A detection method and device for an anti-falling cable at high altitude

Abstract

The present invention relates to the technical field of cable detection, and specifically to a detection method and device for anti-falling cables at high altitudes, including: obtaining the coded waveform of the cable based on a radar array and constructing an observation matrix; constructing a high-dimensional feature space based on the observation matrix and calculating the embedded data; mapping the embedded data into a surface topological structure to calculate a critical threshold and obtaining the eigenvalue of the critical threshold; obtaining the change index of the eigenvalue to establish a defect feature spectrum, inputting the defect feature spectrum into a pulsed neural network for time series coding to generate a defect probability cloud map; extracting the spatial and temporal features in the defect probability cloud map to form a defect type tensor; performing an evolution process on the defect type tensor, and calculating the evolution process based on a numerical solution method to obtain a repair strategy matrix, where the repair strategy matrix includes the adjustment schemes of repair measures at different spatial points and times. This application can effectively prevent and timely respond to cable faults, and reduce potential safety hazards.

Description

Technical Field

[0001] The present invention relates to the technical field of cable detection, and in particular to a detection method and device for a high-altitude anti-fall cable. Background Art

[0002] High-altitude anti-fall cables are widely used in modern urban construction in the fields of power transmission, communication network construction, and safety protection of large buildings. These cables face complex environmental conditions during operation, including wind, rain, temperature changes, ultraviolet radiation and other external factors. They are prone to aging, breakage or loosening. When the cables fail or fail, it may cause serious economic losses and may even endanger people's lives. Therefore, how to achieve real-time and accurate detection of high-altitude cables and timely discover and warn potential defects has always been a technical problem that needs to be solved in the engineering field.

[0003] At present, most of the detection technologies for high-altitude anti-fall cables are through physical detection methods, such as visual inspection, vibration analysis, temperature monitoring or mechanical stress monitoring. Although they can reflect the status of the cables to a certain extent, they have low detection accuracy and cannot fully cover the entire life cycle of the cables. For example, vibration-based detection methods are difficult to capture tiny defects or damage, especially in complex environments. Signals are easily interfered by noise, resulting in insufficient reliability of detection results, making it difficult to accurately identify different types of defects, and unable to meet the high requirements of modern engineering for cable safety detection.

[0004] Therefore, building an efficient, accurate and dynamically updated high-altitude fall protection cable detection and repair system is still a key issue that needs to be urgently solved by current technology. Summary of the invention

[0005] The main purpose of the present invention is to provide a detection method for high-altitude anti-fall cables, aiming to overcome the technical problem of low detection accuracy of existing cable detection methods.

[0006] In order to solve the above technical problems, the present invention proposes a detection method for a high-altitude anti-fall cable, comprising:

[0007] Constructing a radar array with a preset working frequency, acquiring a coding waveform of the cable based on the radar array, and constructing an observation matrix according to the coding waveform;

[0008] Constructing a high-dimensional feature space based on the measurement matrix, and calculating embedded data of the high-dimensional phase space;

[0009] Mapping the embedded data to a surface topology structure, calculating a critical threshold of the surface topology structure, and obtaining a characteristic value of the critical threshold in the surface topology structure;

[0010] Obtain the change index of the eigenvalue, establish a defect feature spectrum according to the change index, input the defect feature spectrum into a pulsed neural network for time series encoding, and generate a defect probability cloud map;

[0011] Map the probability value of each spatial point in the defect probability cloud map to a specific position in three-dimensional space, extract the spatial features and time features in the defect probability cloud map, and form a defect type tensor;

[0012] Perform an evolution process on the defect type tensor, and calculate the evolution process based on a numerical solution method to obtain a repair strategy matrix, where the repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times.

[0013] Further, the steps of constructing a radar array with a preset operating frequency, obtaining an encoded waveform of a cable based on the radar array, and constructing an observation matrix according to the encoded waveform include:

[0014] Perform modulation configuration on a radar array with a preset operating frequency, generate a radar signal based on a set frequency range, and generate a radar signal with a specific periodicity in frequency through the modulation process;

[0015] Parallelly collect the reflected signals of the radar signal based on multiple receiving channels, and perform time synchronization processing on the multiple reflected signals to obtain the encoded waveform;

[0016] Extract the signal intensity information at each moment in the encoded waveform, and convert the reflected signal into a discrete signal sequence according to the signal intensity information;

[0017] Set a time window with a preset number of bits, control the time window to slide on the discrete signal sequence, and generate multiple discrete signals;

[0018] Perform time-frequency transformation on the discrete signal to obtain multiple frequency features, and construct an observation matrix according to the signal amplitude and phase information of each time-frequency point in the frequency features.

[0019] Further, the steps of constructing a high-dimensional feature space based on the observation matrix and calculating the embedded data in the high-dimensional phase space include:

[0020] Perform matrix decomposition on the observation matrix to obtain a set of basis vectors, where each basis vector represents a potential frequency feature;

[0021] Sort the basis vectors in the set of basis vectors, and select the basis vectors whose sorting exceeds a preset value as the main features;

[0022] Perform a linear combination of the multiple main features to form a feature space, and map the data in the feature space through inner product operation to obtain a preliminary representation of the high-dimensional data space;

[0023] Perform phase space reconstruction on the obtained high-dimensional data, and calculate the embedded data according to the dynamic relationship between the dimension of the reconstructed high-dimensional phase space and the original data.

[0024] Further, the step of mapping the embedded data into a surface topological structure, calculating the critical threshold of the surface topological structure, and obtaining the eigenvalue of the critical threshold in the surface topological structure includes:

[0025] Extract the dimensional features of the embedded data, and determine the initial parameters of the surface topological structure based on the dimensional features;

[0026] Based on the topological transformation algorithm, associate each embedded data point in the high-dimensional phase space with a corresponding point on the initial parameters to obtain a set of mapped data points;

[0027] Calculate the local features of the surface topological structure according to the set of mapped data points, and obtain the position features of each mapped data point on the surface topological structure;

[0028] Calculate the overall curvature of the surface topological structure according to the position features, and obtain the high-curvature region and the low-curvature region on the surface;

[0029] Identify the distribution information of the high-curvature region and the low-curvature region, obtain the critical points based on the distribution information, and calculate the critical threshold of the surface topological structure based on the critical points;

[0030] Segment the surface topological structure according to the critical threshold to obtain multiple sub-regions;

[0031] Calculate the density features of the set of mapped data points in each sub-region, and obtain the eigenvalues in each sub-region according to the density features.

