Self-adaptive slope disaster early warning method based on multi-source monitoring data and related device

Abstract

The embodiment of the invention relates to the field of data processing and slope monitoring, and provides a self-adaptive slope disaster early warning method based on multi-source monitoring data and a related device, and the method comprises the steps: obtaining the original multi-source monitoring data of a to-be-detected slope; preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data; performing feature extraction on the target multi-source monitoring data to obtain target feature data; inputting the target characteristic data into a self-adaptive slope disaster early warning model, and using an output calculation result to calculate the stability of the slope to be detected to obtain a stability evaluation result; and carrying out early warning processing and catastrophe prediction on the slope to be detected by adopting the stability evaluation result. The slope disaster early warning model is adaptive to dynamic evolution of original multi-source monitoring data of a to-be-detected slope by automatically adjusting model parameters and structures, analysis and calculation of slope stability evaluation results are realized, catastrophe prediction and early warning processing is performed accordingly, and timeliness and accuracy of prediction and early warning are effectively improved.

Description

Technical Field

[0001] The present application relates to the technical fields of data processing and slope monitoring, and in particular to an adaptive slope disaster early warning method and related devices based on multi-source monitoring data. Background Art

[0002] Traditional slope abnormality warning methods are mostly based on static or preset models, which are difficult to adapt to the dynamic changes of slope status with environmental, time and other factors. Due to the complexity and uncertainty of slope disasters, traditional methods are difficult to achieve high-precision prediction effects. Different monitoring data sources vary greatly, and it is extremely challenging to comprehensively reflect the slope status through effective data fusion. Summary of the invention

[0003] The embodiment of the present application provides an adaptive slope disaster warning method and related devices based on multi-source monitoring data. The slope disaster warning model automatically adjusts the model parameters and structure, adapts to the dynamic evolution of the original multi-source monitoring data of the slope to be detected, realizes the analysis and calculation of the slope stability assessment results, and performs disaster prediction and warning processing based on this, effectively improving the timeliness and accuracy of the prediction and warning.

[0004] A first aspect of an embodiment of the present application provides an adaptive slope disaster early warning method based on multi-source monitoring data, the method comprising:

[0005] Obtaining original multi-source monitoring data of the slope to be detected;

[0006] Preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data;

[0007] Extracting features from the target multi-source monitoring data to obtain target feature data;

[0008] Adopting an adaptive slope disaster early warning model to output the model calculation result through the target characteristic data;

[0009] Using the model output results to perform stability assessment on the slope to be tested, to obtain a stability assessment result;

[0010] The stability assessment result is used to perform early warning processing on the slope to be detected.

[0011] In this example, the original multi-source monitoring data of the slope to be detected is obtained, the original multi-source monitoring data is preprocessed to obtain target multi-source monitoring data, features are extracted from the target multi-source monitoring data to obtain target feature data, an adaptive slope disaster warning model is used to output the model operation results through the target feature data, the model output results are used to perform stability assessment on the slope to be detected to obtain a stability assessment result, and the stability assessment result is used to perform warning processing on the slope to be detected. Therefore, the slope disaster warning model automatically adjusts the model parameters and structure, adapts to the dynamic evolution of the original multi-source monitoring data of the slope to be detected, realizes the analysis and calculation of the slope stability assessment results, and performs disaster prediction and warning processing based on this, thereby effectively improving the timeliness and accuracy of the prediction and warning.

[0012] A second aspect of an embodiment of the present application provides an adaptive slope disaster early warning device based on multi-source monitoring data, the device comprising:

[0013] An acquisition unit, used for acquiring original multi-source monitoring data of the slope to be detected;

[0014] A preprocessing unit, used for preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data;

[0015] A feature extraction unit, used to extract features from the target multi-source monitoring data to obtain target feature data;

[0016] An evaluation unit, the evaluation unit is used to use the adaptive slope disaster early warning model to output the model result through the target feature data; use the model output result to perform stability evaluation on the slope to be detected to obtain a stability evaluation result;

[0017] An early warning unit is used to use the stability assessment result to perform early warning processing on the slope to be detected.

