Debris flow early warning method, system, equipment, medium and product

Abstract

The invention discloses a debris flow early warning method, system and device, a medium and a product, and relates to the field of debris flow early warning, and the method comprises the steps: obtaining a current infrasound signal of the earth surface; calculating a power spectrum entropy value of the current infrasound signal; comparing the power spectrum entropy value of the current infrasound signal with an early warning threshold value; when the power spectrum entropy value of the current infrasound signal is smaller than the early warning threshold value, the peak frequency of the current infrasound signal is calculated, and when the peak frequency is within a set frequency range, it is judged that the current infrasound signal is a debris flow precursor signal; and outputting an early warning signal according to the debris flow precursor signal. According to the invention, early warning can be provided and the accuracy of debris flow early warning can be enhanced.

Description

Technical Field

[0001] The present application relates to the field of debris flow warning, and in particular to a debris flow warning method, system, equipment, medium and product. Background Art

[0002] Debris flow is a common geological disaster, mainly occurring in mountainous and hilly areas, triggered by natural factors such as heavy rainfall, snowmelt, earthquakes, or by unreasonable human activities. When a large amount of solid materials such as mud, sand, and stones mix with water, they slide down the slope or valley at high speed, with the characteristics of strong destructive power, high suddenness, and wide range of damage. Debris flow not only destroys infrastructure such as roads, bridges, and houses, but may also cause casualties. Therefore, how to effectively prevent and monitor debris flow disasters has become an important research topic in the field of disaster prevention and mitigation worldwide.

[0003] Regarding debris flow monitoring and early warning, the warning technologies mainly rely on surface vibration monitoring, mud and water level monitoring, remote sensing monitoring and video monitoring methods. However, the above technologies are easily interfered by other natural or man-made vibration sources, and there is a risk of false alarms or missed alarms, or it is difficult to provide early warnings. Therefore, it is necessary to provide a new debris flow early warning method to reduce the lag of debris flow early warning and enhance the accuracy of debris flow early warning. Summary of the invention

[0004] The purpose of this application is to provide a debris flow warning method, system, equipment, medium and product, which can reduce the lag of debris flow warning and enhance the accuracy of debris flow warning.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a debris flow early warning method, comprising:

[0007] Obtain the current infrasound signal on the surface;

[0008] Calculating a power spectrum entropy value of the current infrasound signal;

[0009] Comparing the power spectrum entropy value of the current infrasound signal with the warning threshold;

[0010] When the power spectrum entropy value of the current infrasound signal is less than the warning threshold, calculating the peak frequency of the current infrasound signal, and determining that the current infrasound signal is a debris flow precursor signal when the peak frequency is within a set frequency range;

[0011] Outputting a warning signal according to the debris flow precursor signal.

[0012] Optionally, obtaining a current infrasound signal on the surface of the earth specifically includes:

[0013] Collect original infrasound signals from the surface;

[0014] The original infrasound signal is filtered and denoised to obtain a current infrasound signal.

[0015] Optionally, filtering and denoising the original infrasound signal to obtain a current infrasound signal specifically includes:

[0016] Performing bandpass filtering on the original infrasound signal to obtain a filtered infrasound signal;

[0017] The environmental noise in the filtered infrasound signal is removed by wavelet transform to obtain the current infrasound signal.

[0018] Optionally, calculating the power spectrum entropy value of the current infrasound signal specifically includes:

[0019] Performing Fourier transform on the current infrasound signal to obtain a frequency domain signal;

[0020] estimating the power spectral density of the current infrasound signal according to the frequency domain signal by using the Welch method;

[0021] The power spectrum entropy value of the current infrasound signal is calculated according to the power spectrum density of the current infrasound signal.

[0022] Optionally, calculating the peak frequency of the current infrasound signal specifically includes:

[0023] Performing short-time Fourier transform or wavelet transform on the current infrasound signal to obtain a time-frequency graph;

[0024] Obtain a power spectrum diagram according to the time-frequency diagram;

[0025] The frequency corresponding to the maximum power value in the power spectrum is determined as the peak frequency of the current infrasound signal.

[0026] Optionally, the method for determining the warning threshold is:

[0027] Collect infrasound signals when there is no debris flow on the surface;

[0028] Calculate the power spectrum entropy value of the infrasound signal when there is no debris flow on the surface, and obtain the power spectrum entropy value range;

[0029] The lower limit of the power spectrum entropy value in the power spectrum entropy value range is determined as the warning threshold.

[0030] In a second aspect, the present application provides a debris flow early warning system, comprising:

[0031] A collection module, used to obtain the current infrasound signal on the surface;

[0032] A power spectrum entropy value calculation module, used to calculate the power spectrum entropy value of the current infrasound signal;

[0033] A power spectrum entropy value comparison module, used to compare the power spectrum entropy value of the current infrasound signal with the warning threshold;

[0034] A determination module, configured to calculate the peak frequency of the current infrasound signal when the power spectrum entropy value of the current infrasound signal is less than the warning threshold, and determine that the current infrasound signal is a debris flow precursor signal when the peak frequency is within a set frequency range;

[0035] The warning signal output module is used to output a warning signal according to the debris flow precursor signal.