[0032] Further, the step of obtaining the change index of the eigenvalue, establishing a defect feature spectrum according to the change index, and inputting the defect feature spectrum into a pulse neural network for time series encoding to generate a defect probability cloud map includes:

[0033] Perform time series analysis on the eigenvalue, calculate the change amount between adjacent eigenvalues, and obtain the change index of the eigenvalue;

[0034] Construct a time series model according to the change index, and convert the time series model into a frequency domain representation to obtain a defect feature spectrum;

[0035] Construct a spiking neural network structure, use the defect feature spectrum as the input signal of the spiking neural network, and perform temporal encoding through the spiking and transmission processes of neurons;

[0036] According to the encoding result, map the defect features at each time point onto a two-dimensional plane to generate a defect probability cloud map, where the density of the defect probability cloud map represents the probability of the defect occurrence.

[0037] Further, the step of mapping the probability value of each spatial point in the defect probability cloud map to a specific position in three-dimensional space, extracting the spatial features and temporal features in the defect probability cloud map, and forming a defect type tensor includes:

[0038] Perform spatial analysis on the defect probability cloud map, use the probability value of each spatial point in the cloud map as coordinate information, map it to a three-dimensional space coordinate system, and obtain a set of specific position points in three-dimensional space;

[0039] Extract spatial features based on the set of specific position points, where the spatial features include the distribution density, spatial distance, and spatial direction of spatial points;

[0040] Extract the temporal features in the defect probability cloud map, where the temporal features include the change trend and periodicity of time points;

[0041] Fuse the spatial features and temporal features to construct a defect type tensor, where the dimensions of the tensor include the spatial dimension and the temporal dimension, and the value of each dimension represents the corresponding feature information.

[0042] Further, the step of performing evolution processing on the defect type tensor and calculating the evolution process based on a numerical solution method to obtain a repair strategy matrix, where the repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times includes:

[0043] Obtain an initial value based on the cable material characteristics, and set the repair parameters of the defect type tensor based on the initial value;

[0044] Construct an evolution model according to the defect type tensor, input the repair parameters into the evolution model, calculate the distribution state of the repair parameters at each time step, and obtain frame-by-frame intermediate evolution results;

[0045] Perform optimization processing on the intermediate evolution results based on the particle swarm optimization algorithm to obtain the optimal combination of repair measure parameters at each spatial point;

[0046] Construct an initial repair matrix based on the combination of repair measure parameters, perform iterative update on the initial repair matrix, and adjust the repair measure parameters according to the repair effect at each time step to obtain a repair strategy matrix.

[0047] The present invention also provides a detection device for high-altitude anti-falling cables, comprising:

[0048] An acquisition module, configured to construct a radar array with a preset operating frequency, obtain a coded waveform of the cable based on the radar array, and construct an observation matrix according to the coded waveform;

[0049] A calculation module, configured to construct a high-dimensional feature space based on the observation matrix and calculate embedded data in the high-dimensional phase space;

[0050] A mapping module, configured to map the embedded data into a surface topological structure, calculate a critical threshold of the surface topological structure, and obtain an eigenvalue of the critical threshold in the surface topological structure;

[0051] A generation module, configured to obtain a change index of the eigenvalue, establish a defect feature spectrum according to the change index, input the defect feature spectrum into a pulsed neural network for time series coding, and generate a defect probability cloud map;

[0052] An extraction module, configured to map the probability value of each spatial point in the defect probability cloud map to a specific position in a three-dimensional space, extract spatial features and time features in the defect probability cloud map, and form a defect type tensor;

[0053] An evolution module, configured to perform evolution processing on the defect type tensor and calculate an evolution process based on a numerical solution method to obtain a repair strategy matrix, where the repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times.

[0054] The present invention also provides a computer device, which includes a processor, a memory, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above method is implemented.

[0055] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above method is implemented.

[0056] Beneficial effects:

[0057] A detection method for high-altitude anti-fall cables proposed in this application can accurately obtain the cable state information with high precision by constructing a radar array with a preset operating frequency and combining the construction of coded waveforms and observation matrices, improving the accuracy and reliability of detection. By constructing a high-dimensional feature space based on the observation matrix and calculating the embedded data in the high-dimensional phase space, multi-dimensional and in-depth analysis of the cable state is realized, and it can accurately capture the possible minor defects or damages that occur during the operation of the cable. The embedded data is mapped into a surface topological structure, the critical threshold and its eigenvalues are calculated, and by analyzing the change index of the eigenvalues, a defect feature spectrum can be established and a defect probability cloud map can be generated, thus realizing the visual expression of the defect type and distribution. By extracting the spatial and temporal features in the defect probability cloud map as a defect type tensor, performing evolution processing and numerical solution calculations on it, and finally generating a repair strategy matrix, it can adaptively formulate spatial and temporal adjustment plans for repair measures, and can also realize the dynamic optimization of the repair strategy, ensuring the efficiency and intelligence of the repair process.

[0058] In summary, the present invention constructs a set of efficient, accurate, and dynamically updated high-altitude anti-fall cable detection and repair systems, which can effectively prevent and timely respond to cable failures and reduce potential safety hazards. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a schematic diagram of the steps of a detection method for a high-altitude anti-fall cable in an embodiment of the present invention;

[0060] Figure 2 is a schematic block diagram of the structure of a detection device for a high-altitude anti-fall cable in an embodiment of the present invention;

[0061] Figure 3 is a schematic block diagram of the structure of a computer device in an embodiment of the present invention.

[0062] The realization, functional features, and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] In order to make the object, technical solution, and advantages of this application clearer, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application.

[0064] Those skilled in the art can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the above" and "the" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present invention means the presence of features, integers, steps, operations, elements, modules and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, modules, components and / or their groups. It should be understood that when an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any one of the modules and all combinations of one or more associated listed items.

[0065] Those skilled in the art can understand that, unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the field to which the present invention belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined as herein.

[0066] Referring to Figure 1 , an embodiment of the present invention provides a method for detecting an anti-falling cable at high altitude, and the method includes:

[0067] S1: Construct a radar array with a preset operating frequency, obtain the encoded waveform of the cable based on the radar array, and construct an observation matrix according to the encoded waveform;

[0068] In step S1, the radar array is an array system composed of multiple radar transmitting and receiving units, which can simultaneously detect target objects from multiple angles and frequencies. A frequency band that can penetrate or bypass external environmental interference while detecting the cable is set as the operating frequency of the radar array. The radar array obtains the reflected echo by sending and receiving signals. These echo signals contain the reflection characteristics of the cable, such as reflection intensity, reflection time, and waveform changes, which directly reflect the state of the cable and possible defects on its surface and inside. Encode the waveform signals received by the radar array, perform mathematical transformation on the original waveform, and modulate the reflected signals through a specific algorithm so that different defect types can present different coding characteristics in the waveform, and extract the key features in the signals. Based on these encoded waveforms, construct an observation matrix. By organizing the multi-dimensional waveform data obtained from the radar array into a matrix form, it aims to comprehensively describe the overall state of the cable, including the detection signals of each radar unit, covering information such as the correlation and temporal variation between various signals. In actual operation, the construction of the observation matrix can be based on multiple parameters, such as the arrangement of the radar array, the receiving range of each unit, the allocation of different frequency signals, etc., thus constructing a data basis.