[0018] A third aspect of an embodiment of the present application provides a terminal, comprising a processor, an input device, an output device and a memory, wherein the processor, input device, output device and memory are interconnected, wherein the memory is used to store a computer program, the computer program comprises program instructions, and the processor is configured to call the program instructions to execute the step instructions in the first aspect of the embodiment of the present application.

[0019] The fourth aspect of the embodiments of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute part or all of the steps described in the first aspect of the embodiments of the present application.

[0020] A fifth aspect of the embodiments of the present application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps described in the first aspect of the embodiments of the present application. The computer program product may be a software installation package. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 A schematic flow chart of an adaptive slope disaster early warning method based on multi-source monitoring data is provided for an embodiment of the present application;

[0023] Figure 2 A schematic diagram of the structure of a terminal provided in an embodiment of the present application;

[0024] Figure 3 A structural schematic diagram of an adaptive slope disaster warning device based on multi-source monitoring data is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0026] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0027] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments.

[0028] In order to better understand the adaptive slope disaster early warning method based on multi-source monitoring data provided by the embodiment of the present application, the slope disaster early warning method in the existing scheme is briefly introduced below. In the existing scheme, a static or preset model is usually used to perform slope early warning based on the collected monitoring data. It is difficult to adapt to the dynamic changes of the slope state with factors such as the environment and time, and due to the complexity and uncertainty of slope disasters, the accuracy of the slope early warning using the existing static or preset model is insufficient, resulting in poor accuracy.

[0029] In order to solve the above technical problems, the embodiment of the present application provides an adaptive slope disaster warning method based on multi-source monitoring data. The slope disaster warning model automatically adjusts the model parameters and structure to adapt to the dynamic evolution of the original multi-source monitoring data of the slope to be detected, realizes the analysis and calculation of the slope stability assessment results, and performs disaster prediction and warning processing based on this, which effectively improves the timeliness and accuracy of the prediction and warning. .

[0030] See also Figure 1 , Figure 1 The present application provides a flowchart of an adaptive slope disaster early warning method based on multi-source monitoring data. Figure 1 As shown, the method includes:

[0031] 101. Obtain original multi-source monitoring data of the slope to be tested.

[0032] Among them, the original multi-source monitoring data include slope displacement, slope stress, groundwater level, rainfall, soil moisture and temperature, etc.

[0033] 102. Preprocess the original multi-source monitoring data to obtain target multi-source monitoring data.

[0034] Since the data collection process may be affected by environmental noise, equipment errors and other factors, preprocessing is required to ensure the quality and reliability of the data used in subsequent analysis. The preprocessing process includes data cleaning (removing noise and outliers), data denoising (further reducing noise through Fourier transform and other techniques), data standardization (ensuring that data of different dimensions can be compared on the same scale) and data smoothing (reducing data fluctuations and highlighting trend changes).

[0035] Specifically, a method for preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data includes:

[0036] A1. Using a median filtering method to filter the original multi-source monitoring data to obtain first multi-source monitoring data;

[0037] A2. performing outlier removal processing on the first multi-source monitoring data to obtain second multi-source monitoring data;

[0038] A3, performing denoising processing on the second multi-source monitoring data to obtain third multi-source monitoring data;

[0039] A4. Standardize and smooth the third multi-source monitoring data to obtain the target multi-source monitoring data.

[0040] The original multi-source monitoring data may be filtered using the method shown in the following formula to obtain the first multi-source monitoring data:

[0041] y N [i] = median(y O [ik:i+k+1])

[0042] Where: y N [i] represents the i-th data point in the original multi-source monitoring data; k represents the size of the filter window, which is a positive integer and determines the smoothness of the filter; y o [i] represents the data point after median filtering, which can remove random noise, that is, the first multi-source detection data; median is a mathematical function, which means finding the median of a group of numbers. The above formula represents the original data sequence y N The median of the data in a window with a length of 2k+1 and the ith data point as the center is calculated to obtain the corresponding first multi-source monitoring data.

