

[0081] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment, whereby a series of operational steps are performed 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.

[0082] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that, based on the technical solutions disclosed in the present invention, various modifications or variations that can be made by those skilled in the art without creative effort should be included within the scope of protection of the present invention.

Claims

1. A three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning, characterized in that, include: Acquire video data of the motion posture of the ballastless track to be identified; The motion posture video data of the ballastless track to be identified is input into a pre-trained three-element diagnostic model for the impact of foundation deformation on the ballastless track. This model is then processed to obtain the results of identifying the impact of foundation deformation on the ballastless track and to issue early warnings. Subsequently, after receiving early warning information, some classic samples labeled as "warning" are added to the training set to retrain the three-element diagnostic model for the impact of foundation deformation on the ballastless track, in order to further refine the graded identification label results. The training of the three-element diagnostic model for the impact of foundation deformation on the ballastless track involves three sub-models: the motion posture video sub-model, the mode shape posture function sub-model, and the vibration oscillate function sub-model. These sub-models are trained using multiple motion posture video samples, function samples, and safety labels. The three-element diagnostic model for the impact of foundation deformation on the ballastless track outputs the final "safety label" only when all three sub-models are deemed safe. The posture function is established by obtaining the intersection points of the neutral axis and symmetry axis of each section of the rail and ballastless track slab from the motion posture video, forming curves, and fitting these curves in an established Cartesian coordinate system to obtain preliminary results. Function, denoted as After expanding the y-axis, the final result is obtained. Function, denoted as The oscillate function is built by replacing the time parameter t in the motion posture video with... x in the function, we get .

2. The three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning according to claim 1, characterized in that, Multiple videos of the ballastless track structure under normal geological conditions and in operation were filmed and uploaded. , As a baseline sample, it is included in the training set and labeled as safe, for the ternary diagnostic model of the impact of foundation deformation on ballastless track to learn.

3. The three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning according to claim 1, characterized in that, After the diagnostic warning information is output, the basic deformation status of the target section is manually investigated. If it is found that the basic deformation reaches the operational safety threshold or is of great significance, it is regarded as a classic sample and placed into the training set to facilitate the subsequent clustering process, that is, to refine the hierarchical identification and diagnosis of the basic deformation status of the target section. Among them, the great significance includes, but is not limited to: the critical value of locomotive operation safety and comfort, cracks in the ballastless track structure or damage and failure of the track structure.

4. The three-dimensional diagnostic method for the influence of basic deformation on ballastless track based on supervised learning according to claim 1, characterized in that, The training method for the ternary diagnostic model of the impact of basic deformation on ballastless track is supervised learning, including but not limited to neural network algorithms, random forest algorithms, or support vector machine algorithms.

5. A three-dimensional diagnostic system for the influence of basic deformation on ballastless track based on supervised learning, characterized in that, include: The acquisition module is used to acquire video data of the motion posture of the ballastless track to be identified; The processing module is used to input the motion posture video data of the ballastless track to be identified into a pre-trained three-element diagnostic model of the impact of foundation deformation on the ballastless track for processing, to obtain the identification results of the impact of foundation deformation on the ballastless track and to issue an early warning. After receiving the early warning information, some classic samples labeled as early warning are added to the training set to retrain the three-element diagnostic model of the impact of foundation deformation on the ballastless track, so as to refine the graded identification label results later. The training of the three-element diagnostic model of the impact of foundation deformation on the ballastless track includes motion posture video sub-models, mode shape posture function sub-models, and vibration oscillate function sub-models, which are obtained from multiple motion posture video samples, function samples, and safety labels. The three-element diagnostic model of the impact of foundation deformation on the ballastless track outputs the final "safety label" only when all three sub-models are determined to be safe. The posture function establishment process is as follows: the intersection points of the neutral axis and symmetry axis of each section of the rail and ballastless track slab are obtained from the motion posture video, forming a curve, and then fitted in an established Cartesian coordinate system to obtain a preliminary... Function, denoted as After expanding the y-axis, the final result is obtained. Function, denoted as The oscillate function is built by replacing the time parameter t in the motion posture video with... x in the function, we get .

6. The three-dimensional diagnostic system for the influence of basic deformation on ballastless track based on supervised learning according to claim 5, characterized in that, The acquisition module includes: a set of non-parallel high-definition cameras and a processing module pre-installed near the section being tested, capable of acquiring complete motion posture video of the section; the high-definition cameras are powered independently, and when the power of the high-definition camera is below 5%, a power alarm is triggered, issuing a reminder to replace the power supply; the processing module performs pre-processing and uploading functions after the motion posture video is captured; when the high-definition camera captures track vibration, it triggers the start of recording, and when the high-definition camera captures track cessation of vibration, it triggers the end of recording, thereby acquiring a single motion posture video sample.

7. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the ternary diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described in any one of claims 1-4.

8. A computer device, characterized in that, The system includes a memory and a processor, which communicate with each other. The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the ternary diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described in any one of claims 1-4.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the ternary diagnostic method for the influence of basic deformation on ballastless track based on supervised learning as described in any one of claims 1-4.

Images