[0069] S2: Based on the observation matrix, construct a high-dimensional feature space and calculate the embedded data in the high-dimensional phase space;

[0070] In step S2, the data information contained in the observation matrix is often multi-dimensional. Mapping these multi-dimensional data into a high-dimensional feature space, the high-dimensional feature space is a mathematical construct that maps the original data to a higher dimension through a high-dimensional coordinate system, thereby retaining more information and structural features. By constructing a high-dimensional feature space, complex patterns and laws hidden in the observed data can be captured. Specifically, constructing a high-dimensional feature space is based on phase space reconstruction, which is a mathematical method for recovering the dynamic characteristics of a system from time series data. In this embodiment, the time series data obtained from the radar array is used as input, and the delay embedding technique is utilized to construct the phase space. By constructing the waveform data of the observation matrix into a high-dimensional space at different time delays, the dynamic characteristics of the original signal can be retained, representing the time dependence and possible non-linear behavior of the signal. For example, if cracks or other defects appear on the surface of the cable, the time-varying characteristics of the reflected signal will exhibit different patterns in the embedding space, thereby identifying the specific location and type of the defect. The calculation process of the embedded data can be based on some mathematical tools, such as Takens' theorem, which provides a theoretical basis for reconstructing the phase space from the observed data. By selecting appropriate embedding dimensions and delay times, more cable state information can be captured in the high-dimensional space. The mathematical formula for this process can be expressed as converting a one-dimensional time series into a multi-dimensional phase space vector through delay mapping. The position of each data point in the phase space represents a specific state of the cable. By analyzing the distribution of these positions, the health state of the cable and whether there are potential defects can be inferred.

[0071] S3: Map the embedded data into a surface topological structure, calculate the critical threshold of the surface topological structure, and obtain the eigenvalue of the critical threshold in the surface topological structure;

[0072] In step S3, the embedding data of the high-dimensional phase space is calculated in step S2. The embedding data contains the characteristic information of the cable in different states. Mapping these high-dimensional embedding data into the surface topological structure of a mathematical structure enables the analysis of the variation law of these data by mathematical means. Specifically, mapping the embedding data into the surface topological structure transforms the high-dimensional data into a structure that can be represented in two-dimensional or three-dimensional space. The topological structure is a mathematical concept that describes the shape, connectivity, and structure of objects in space. By using topological methods, the embedding data is constructed into a surface, which can reflect the different states or characteristics of the cable. The critical threshold refers to certain change or transition points in the surface topological structure under specific conditions. These points represent the transition of the cable's health state from normal to abnormal. By calculating the critical threshold, it is possible to determine which regions or parts of the cable have potential defects. When the embedding data is mapped into the surface topological structure, these critical points can be found on the surface, and these points correspond to different degrees of cable damage. For example, information such as the distance and density change between data points can be used to set these thresholds. The critical threshold reflects the boundary where mutations occur under specific circumstances and can identify regions that may have problems. Specifically, when a data point exceeds a certain critical threshold, it means that a certain part of the cable has potential defects, and these defects may affect the normal operation of the cable. Extract the eigenvalue corresponding to this critical point. The eigenvalue reflects the morphological characteristics of the surface topological structure near these critical points. The eigenvalue may represent that a certain part of the cable has been deformed or worn, while other eigenvalues may correspond to excessive strain or vibration at certain points of the cable.

[0073] S4: Obtain the change index of the eigenvalue, establish a defect feature spectrum according to the change index, input the defect feature spectrum into a pulsed neural network for time series encoding, and generate a defect probability cloud map;

[0074] In step S4, by further analyzing the eigenvalues, the changing trends of these eigenvalues can be observed, thereby obtaining a change index. The change index reflects the rate and amplitude of the eigenvalue change over time or spatial position, indicating the possible structural changes that may occur in the cable during use, and a defect feature spectrum is constructed using the index. The defect feature spectrum is a spectrogram reflecting the health status of the cable, and the possible defect types at different times or spatial positions of the cable are represented by the changes in eigenvalues. By constructing the defect feature spectrum, the potential problems of the cable can be visualized, and the defect feature spectrum is input into a spiking neural network for temporal coding. The spiking neural network is a neural network model based on biological principles, which uses discrete spike signals for information processing. The spiking neural network converts the defect feature spectrum into temporal information through temporal coding, thereby capturing the evolution process of the defect in the time dimension and generating a defect probability cloud map. The defect probability cloud map is an image in both spatial and temporal dimensions, representing the probability of a defect existing at each spatial point, and the probability value can be calculated based on the data obtained by the radar array and the output result of the spiking neural network. Each point in the cloud map represents the health status of a certain position of the cable, and the probability value represents the probability of a defect occurring at that position.

[0075] S5: Map the probability value of each spatial point in the defect probability cloud map to a specific position in three-dimensional space, extract the spatial features and temporal features in the defect probability cloud map, and form a defect type tensor.

[0076] In step S5, the probability value of each spatial point in the cloud map is mapped to a specific position in three-dimensional space, and the spatial distribution of each defect is introduced into a three-dimensional coordinate system. In this coordinate system, a multi-dimensional space reflecting the cable state is constructed by combining the spatial position and temporal characteristics, and the spatial features and temporal features are extracted from it. The spatial features reflect the spatial distribution of possible defects in each part of the cable, such as stress concentration areas in certain specific parts of the cable, or structural inhomogeneities that may be caused by construction technology or material problems. The temporal features reflect the dynamic characteristics of these defects changing over time, such as the aging, wear of the cable, or the influence of external environmental changes on the cable. By extracting the features in these two aspects, the damage state of the cable can be comprehensively understood, identifying where the damage occurs, the evolution process of the damage, and the severity of the damage at different time points, and a defect type tensor is formed by combining the spatial features and temporal features. A tensor is a multi-dimensional data structure that can integrate spatial and temporal characteristics and serve as the basis for analysis and processing.