[0043] Based on the IQR (interquartile range) method, the first multi-source monitoring data is screened for outliers. After the outliers are screened out, the outliers are deleted to obtain the second multi-source monitoring data. Specifically, the abnormal threshold range is set to [Q 1 -1.5×IQR,Q 3 +1.5×IQR], if it exceeds this range, it is considered as data abnormality, where:

[0044] IQR=Q 3 -Q 1 ,

[0045] Where: Q 1 , Q 3They represent the first quartile (25% quantile) and the third quartile (75% quantile) of the first multi-source monitoring data respectively; IQR represents the interquartile range, which is used to measure the degree of dispersion of the data; Q 1 -1.5×IQR represents the lower bound of outliers, Q 3 +1.5×IQR represents the upper bound of outliers.

[0046] When denoising the second multi-source monitoring data, Fourier transform can be used for denoising, specifically:

[0047] Y = FFT(y N )

[0048] Keep the low frequency part, that is:

[0049] Y f =Y×Y fm

[0050] Then perform the inverse transformation, that is:

[0051] y d =IFFT(Y f )

[0052] Where, Y represents the Fourier transform result of the original data monitoring data; y N represents the original monitoring data sequence; Y f Represents the filtered low-frequency data; Y fm Where fm is the filter mask, Y fm Indicates the frequency of the data to be retained; d Represents the third multi-source monitoring data obtained after denoising.

[0053] The method for standardizing and smoothing the third multi-source monitoring data is as follows:

[0054] The Z-score standardization method is used for data standardization:

[0055] y n [i]=(y d [i]-mean(y d [i])) / std(y d [i])

[0056] Where: y d [i] represents the i-th element in the third multi-source monitoring data; mean is the mean calculation function, which means calculating the mean of all elements in the data set; std represents the standard deviation calculation function, which means calculating the standard deviation of all elements in the data set; y n Represents the normalized data.

[0057] The sliding average is used to smooth the data to reduce random fluctuations in the data and reveal the long-term trend and cyclical changes of the data, namely:

[0058]

[0059] In the formula, y s [i] represents the data value at position i after sliding average smoothing; I represents half of the window size (because the window is centrally symmetrical, the actual range considered is from il to i+l, a total of 2×I+1 points); y n Represents the standardized data, that is, the target multi-source monitoring data.

[0060] In this example, the original multi-source monitoring data is preprocessed to ensure data quality and reduce the impact of noise and outliers on subsequent fitting and prediction. The preprocessed data can better reflect the real changes in the slope status.

[0061] 103. Perform feature extraction on the target multi-source monitoring data to obtain target feature data.

[0062] Among them, the target characteristic data includes data change trend information, data cycle information and emergency event information.

[0063] A method for extracting features from the target multi-source monitoring data to obtain target feature data includes:

[0064] B1. Extracting data change trends of the target multi-source monitoring data to obtain data change trend information;

[0065] B2. Using discrete Fourier transform to extract data period information from the target multi-source monitoring data;

[0066] B3. Use preset sensitivity information, data reference values ​​and target multi-source monitoring data to identify emergencies and obtain the emergencies information.

[0067] Specifically, the method for extracting the data change trend of the target multi-source monitoring data and obtaining the data change trend information is as follows:

[0068] The data change trend information can be fitted with a trend line using a linear regression model. By calculating the slope of the trend line and taking its absolute value, the strength of the data change trend can be quantified. The larger the absolute value of the slope, the stronger the trend.

[0069]

[0070] In the formula, t i represents the time variable, y i Indicates the corresponding target multi-source monitoring data value; and Represent the mean of time and data respectively, n is the number of data points; T represents the slope of the trend line.

[0071] The method of extracting data periodic information from the target multi-source monitoring data by using discrete Fourier transform is as follows: using discrete Fourier transform to extract the periodic components of the dynamic changes of the data, converting the signal from the time domain to the frequency domain, decomposing the time series data into the sum of sine waves and cosine waves of different frequencies, and identifying the main periodic components:

[0072]

[0073] Where Y(f) is the Fourier coefficient at frequency f; y(t) is the time series data, where t is the time index, ranging from 0 to n-1, and n is the total number of data points; f is the analysis frequency, and its range depends on the sampling rate of the data and the selected frequency range; e -2πift is a complex exponential function used to map time domain signals to frequency domain, where i is the imaginary unit and π is the circumference of a circle.

[0074] The main periodic components in the time series data can be identified by observing the amplitude spectrum of Y(f) (i.e., the change of |Y(f)| with f). The peak in the amplitude spectrum usually corresponds to the periodic component in the data. The position of the peak indicates the frequency of the periodic component, and the peak size reflects the intensity of the corresponding periodic component.

[0075] The method of using preset sensitivity information, data reference value and target multi-source monitoring data to identify emergencies and obtain the emergencies information is as follows:

[0076] Identification of sudden events is a crucial link in slope monitoring and disaster warning systems. Sudden events usually refer to events that occur suddenly in a short period of time and have a significant impact on slope stability, such as heavy rainfall, earthquakes, landslides, etc. They are sudden, abnormal and potentially dangerous. It is necessary to detect events that may indicate significant changes in slope stability from monitoring data in a timely and accurate manner, which is crucial for subsequent early warnings and response measures. Potential events are detected by analyzing abnormal changes in monitoring data. Abnormal changes may manifest as a sharp rise or fall in data values, a significant increase in data volatility, etc. CUSUM control charts are used to identify sudden events, and signals are amplified by accumulating deviations so that small deviations can also be detected after accumulating to a certain extent. The following formula is used to calculate slope deviation information:

[0077] S t =max(0, S t-1 +(y t -k)-h),

[0078] In the formula, St Represents the slope offset information at time t; y t It represents the target multi-source monitoring data value at time t; k represents the data reference value, which is usually taken as the mean or median of the monitoring data of the process in a stable state; h is the control limit or decision interval (preset sensitivity information), which is used to control the sensitivity of the cumulative sum; the max function is to ensure that the cumulative sum is not negative.

[0079] The specific principle is: Initialize, accumulate and S 0 Initialized to 0; calculate the cumulative sum, for each new monitoring data point y t , calculate its deviation from the target value k and subtract the control limit h (if the deviation is positive) or add the control limit h (but usually only positive deviations are of interest, as negative deviations may indicate that the process is returning to a more stable state). Then, add this result to the previous cumulative sum S t-1 On, we get the new cumulative sum S t ; Detection event, if the accumulation and S t If the preset threshold value (usually equal to or slightly greater than a certain multiple of the control limit h) is exceeded, it is considered that a sudden event has occurred; reset the cumulative sum. After the event is detected, the cumulative sum S t Reset to 0 or some smaller value to restart accumulating new deviations.

[0080] Dynamic change feature extraction is the bridge between data preprocessing and adaptive algorithm construction. By extracting key features, the original data is converted into a form that is beneficial to subsequent analysis. The key features reflect the current status of the slope state and reveal the dynamic change rules and potential trends, providing important input information for the subsequent adaptive algorithm construction.

[0081] 104. Use an adaptive slope disaster early warning model to evaluate the stability of the slope to be detected through the target feature data to obtain a model output result.

[0082] 105. Perform stability assessment on the slope to be tested using the model output result to obtain a stability assessment result.

[0083] The model output result is used to perform stability assessment on the slope to be detected to obtain a stability assessment result; and the stability assessment result is used to perform early warning processing on the slope to be detected.

[0084] Before using the adaptive slope disaster early warning model for stability assessment, it is necessary to build the model. The specific construction method is as follows:

[0085] By building a deep learning model that combines LSTM with an attention mechanism, it can automatically learn the complex relationship between dynamic features and predict the future state or stability level of the slope based on this relationship. The adaptive early warning algorithm can automatically adjust its internal parameters according to changes in input data to better fit the dynamic change pattern in the data.

[0086] 1) Dynamic model construction: Based on the long short-term memory network (LSTM) combined with the attention mechanism or the hybrid model of the convolutional neural network (CNN) and LSTM (ConvLSTM), the output of the LSTM layer is defined by weighting the attention layer to highlight important information.