[0077] S6: Perform an evolution process on the defect type tensor and calculate the evolution process based on a numerical solution method to obtain a repair strategy matrix, where the repair strategy matrix includes the adjustment schemes of repair measures at different spatial points and times.

[0078] In step S6, by performing evolution processing on the defect type tensor, the variation law of the defect over time can be captured. Evolution processing refers to analyzing the state changes of a system over a certain period of time through certain mathematical methods or computational models and predicting the future behavior of the system. In the embodiment, the goal of the evolution processing is to identify the evolution trend of the cable defect by calculating the evolution path of the defect type tensor, such as whether certain defects will gradually expand or worsen over time, whether new damage points will appear, etc. Numerical solutions (such as finite element analysis, dynamic models, or other numerical simulation methods) can be used to predict the evolution of the defect. In a specific implementation, the defect type tensor can be regarded as a high-dimensional data set, and numerical solutions can be used to gradually deduce it. The numerical solutions can be based on different computational models, such as the difference method, the integral method, etc. Through these numerical methods, the possible distribution of the defect at future time nodes can be deduced, so as to reasonably predict the future maintenance and repair requirements. For example, if the defect at a certain position has gradually worsened over a past period of time, the numerical solution can help deduce the defect development trend at this position in the next few months and formulate corresponding repair measures, and convert these evolution data into a specific repair strategy matrix. The repair strategy matrix is a multi-dimensional matrix, where each element represents what repair measures should be taken at a specific spatial position and time point. The repair measures can include replacing the cable, strengthening certain stressed parts, adjusting the installation angle of the cable, performing local repair processing, etc. According to the different dimensions of the repair strategy matrix, different adjustment plans can be proposed at different times and in different spaces to ensure the safety and service life of the entire cable. For example, assume that during a certain inspection period, it is detected that the defect probability at the cable joint position gradually increases, and this defect is expandable and may cause more serious damage. Through evolution processing, the defect evolution trend at this joint position in the next few months can be obtained, such as the expansion speed and influence range of the defect. Based on this information, the repair strategy matrix will propose repair measures for this position, such as strengthening the joint position in the short term or replacing it in a longer time.

[0079] In one embodiment, the steps of constructing a radar array with a preset operating frequency, obtaining a coded waveform of the cable based on the radar array, and constructing an observation matrix according to the coded waveform include:

[0080] Modulating and configuring the radar array with a preset operating frequency, generating a radar signal based on a set frequency range, and generating a radar signal with a specific periodicity in frequency through the modulation process;

[0081] Parallelly collecting the reflected signals of the radar signal based on multiple receiving channels and performing time synchronization processing on the multiple reflected signals to obtain the coded waveform;

[0082] Extract the signal intensity information at each moment in the encoded waveform, and convert the reflected signal into a discrete signal sequence according to the signal intensity information;

[0083] Set a time window with a preset number of bits, and control the time window to slide on the discrete signal sequence to generate a plurality of discrete signals;

[0084] Perform time-frequency transformation on the discrete signals to obtain a plurality of frequency characteristics, and construct an observation matrix according to the signal amplitude and phase information of each time-frequency point in the frequency characteristics.

[0085] In the above embodiments, during detection, the operating frequency of the radar array is modulated and configured. Through the modulation process, the frequency of the radar signal exhibits a specific periodicity, which enables the radar signal to effectively interact with the cable when penetrating the cable, thereby capturing the reflection signals on its surface and inside. Specifically, the frequency range of the radar signal is set according to the physical characteristics of the cable to be detected and the required detection accuracy. Multiple receiving channels of the radar array collect the reflection signals of the radar signal in parallel to improve the data acquisition efficiency. The reflection signal reflects the information of the radar signal reflected from the cable surface and its internal structure. The reflection signals are subjected to time synchronization processing to align the signals collected by multiple receiving channels in time, so that they can be analyzed within the same time window, thereby ensuring the consistency and reliability of the data, and obtaining an encoded waveform containing the time information of different reflection signals. The signal intensity information at each moment is extracted. The signal intensity is an important indicator for measuring the signal reflection intensity and reflects the strength of the interaction between the radar signal and the object. By extracting this intensity information, the energy distribution of the reflection signal can be obtained, and the synchronized reflection signal can be converted into a discrete signal sequence, which is obtained by sampling the signal in the time domain and can convert the continuous signal into a form suitable for digital processing. For example, the signal intensity value of the synchronized reflection signal at a certain moment is a real number, such as 1.23. After sampling, 1000 samplings are performed at this moment, and a set of signal intensity values (such as 1.23, 1.25, 1.20, etc.) are obtained. Then these signal intensity values are quantized into discrete levels (such as: 1, 2, 0, etc.) to improve the efficiency of data processing. A time window with a preset number of bits is set, and the time window is controlled to slide on the discrete signal sequence, thereby generating multiple discrete signals, dividing the entire signal sequence into multiple short-time segments, and the signals within each segment can be analyzed independently. These discrete signals are subjected to time-frequency transformation to obtain multiple frequency characteristics, thereby obtaining the characteristic information of the signal at different time points and frequencies. The time-frequency transformation can adopt methods such as wavelet transformation and short-time Fourier transformation to identify the frequency components in the signal and their changing laws over time. According to the signal amplitude and phase information of the time-frequency points in these frequency characteristics, an observation matrix can be constructed. The observation matrix is a multi-dimensional data structure that organizes the time-frequency information in the matrix in a certain way. Through this matrix, the characteristics of the signal can be captured from multiple dimensions, such as the amplitude, phase, and frequency components of the signal.

[0086] In one embodiment, the step of constructing a high-dimensional feature space based on the observation matrix and calculating the embedded data of the high-dimensional phase space includes:

[0087] Performing matrix decomposition on the observation matrix to obtain a set of basis vectors, where each basis vector represents a potential frequency feature;

[0088] Sort the basis vectors in the set of basis vectors, and select the basis vectors whose sorting exceeds a preset value as the main features;

[0089] Perform a linear combination of multiple said main features to form a feature space, and map the data in the feature space through an inner product operation to obtain a preliminary representation of the high-dimensional data space;

[0090] Perform phase space reconstruction on the obtained high-dimensional data, and calculate the embedded data according to the dynamic relationship between the dimension of the reconstructed high-dimensional phase space and the original data.

[0091] In the above embodiment, the observation matrix is factorized, and a large matrix is factorized into the product of multiple small matrices to obtain a set of basis vectors. Each basis vector represents a potential frequency feature in the observation matrix. The basis vector can be understood as an abstract representation of the original data matrix, extracting the main structural features in the matrix. Through these basis vectors, the frequency characteristics and change trends existing in the signal can be understood more clearly. After the matrix factorization, these basis vectors are sorted to screen out the most representative and relevant basis vectors. The sorting can be based on the contribution degree of the basis vectors, or the importance of the frequency features represented by the basis vectors in the entire dataset. For example, if the frequency feature corresponding to a certain basis vector shows a strong signal intensity at multiple times and frequency ranges, then this basis vector will be considered an important feature. After sorting, select the basis vectors whose sorting exceeds a preset value as the main features. This preset value can be a set threshold, indicating that the basis vectors with greater contributions are selected to enter the final feature set, so that the information most relevant to the cable state is retained. Multiple main features are linearly combined to form a new feature space. By the process of combining different basis vectors according to certain weights, these weights are usually determined according to the importance of the basis vectors. The result of the linear combination is a new vector space, which can integrate all the main features, provide a more comprehensive perspective to describe the state of the cable, and map the data through an inner product operation to obtain a preliminary representation of the high-dimensional data space. Mapping the data from one space to another can better capture the potential structures and patterns in the data, and perform phase space reconstruction on the high-dimensional data. Phase space reconstruction is used to analyze the time series data of a dynamic system. Through phase space reconstruction, the dynamic relationship between the high-dimensional data space and the original data can be understood, and thus the embedded data can be calculated. The embedded data refers to the data obtained through phase space reconstruction, which can effectively reflect the dynamic evolution process of the system and represent the dynamic characteristics of the cable state.

[0092] In one embodiment, the steps of mapping the embedded data to a surface topological structure, calculating the critical threshold of the surface topological structure, and obtaining the eigenvalue of the critical threshold in the surface topological structure include:

[0093] Extract the dimensional features of the embedded data, and determine the initial parameters of the surface topological structure based on the dimensional features;

[0094] Based on the topological transformation algorithm, associate each embedded data point in the high-dimensional phase space with a corresponding point on the initial parameters to obtain a set of mapped data points;

[0095] Calculate the local features of the surface topological structure according to the set of mapped data points to obtain the position features of each mapped data point on the surface topological structure;

[0096] Calculate the overall curvature of the surface topological structure according to the position features, and obtain the high-curvature region and the low-curvature region on the surface;

[0097] Identify the distribution information of the high-curvature region and the low-curvature region, obtain the critical points based on the distribution information, and calculate the critical threshold of the surface topological structure based on the critical points;

[0098] Segment the surface topological structure according to the critical threshold to obtain multiple sub-regions;

[0099] Calculate the density features of the set of mapped data points in each sub-region, and obtain the eigenvalue in each sub-region according to the density features.

[0100] In the above embodiments, the embedded data reflects the state changes of the high-altitude fall prevention cable, including signal characteristics such as vibration and bending, and extracts the dimensional characteristics of these data. The extraction of dimensional characteristics can be understood as obtaining the complexity and structure of the embedded data in a multi-dimensional space. For example, methods such as principal component analysis (PCA) can effectively extract the main dimensions of the data and understand its distribution pattern in the high-dimensional space. Based on these dimensional characteristics, the initial parameters of the surface topological structure can be determined, that is, the numerical values closely related to the distribution characteristics and change trends of the data, such as the starting point, preliminary curvature information, etc. According to the topological transformation algorithm, each embedded data point in the high-dimensional phase space is associated with a corresponding point on the initial parameters to form a set of mapped data points. The topological transformation can map the data from the original high-dimensional space to the surface topological structure. In this process, the mapping relationship of each data point depends on its position in the original space and its relative relationship with the initial parameters. Through this mapping, a new set of data points can be obtained, and these data points show the distribution characteristics and change rules of the data in the surface topological structure. The set of mapped data points can be used to calculate the local characteristics of the surface topological structure. The local characteristics refer to the characteristics of each data point at the position after mapping, reflecting the local geometric properties of the point on the surface. For example, the area where the data point is located may have different densities or geometric forms, which can reflect a certain state or change of the cable. Based on these local characteristics, the position characteristics of each mapped data point in the surface topological structure can be obtained, and the overall curvature can be calculated. The overall curvature refers to the degree of bending of the surface at a certain point or a small area. In this embodiment, calculating the overall curvature of the surface can identify the high-curvature regions and low-curvature regions in the surface. The high-curvature regions represent the parts where the surface bends or deforms more significantly, corresponding to the damage or abnormality of the cable; the low-curvature regions represent the parts with less surface change, corresponding to the normal or stable state. Based on the distribution information of the high-curvature regions and low-curvature regions, the critical points are identified. The critical points refer to the boundaries between the high-curvature and low-curvature regions in the surface topological structure, and based on the critical points, the critical threshold of the surface topological structure is calculated. The calculation method can be to perform weighted averaging on the curvature values at the critical points or take statistical quantities such as the maximum value, minimum value, etc. as the critical threshold, and an effective index can be obtained to judge whether the cable is in a safe state or whether maintenance is required. Based on the critical threshold, the surface topological structure is segmented, and the surface is divided into multiple sub-regions. The characteristics within each sub-region are relatively consistent, and each sub-region corresponds to different physical or geometric properties. Through this segmentation, information such as the density characteristics of different regions on the surface can be separated, so as to conduct targeted analysis on each sub-region.Calculate the density feature of each sub-region. The density feature refers to the degree of concentration of data points within each sub-region, reflecting the distribution of data points in the local area. Areas with higher density may indicate more obvious features in that area, while areas with lower density may indicate relatively stable or less variable features in that area. By calculating the density feature of each sub-region, the eigenvalue of each sub-region can be further obtained. The calculation of the eigenvalue is a quantitative index obtained through comprehensive evaluation based on the density feature and other attributes of the sub-region, such as average curvature, geometric shape, etc.

[0101] In another embodiment, the calculation expression of the above embodiment is:

[0102]

[0103] where, represents the final eigenvalue, reflecting the state change of the cable; is the weighting coefficient, representing the weight of the contribution of each data point after mapping to the eigenvalue, depending on the relative importance or sensitivity of the data; n is the size of the mapped data point set, representing the number of data points after mapping; is the density feature vector of the i-th data point, which contains the local density information of the data point within the sub-region (i.e., the degree of distribution within that region), and can be expressed as: , where, is the neighborhood set of the i-th point, is the square of the Euclidean distance between data point i and its neighborhood point j; is the topological transformation feature of the i-th data point, based on the transformation relationship after embedding high-dimensional data into the surface topological structure, and can be expressed as: , where, is the coordinate of data point i in the original high-dimensional space, is the initial parameter (such as the starting point), is the parameter obtained through topological transformation (such as preliminary curvature information); is the curvature feature of the i-th data point, representing the overall curvature change after it is mapped to the surface topological structure, and is calculated based on local features and overall geometric shape, and can be expressed as: , where, is the Laplace operator at the position of data point i, reflecting the local curvature change of the surface; is the density region feature of the i-th data point, and can be expressed as: , where, m is the number of sample points in the region, C i,k represents the curvature value of each point in the region. Through the above comprehensive formula, we can combine the local and global features of high-dimensional data, map them into the surface topological structure, and finally calculate the state eigenvalue of the cable.