[0087] T o =LSTM(x t-1 ,h t-1 ),

[0088] In the formula, x t-1 h represents the input sequence value at time t-1, usually a vector containing multiple features (such as trend strength, periodic components, and sudden events in step 2); t-1 represents the hidden state of LSTM at time t-1, which is also a vector containing the sequence information up to the current time; LSTM is the function of the LSTM layer, which calculates the hidden state at the current time based on the current input and the hidden state at the previous time; T o Represents the output data of the LSTM layer (that is, the hidden layer output data).

[0089] T AW =softmax(W a tanh(U a ·T o +b a )),

[0090] Where W a and U a are the attention weight matrices of the attention layer, which are used to learn to convert the output data of the LSTM layer into attention scores; b a represents the bias vector of the attention layer; tanh is the hyperbolic tangent activation function, which is used to introduce nonlinearity; softmax is the softmax function, which is used to normalize the attention score to a probability distribution, i.e., the attention weight; T AW is the attention weight value.

[0091] T WO =T AW ⊙T o ,

[0092] In the formula, ⊙ represents element-by-element multiplication, which is used to perform weighted operations on the output data of the LSTM layer according to the attention weight value, TWO Output results for the model.

[0093] The output of the adaptive model algorithm is used to calculate the real-time assessment value of slope stability (slope stability can be calculated based on the limit equilibrium method), and compared with the set warning threshold to trigger the corresponding warning mechanism. The algorithm model can automatically adjust its parameters and structure to adapt to the dynamic changes of monitoring data, and realize accurate prediction of slope stability and advance warning of disasters.

[0094] After the model is built, model validation is required:

[0095] ① Based on the confusion matrix, the specificity of different parameters in the above model is calculated, and the model is verified through the precision-recall trade-off, including:

[0096]

[0097] In the formula, TPR is the true positive rate or recall rate, which indicates the proportion of samples that are correctly classified as positive samples among all truly positive samples; TP is the true positive example, which indicates the number of positive samples correctly predicted by the model as positive; FN is the false negative example, which indicates the number of positive samples incorrectly predicted by the model as negative; FPR is the false positive rate, which indicates the proportion of samples that are incorrectly classified as positive among all truly negative samples; FP is the false positive example, which indicates the number of negative samples incorrectly predicted by the model as positive; TN is the true negative example, which indicates the number of negative samples correctly predicted by the model as negative; S indicates the proportion of samples that are correctly classified as negative among all truly negative samples.

[0098] ②Use F β Comprehensive evaluation model performance:

[0099]

[0100] In the formula, P is the precision, It indicates the proportion of samples that are actually positive among the samples predicted by the model to be positive; β is a user-defined parameter used to control the trade-off between precision and recall.

[0101] Specifically, a method for outputting a model calculation result of the slope to be detected by using the adaptive slope disaster early warning model through the target feature data includes:

[0102] C1. constructing a vector of the target feature data to obtain a target feature data vector;

[0103] C2, obtaining hidden layer output data by performing calculations on the hidden state vector of the hidden layer of the adaptive slope disaster early warning model and the target feature data vector;

[0104] C3, using the attention weight matrix of the attention layer of the adaptive slope disaster early warning model to calculate the attention weight of the hidden layer output data to obtain the attention weight value;

[0105] C4. Using the adaptive slope disaster warning model, the attention weight value is used to perform weighted operation on the hidden layer output data to obtain a model operation result.

[0106] The model output result is used to perform stability assessment on the slope to be detected to obtain a stability assessment result.

[0107] The feature vector includes trend strength, periodic components, emergencies, etc. The above steps C2-C4 can be processed by referring to the calculation method of the network layer corresponding to the adaptive slope disaster early warning model in the above embodiment to obtain the corresponding results. The hidden layer of the adaptive slope disaster early warning model is the LSTM layer.

[0108] 106. Use the stability assessment result to perform early warning processing on the slope to be detected.

[0109] Specifically, a method for performing early warning processing on the slope to be detected using the stability assessment result includes:

[0110] D1. extracting a stability evaluation value from the stability evaluation result;

[0111] D2. If the stability evaluation value is greater than or equal to a preset stability warning threshold, a warning process is performed;

[0112] D3. If the stability evaluation value is less than the preset stability warning threshold, no warning processing is performed.