[0104] In one embodiment, the steps of obtaining a change index of the eigenvalue, establishing a defect feature spectrum according to the change index, and inputting the defect feature spectrum into a spiking neural network for temporal encoding to generate a defect probability cloud map include:

[0105] Perform time series analysis on the eigenvalue, calculate the change amount between adjacent eigenvalues, and obtain the change index of the eigenvalue;

[0106] Construct a time series model according to the change index, convert the time series model into a frequency domain representation, and obtain a defect feature spectrum;

[0107] Construct a spiking neural network structure, use the defect feature spectrum as the input signal of the spiking neural network, and perform temporal encoding through the process of spike generation and transmission of neurons;

[0108] According to the encoding result, map the defect features at each time point onto a two-dimensional plane to generate a defect probability cloud map, where the density of the defect probability cloud map represents the probability of the defect occurrence.

[0109] In the above embodiments, time series analysis is performed on the eigenvalue to obtain the dynamic change over time, the change amount between adjacent eigenvalues is calculated, and this change amount reflects the difference or variation between every two consecutive data points, thereby obtaining a change index, which reflects the change amplitude and change rate of the eigenvalue. Based on these change indices, a time series model is constructed and transformed into a frequency domain representation. Through mathematical modeling, the regularity in the eigenvalue change is extracted to understand the behavior pattern of the cable in the time series. By performing frequency domain transformation on these time series models, the eigenvalue change can be analyzed from another perspective to obtain a defect feature spectrum, which reflects the change characteristics of the cable in different frequency ranges. The spectrum is input into a Spiking Neural Network (SNN) for time series encoding, simulating the working mode of the biological nervous system, and processing information through the firing and transmission of pulses. In this embodiment, the input signal of the spiking neural network is the defect feature spectrum, representing the change characteristics of the cable at different frequencies. After the neurons in the neural network receive these input signals, they process the information through the firing of pulses, and these pulses are transmitted between neurons. After a series of activation and inhibition processes of neurons, a time series encoding result is finally generated, reflecting the time structure characteristics of the input signal through the firing of pulses of neurons, and being able to retain the time structure characteristics of the signal. Through time series encoding of the signal, the spiking neural network can simulate the dynamic evolution process of the defect in the model. According to the encoding result of the spiking neural network, the defect characteristics at each time point are mapped to a two-dimensional plane to generate a defect probability cloud map. In this process, the characteristics at each time point occupy a position in the two-dimensional plane, and the area with a higher density represents the area with a higher probability of the defect occurring.

[0110] In one embodiment, the step of mapping the probability value of each spatial point in the defect probability cloud map to a specific position in three-dimensional space, and extracting the spatial and time characteristics in the defect probability cloud map to form a defect type tensor includes:

[0111] Perform spatial analysis on the defect probability cloud map, use the probability value of each spatial point in the cloud map as coordinate information, and map it to a three-dimensional space coordinate system to obtain a set of specific position points in three-dimensional space;

[0112] Extract spatial characteristics based on the set of specific position points, and the spatial characteristics include the distribution density, spatial distance, and spatial direction of the spatial points;

[0113] Extract the time characteristics in the defect probability cloud map, and the time characteristics include the change trend and periodicity of the time points;

[0114] Fuse the spatial features and temporal features to construct a defect type tensor, where the dimensions of the tensor include the spatial dimension and the temporal dimension, and the values of each dimension represent the corresponding feature information.

[0115] In the above embodiment, the probability value of each spatial point in the cloud map is mapped to a specific position in the three-dimensional space. Through spatial analysis, according to its position in the cloud map, the probability value of each spatial point is mapped to a three-dimensional coordinate system to form a point set in the three-dimensional space. Based on these point sets mapped to specific positions in the three-dimensional space, spatial features are extracted, mainly including information such as the distribution density, spatial distance, and spatial direction of spatial points. The spatial distance refers to the distance between different defect points, which can help analyze the extent of defect expansion in space. If the distance between points is small, it may indicate that these defects are correlated and form a certain pattern or trend. The spatial direction involves the directionality of the distribution of defect points. If the number of defect points in a certain direction is significantly more than that in other directions, it may reflect specific damage or stress conditions in a certain direction, and temporal features are extracted. The temporal features are mainly reflected in the change trend and periodicity of time points, and the change law of defects over time can be obtained, so as to predict possible faults or risks of the cable in a future period. Fuse the spatial features and temporal features to construct a defect type tensor, whose dimensions include the spatial dimension and the temporal dimension. The spatial dimension corresponds to the spatial distribution information of the defect, and the temporal dimension represents the change of the defect at different time points. In this tensor, the value of each dimension represents the corresponding feature information. For example, a value on the spatial dimension represents the distribution density of a certain position point, and a value on the temporal dimension represents the change trend of the defect at this time point. By combining these two types of features to form a defect type tensor, the spatial and temporal characteristics of cable defects can be comprehensively described, and then through further analysis of the tensor, the defect type, development trend, and occurrence probability of the cable can be accurately predicted.

[0116] In one embodiment, the step of evolving the defect type tensor and calculating the evolution process based on a numerical solution to obtain a repair strategy matrix, where the repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times, includes:

[0117] Obtain an initial value based on the cable material properties, and set the repair parameters of the defect type tensor based on the initial value;

[0118] Construct an evolution model according to the defect type tensor, input the repair parameters into the evolution model, calculate the distribution state of the repair parameters at each time step, and obtain the evolution intermediate results frame by frame;

[0119] Perform optimization processing on the evolution intermediate results based on the particle swarm optimization algorithm to obtain the optimal combination of repair measure parameters at each spatial point;

[0120] Construct an initial repair matrix based on the combination of the repair measure parameters, iteratively update the initial repair matrix, and adjust the repair measure parameters according to the repair effect at each time step to obtain a repair strategy matrix.