[0113] The stability assessment results are used to calculate the slope stability assessment value (slope stability can be calculated based on the limit equilibrium method) and compared with the set stability warning threshold to trigger the corresponding warning mechanism. The algorithm model can automatically adjust its parameters and structure to adapt to the dynamic changes of monitoring data, and realize accurate prediction of slope stability and advance warning of disasters.

[0114] In this example, the original multi-source monitoring data of the slope to be detected is obtained; the original multi-source monitoring data is preprocessed to obtain the target multi-source monitoring data; the target multi-source monitoring data is feature extracted to obtain the target feature data; the target feature data is input into the adaptive slope disaster early warning model, and the calculation result outputted is used to calculate the stability of the slope to be detected to obtain the stability evaluation result; the stability evaluation result is used to perform early warning processing and disaster prediction on the slope to be detected. Therefore, the slope disaster early warning model automatically adjusts the model parameters and structure, adapts to the dynamic evolution of the original multi-source monitoring data of the slope to be detected, realizes the analysis and calculation of the slope stability evaluation results, and performs disaster prediction and early warning processing based on this, which effectively improves the timeliness and accuracy of the prediction and early warning. .

[0115] For the above embodiments, please refer to Figure 2 , Figure 2 A schematic diagram of the structure of a terminal provided in an embodiment of the present application, such as Figure 2 As shown, it includes a processor, an input device, an output device and a memory, which are connected to each other, wherein the memory is used to store a computer program, the computer program includes program instructions, the processor is configured to call the program instructions, and the program includes instructions for executing the following steps;

[0116] Obtaining original multi-source monitoring data of the slope to be detected;

[0117] Preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data;

[0118] Extracting features from the target multi-source monitoring data to obtain target feature data;

[0119] Adopting an adaptive slope disaster early warning model to output the model calculation result through the target characteristic data;

[0120] Using the model output results to perform stability assessment on the slope to be tested, to obtain a stability assessment result;

[0121] The stability assessment result is used to perform early warning processing on the slope to be detected.

[0122] The above mainly introduces the scheme of the embodiment of the present application from the perspective of the execution process on the method side. It is understandable that in order to realize the above functions, the terminal includes a hardware structure and / or software module corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of each example described in the embodiments provided herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0123] The embodiment of the present application can divide the terminal into functional units according to the above method example. For example, each functional unit can be divided according to each function, or two or more functions can be integrated into one processing unit. The above integrated unit can be implemented in the form of hardware or in the form of software functional units. It should be noted that the division of units in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0124] In line with the above, see Figure 3 , Figure 3 The present invention provides a schematic structural diagram of an adaptive slope disaster warning device based on multi-source monitoring data. Figure 3 As shown, the device comprises:

[0125] An acquisition unit 301 is used to acquire original multi-source monitoring data of the slope to be detected;

[0126] A preprocessing unit 302 is used to preprocess the original multi-source monitoring data to obtain target multi-source monitoring data;

[0127] A feature extraction unit 303 is used to extract features from the target multi-source monitoring data to obtain target feature data;

[0128] An evaluation unit 304 is used to use the adaptive slope disaster early warning model to perform stability evaluation on the slope to be detected through the target feature data to obtain a stability evaluation result; and use the model output result to perform stability evaluation on the slope to be detected to obtain a stability evaluation result;

[0129] The early warning unit 305 is used to perform early warning processing on the slope to be detected using the stability assessment result.

[0130] In a possible implementation, the preprocessing unit 302 is specifically configured to:

[0131] Using a median filtering method to filter the original multi-source monitoring data to obtain first multi-source monitoring data;

[0132] Performing outlier removal processing on the first multi-source monitoring data to obtain second multi-source monitoring data;

[0133] Performing denoising processing on the second multi-source monitoring data to obtain third multi-source monitoring data;

[0134] The third multi-source monitoring data is standardized and smoothed to obtain the target multi-source monitoring data.

[0135] In a possible implementation, the target feature data includes data change trend information, data cycle information and emergency event information, and the feature extraction unit 303 is specifically used to:

[0136] Extracting data change trends from the target multi-source monitoring data to obtain data change trend information;

[0137] Using discrete Fourier transform to extract data period information from the target multi-source monitoring data;

[0138] The preset sensitivity information, data reference value and target multi-source monitoring data are used to perform emergency event identification to obtain the emergency event information.

[0139] In a possible implementation, the evaluation unit 304 is specifically configured to:

[0140] Performing vector construction on the target feature data to obtain a target feature data vector;

[0141] Obtaining hidden layer output data by performing calculations on the hidden state vector of the hidden layer of the adaptive slope disaster early warning model and the target feature data vector;

[0142] Using the attention weight matrix of the attention layer of the adaptive slope disaster early warning model to calculate the attention weight of the hidden layer output data, to obtain the attention weight value;

[0143] The adaptive slope disaster early warning model uses the attention weight value to perform weighted operation on the hidden layer output data to obtain a model output result; the model output result is used to perform stability assessment on the slope to be detected to obtain a stability assessment result; and an early warning unit is used to use the stability assessment result to perform early warning processing on the slope to be detected.

[0144] In a possible implementation, in terms of calculating the attention weight of the hidden layer output data using the attention weight matrix of the attention layer of the adaptive slope hazard warning model to obtain the attention weight value, the evaluation unit 304 is specifically used to:

[0145] The attention weight value is calculated by the method shown in the following formula to obtain the attention weight value:

[0146] T AW =softmax(W a tanh(U a ·T O +b a )),

[0147] Among them, W a and U a are the weight matrices of the attention layer, which are used to learn how to convert the hidden layer output data into attention scores, and b a represents the bias vector of the attention layer, tanh is the hyperbolic tangent activation function used to introduce nonlinearity, and softmax is the softmax function used to normalize the attention score to a probability distribution; T AW is the attention weight.

[0148] In a possible implementation, the early warning unit 305 is specifically configured to:

[0149] extracting a stability evaluation value from the stability evaluation result;

[0150] If the stability evaluation value is greater than or equal to the preset stability warning threshold, a warning process is performed;

[0151] If the stability evaluation value is less than the preset stability warning threshold, no warning processing is performed.

[0152] An embodiment of the present application also provides a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, and the computer program enables a computer to execute part or all of the steps of any one of the adaptive slope hazard warning methods based on multi-source monitoring data recorded in the above method embodiments.

[0153] An embodiment of the present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program enables a computer to execute part or all of the steps of any one of the adaptive slope hazard warning methods based on multi-source monitoring data recorded in the above method embodiments.

[0154] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0155] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0156] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the device or unit can be electrical or other forms.

[0157] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0158] In addition, the functional units in the various embodiments of the application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware or in the form of software program modules.

[0159] If the integrated unit is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, disk or optical disk and other media that can store program codes.

[0160] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which can include: a flash drive, a read-only memory, a random access memory, a magnetic disk or an optical disk, etc.

[0161] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. An adaptive slope disaster early warning method based on multi-source monitoring data, characterized in that: The method comprises: Obtaining original multi-source monitoring data of the slope to be detected; Preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data; Extracting features from the target multi-source monitoring data to obtain target feature data; Adopting an adaptive slope disaster early warning model to output the model calculation result through the target characteristic data; Using the model output results to perform stability assessment on the slope to be tested, to obtain a stability assessment result; The stability assessment result is used to perform early warning processing on the slope to be detected.

2. The adaptive slope disaster early warning method based on multi-source monitoring data according to claim 1 is characterized in that: Preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data includes: Using a median filtering method to filter the original multi-source monitoring data to obtain first multi-source monitoring data; Performing outlier removal processing on the first multi-source monitoring data to obtain second multi-source monitoring data; Performing denoising processing on the second multi-source monitoring data to obtain third multi-source monitoring data; The third multi-source monitoring data is standardized and smoothed to obtain the target multi-source monitoring data.