[0121] In the above embodiment, initial values are obtained based on the characteristics of the cable material. Each type of cable material has different physical and chemical characteristics, which determine the damaged condition of the cable and its repair method. For example, some materials will experience microcrack propagation when subjected to pressure or tension, and some will exhibit brittle fracture due to long-term exposure to high or low temperatures. These factors will affect the formulation of the repair strategy. Therefore, the initial repair parameters are determined from the material characteristics of the cable. The parameters can include a preliminary assessment of aspects such as material strength, fatigue life, and damage tolerance. Based on these initial repair parameters, an evolution model is constructed. The goal of this evolution model is to simulate the development process of cable defects over time and how to take different repair measures at different times to prevent the defects from deteriorating, covering the dynamic characteristics of defect propagation. Considering the passage of time, the defects may propagate at different speeds at different spatial points. By inputting the repair parameters into the evolution model, the distribution state of these repair parameters at each time step can be calculated, and then the frame-by-frame intermediate results of the evolution can be obtained, reflecting the change trend of cable defects and the corresponding repair requirements at different time points. The intermediate results obtained as the evolution process progresses need to be further optimized to find the optimal repair measures. The particle swarm optimization algorithm is an optimization technique that simulates swarm behavior and can quickly find the optimal solution in a large solution space. In this embodiment, the particle swarm represents different combinations of repair measures, and each particle represents a possible repair plan. The particles move in the search space and continuously adjust their positions to find the repair measures that optimize the results of the evolution model. At each time step, the particle swarm optimization algorithm dynamically adjusts the parameter combination of the repair measures by evaluating the error between the intermediate results of the evolution and the repair target, and finally determines the most suitable repair plan at each spatial point and constructs an initial repair matrix based on these parameters. Record and adjust the specific content of the repair measures at each time step during the entire repair process. The construction of the matrix is an iterative process. When constructing the initial repair matrix, the content is based on the preliminary plan of the optimal repair measures. As time goes by and the feedback of the repair effect, the parameters in the repair matrix will be continuously updated and adjusted. Each update will recalculate the effectiveness of the repair measures according to the feedback of the actual repair effect during the evolution process and adjust the parameters accordingly to ensure that the most suitable repair strategy can be provided according to the defect state of the cable at each time step, so that the repair strategy can be gradually optimized to cope with the dynamically changing defect situation, and finally, through iterative update, a repair strategy more suitable for the defects in this area is obtained.

[0122] Refer toFigure 2 , a detection device for high-altitude anti-falling cables, comprising:

[0123] An acquisition module 100, configured to construct a radar array with a preset operating frequency, obtain a coded waveform of the cable based on the radar array, and construct an observation matrix according to the coded waveform;

[0124] A calculation module 200, configured to construct a high-dimensional feature space based on the observation matrix, and calculate embedded data in the high-dimensional phase space;

[0125] A mapping module 300, configured to map the embedded data into a surface topological structure, calculate a critical threshold of the surface topological structure, and obtain an eigenvalue of the critical threshold in the surface topological structure;

[0126] A generation module 400, configured to obtain a change index of the eigenvalue, establish a defect feature spectrum according to the change index, input the defect feature spectrum into a pulse neural network for time series coding, and generate a defect probability cloud map;

[0127] An extraction module 500, configured to map the probability value of each spatial point in the defect probability cloud map to a specific position in a three-dimensional space, extract spatial features and time features in the defect probability cloud map, and form a defect type tensor;

[0128] An evolution module 600, configured to perform evolution processing on the defect type tensor, and calculate an evolution process based on a numerical solution method to obtain a repair strategy matrix, where the repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times.

[0129] In this embodiment, a coded waveform of the cable is obtained by using a radar array at a preset operating frequency, and an observation matrix is constructed based on these waveforms, so as to provide preliminary data on the cable state. Then, by analyzing the observation matrix, a high-dimensional feature space is constructed, and embedded data is calculated within this space. These embedded data are mapped into a surface topological structure, and a critical threshold and an eigenvalue are calculated to identify potential defect areas. Further, according to the change index of the eigenvalue, a defect feature spectrum is established and input into a pulse neural network for time series coding, and finally a defect probability cloud map is generated, showing the defect probabilities at different spatial points. By mapping these probability values into a three-dimensional space and extracting spatial and time features, a defect type tensor is formed to describe the spatial distribution and time variation of the defects. Finally, a numerical solution method is used to perform evolution processing on the defect type tensor, deduce the evolution process of the defects, and generate a repair strategy matrix accordingly, providing corresponding repair schemes for different times and spatial points.

[0130] Refer to Figure 3 , in the embodiment of the present application, a computer device is further provided. The computer device may be a server, and its internal structure may be asFigure 3 As shown in Figure 3 . The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Among them, the processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data and other information related to this application. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a detection method for high-altitude anti-falling cables.

[0131] An embodiment of this application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements a detection method for high-altitude anti-falling cables, including the steps of: constructing a radar array with a preset operating frequency, obtaining the encoded waveform of the cable based on the radar array, and constructing an observation matrix according to the encoded waveform; constructing a high-dimensional feature space based on the observation matrix, and calculating the embedded data in the high-dimensional phase space; mapping the embedded data into a surface topological structure, calculating the critical threshold of the surface topological structure, and obtaining the eigenvalue of the critical threshold in the surface topological structure; obtaining the change index of the eigenvalue, establishing a defect feature spectrum according to the change index, inputting the defect feature spectrum into a pulse neural network for time series coding, and generating a defect probability cloud map; mapping the probability value of each spatial point in the defect probability cloud map to a specific position in three-dimensional space, extracting the spatial and temporal features in the defect probability cloud map to form a defect type tensor; performing an evolution process on the defect type tensor, and calculating the evolution process based on a numerical solution method to obtain a repair strategy matrix, where the repair strategy matrix includes the adjustment plan of repair measures at different spatial points and times.