3. The adaptive slope disaster early warning method based on multi-source monitoring data according to claim 2 is characterized in that: The target feature data includes data change trend information, data cycle information and emergency event information. Feature extraction is performed on the target multi-source monitoring data to obtain target feature data, including: Extracting data change trends from the target multi-source monitoring data to obtain data change trend information; Using discrete Fourier transform to extract periodic information from the target multi-source monitoring data; The preset sensitivity information, data reference value and target multi-source monitoring data are used to perform emergency event identification to obtain the emergency event information.

4. The adaptive slope disaster early warning method based on multi-source monitoring data according to any one of claims 1 to 3, characterized in that: The method of using the adaptive slope disaster early warning model to perform model calculation on the slope to be detected through the target feature data and outputting the result includes: Performing vector construction on the target feature data to obtain a target feature data vector; Obtaining hidden layer output data by performing calculations on the hidden state vector of the hidden layer of the adaptive slope disaster early warning model and the target feature data vector; Using the attention weight matrix of the attention layer of the adaptive slope disaster early warning model to calculate the attention weight of the hidden layer output data, to obtain the attention weight value; Using the attention weight value to perform weighted operation on the hidden layer output data through the adaptive slope disaster early warning model, an output result is obtained; The stability of the slope to be detected is calculated using the model output results to obtain a stability assessment result, and slope early warning processing is performed based on the result.

5. The adaptive slope disaster early warning method based on multi-source monitoring data according to claim 4 is characterized in that: The attention weight matrix of the attention layer of the adaptive slope disaster early warning model is used to calculate the attention weight of the hidden layer output data to obtain the attention weight value, including: The attention weight value is calculated by the method shown in the following formula to obtain the attention weight value: T AW =softmax(W a ·tanh(U a ·T o +b a )), Among them, W a and U a are the weight matrices of the attention layer, which are used to learn how to convert the hidden layer output data into attention scores, and b a represents the bias vector of the attention layer, tanh is the hyperbolic tangent activation function used to introduce nonlinearity, and softmax is the softmax function used to normalize the attention score to a probability distribution; T AW is the attention weight.

6. The adaptive slope disaster early warning method based on multi-source monitoring data according to claim 5 is characterized in that: The using the stability assessment result to perform early warning processing on the slope to be detected includes: extracting a stability evaluation value from the stability evaluation result; If the stability evaluation value is greater than or equal to the preset stability warning threshold, a warning process is performed; If the stability evaluation value is less than the preset stability warning threshold, no warning processing is performed.

7. An adaptive slope disaster early warning device based on multi-source monitoring data, characterized in that: The device comprises: An acquisition unit, used for acquiring original multi-source monitoring data of the slope to be detected; A preprocessing unit, used for preprocessing the original multi-source monitoring data to obtain target multi-source monitoring data; A feature extraction unit, used to extract features from the target multi-source monitoring data to obtain target feature data; An evaluation unit is used to use an adaptive slope disaster early warning model to output a model result through the target feature data; and use the model output result to perform stability evaluation on the slope to be detected to obtain a stability evaluation result; An early warning unit is used to use the stability assessment result to perform early warning processing on the slope to be detected.

8. The adaptive slope disaster early warning device based on multi-source monitoring data according to claim 7 is characterized in that: The pre-processing unit is specifically used for: Using a median filtering method to filter the original multi-source monitoring data to obtain first multi-source monitoring data; Performing outlier removal processing on the first multi-source monitoring data to obtain second multi-source monitoring data; Performing denoising processing on the second multi-source monitoring data to obtain third multi-source monitoring data; The third multi-source monitoring data is standardized and smoothed to obtain the target multi-source monitoring data.

9. A terminal, characterized in that: The method comprises a processor, an input device, an output device and a memory, wherein the processor, the input device, the output device and the memory are interconnected, wherein the memory is used to store a computer program, the computer program comprises program instructions, and the processor is configured to call the program instructions to execute the adaptive slope hazard early warning method based on multi-source monitoring data as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which includes program instructions. When the program instructions are executed by a processor, the processor executes the adaptive slope hazard early warning method based on multi-source monitoring data as described in any one of claims 1 to 6.

Images