[0132] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided in this application and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0133] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A method for detecting a high altitude anti-fall cable, characterized in that: The method comprises: Constructing a radar array with a preset working frequency, acquiring a coding waveform of the cable based on the radar array, and constructing an observation matrix according to the coding waveform; Constructing a high-dimensional feature space based on the measurement matrix, and calculating embedded data of the high-dimensional phase space; Mapping the embedded data into a surface topological structure, calculating a critical threshold of the surface topological structure, and obtaining a characteristic value of the critical threshold in the surface topological structure; Obtaining a variation index of the characteristic value, establishing a defect characteristic spectrum according to the variation index, inputting the defect characteristic spectrum into a pulse neural network for time series encoding, and generating a defect probability cloud map; Mapping the probability value of each spatial point in the defect probability cloud map to a specific position in the three-dimensional space, extracting the spatial features and time features in the defect probability cloud map, and forming a defect type tensor; Performing evolution processing on the defect type tensor, and calculating the evolution process based on a numerical solution to obtain a repair strategy matrix, wherein the repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times; The step of mapping the embedded data into a surface topological structure, calculating a critical threshold of the surface topological structure, and obtaining a characteristic value of the critical threshold in the surface topological structure comprises: Extracting dimensional features of the embedded data, and determining initial parameters of the surface topology structure based on the dimensional features; Based on a topological transformation algorithm, each embedded data point in the high-dimensional phase space is associated with a corresponding point on the initial parameter to obtain a mapping data point set; Calculate the local features of the surface topology structure according to the mapping data point set to obtain the position features of each mapping data point on the surface topology structure; Calculate the overall curvature of the surface topological structure according to the position features, and obtain high curvature areas and low curvature areas on the surface; Identifying distribution information of the high curvature region and the low curvature region, obtaining a critical point based on the distribution information, and calculating a critical threshold of a surface topological structure based on the critical point; Segmenting the surface topology structure according to the critical threshold to obtain a plurality of sub-regions; Calculating the density feature of each sub-region mapping data point set, and obtaining the feature value in each sub-region according to the density feature; The step of obtaining a variation index of the characteristic value, establishing a defect characteristic spectrum according to the variation index, inputting the defect characteristic spectrum into a pulse neural network for time series encoding, and generating a defect probability cloud map comprises: Performing time series analysis on the eigenvalues, calculating the variation between adjacent eigenvalues, and obtaining a variation index of the eigenvalues; Constructing a time series model according to the variation index, converting the time series model into a frequency domain representation, and obtaining a defect characteristic spectrum; Constructing a spiking neural network structure, taking the defect characteristic spectrum as an input signal of the spiking neural network, and performing temporal encoding through the pulse emission and transmission process of neurons; According to the encoding result, the defect characteristics at each time point are mapped onto a two-dimensional plane to generate a defect probability cloud map, wherein the density of the defect probability cloud map represents the probability of the defect occurring.

2. The detection method of the high altitude anti-fall cable according to claim 1 is characterized in that: The step of constructing a radar array with a preset working frequency, acquiring a coded waveform of the cable based on the radar array, and constructing an observation matrix according to the coded waveform includes: Modulating and configuring a radar array of a preset operating frequency, generating a radar signal based on a set frequency range, and generating a radar signal with a specific periodicity in frequency through a modulation process; Based on multiple receiving channels, the reflected signals of the radar signal are collected in parallel, and the multiple reflected signals are subjected to time synchronization processing to obtain the coded waveform; Extracting signal strength information at each moment in the coded waveform, and converting the reflected signal into a discrete signal sequence according to the signal strength information; Setting a time window of preset bits, controlling the time window to slide on the discrete signal sequence, and generating a plurality of discrete signals; The discrete signal is subjected to time-frequency transformation to obtain a plurality of frequency features, and an observation matrix is ​​constructed according to the signal amplitude and phase information of each time-frequency point in the frequency features.

3. The detection method of the high altitude anti-fall cable according to claim 1 is characterized in that: The step of constructing a high-dimensional feature space based on the measurement matrix and calculating the embedded data of the high-dimensional phase space comprises: Performing matrix decomposition on the observation matrix to obtain a set of basis vectors, wherein each basis vector represents a potential frequency feature; Sorting basis vectors in the basis vector set, and selecting basis vectors whose order exceeds a preset value as main features; Linearly combining a plurality of the main features to form a feature space, mapping the data in the feature space through an inner product operation to obtain a preliminary representation of a high-dimensional data space; The obtained high-dimensional data is reconstructed in phase space, and the embedded data is calculated based on the dynamic relationship between the dimension of the reconstructed high-dimensional phase space and the original data.

4. The detection method of the high altitude anti-fall cable according to claim 1 is characterized in that: The step of mapping the probability value of each spatial point in the defect probability cloud map to a specific position in the three-dimensional space, extracting the spatial features and time features in the defect probability cloud map, and forming a defect type tensor comprises: Performing spatial analysis on the defect probability cloud map, taking the probability value of each spatial point in the cloud map as coordinate information, mapping it to a three-dimensional space coordinate system, and obtaining a specific position point set in the three-dimensional space; Extracting spatial features based on the specific location point set, the spatial features including distribution density, spatial distance and spatial direction of spatial points; Extracting time features from the defect probability cloud map, wherein the time features include a change trend and periodicity of a time point; The spatial features and the temporal features are fused to construct a defect type tensor, wherein the dimensions of the tensor include a spatial dimension and a temporal dimension, and the value of each dimension represents corresponding feature information.

5. The detection method of the high altitude anti-fall cable according to claim 1 is characterized in that: The step of performing evolution processing on the defect type tensor and calculating the evolution process based on a numerical solution to obtain a repair strategy matrix, wherein the repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times, includes: Acquire an initial value based on cable material characteristics, and set a repair parameter of the defect type tensor based on the initial value; An evolution model is constructed according to the defect type tensor, the repair parameters are input into the evolution model, the distribution state of the repair parameters at each time step is calculated, and the intermediate evolution results are obtained frame by frame; Optimizing the intermediate results of the evolution based on a particle swarm optimization algorithm to obtain an optimal combination of repair measure parameters at each spatial point; An initial repair matrix is ​​constructed based on the combination of repair measure parameters, the initial repair matrix is ​​iteratively updated, and the repair measure parameters are adjusted according to the repair effect at each time step to obtain a repair strategy matrix.

6. A detection device for a high altitude anti-fall cable, applied to the method according to any one of claims 1 to 5, characterized in that: include: An acquisition module, used to construct a radar array of a preset working frequency, acquire a coding waveform of the cable based on the radar array, and construct an observation matrix according to the coding waveform; A calculation module, used for constructing a high-dimensional feature space based on the observation matrix, and calculating the embedded data of the high-dimensional phase space; A mapping module, used for mapping the embedded data into a surface topological structure, calculating a critical threshold of the surface topological structure, and obtaining a characteristic value of the critical threshold in the surface topological structure; A generation module, used for obtaining a variation index of the characteristic value, establishing a defect characteristic spectrum according to the variation index, inputting the defect characteristic spectrum into a pulse neural network for time series encoding, and generating a defect probability cloud map; An extraction module is used to map the probability value of each spatial point in the defect probability cloud map to a specific position in the three-dimensional space, extract the spatial features and time features in the defect probability cloud map, and form a defect type tensor; The evolution module is used to perform evolution processing on the defect type tensor and obtain a repair strategy matrix based on the numerical solution to calculate the evolution process. The repair strategy matrix includes adjustment schemes of repair measures at different spatial points and times.

7. A computer device, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images