[0036] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-mentioned debris flow warning steps.

[0037] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned debris flow warning steps.

[0038] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned debris flow warning steps.

[0039] According to the specific embodiments provided in this application, this application has the following technical effects:

[0040] The present application provides a debris flow early warning method, system, equipment, medium and product, which collect the current infrasound signal on the surface as the basis for early warning. Compared with traditional debris flow monitoring, infrasound signals are more stable and have strong long-distance transmission capabilities, and can provide early warnings. Infrasound monitoring can effectively distinguish debris flow signals from other types of vibration or noise sources, reducing the false alarm problem in traditional monitoring; power spectrum entropy is introduced as an abnormality detection indicator, and the peak frequency is used to further verify whether the source of the abnormal signal meets the characteristics of the debris flow. Power spectrum entropy and peak frequency are used as dual criteria for dual verification, which not only improves the accuracy of debris flow early warning, but also can better distinguish debris flow signals from other noise signals, reduce false alarm rates, enhance the ability to identify debris flow precursor signals, and provide advanced early warnings with higher accuracy and anti-interference capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] 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 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.

[0042] Figure 1 A schematic diagram of a flow chart of a debris flow early warning method in an embodiment of the present application;

[0043] Figure 2 A time sequence diagram, a time-frequency diagram, and a power spectrum diagram of an infrasound signal provided in an embodiment of the present application when a debris flow has not occurred;

[0044] Figure 3 A time sequence diagram, a time-frequency diagram, and a power spectrum diagram of an infrasound signal when a debris flow disaster occurs provided in an embodiment of the present application;

[0045] Figure 4 A schematic diagram of the functional modules of a debris flow warning system provided in one embodiment of the present application;

[0046] Figure 5 A flowchart of a debris flow warning system provided in accordance with an embodiment of the present application;

[0047] Figure 6 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0048] 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.

[0049] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0050] Among the relevant debris flow monitoring methods, surface vibration monitoring determines whether a debris flow has occurred by deploying seismic sensors in debris flow-prone areas and detecting seismic waves or vibration signals. This method can provide relatively timely warnings for fast and violent debris flows, but it is easily interfered by other natural or man-made vibration sources and there is a risk of false alarms or missed alarms. Mud and water level monitoring is to install water level sensors at common locations or basins of debris flows and provide early warnings by monitoring changes in mud and water levels. Although this method is intuitive and effective, the suddenness and rapid expansion of debris flows often cause water level change signals to lag behind, making it difficult to provide early warnings. Remote sensing monitoring and video monitoring monitor debris flows through satellite remote sensing or video monitoring. This method is suitable for large-scale observations, but its accuracy and real-time performance are not high, and it depends on weather and equipment conditions, making it difficult to provide continuous and real-time warnings. The defects of the above-mentioned related schemes include:

[0051] Lag: Methods such as mud and water level monitoring often detect obvious signals only after the debris flow has already started, and cannot provide sufficient warning time.

[0052] Many interferences: Surface vibration monitoring is easily interfered by non-debris flow sources such as traffic and construction activities, resulting in false alarms.

[0053] Limited coverage: Remote sensing and video surveillance are often only targeted at specific areas and require a large amount of equipment investment, and it is difficult to meet the requirements of real-time and coverage.

[0054] Therefore, based on the above problems, the present application provides a debris flow warning method, system, equipment, medium and product, which can realize an early warning, have higher accuracy and real-time performance, and cover a wide range of debris flow warnings.

[0055] In an exemplary embodiment, Figure 1 As shown, a debris flow early warning method is provided, comprising the following steps 101 to 105. Among them:

[0056] Step 101: Acquire the current infrasound signal on the ground surface.

[0057] Step 102: Calculate the power spectrum entropy value of the current infrasound signal.

[0058] Step 103: Compare the power spectrum entropy value of the current infrasound signal with the warning threshold.

[0059] Step 104: When the power spectrum entropy value of the current infrasound signal is less than the warning threshold, the peak frequency of the current infrasound signal is calculated, and when the peak frequency is within a set frequency range, the current infrasound signal is determined to be a debris flow precursor signal. The set frequency range is the peak frequency range of the infrasound signal when debris flow occurs on the surface.

[0060] Step 105: Outputting a warning signal according to the debris flow precursor signal.

[0061] In another exemplary embodiment of the present application, step 101 specifically includes:

[0062] The original infrasound signal of the ground surface is collected; the original infrasound signal is filtered and denoised to obtain a current infrasound signal.

[0063] The filtering and denoising of the original infrasound signal to obtain the current infrasound signal specifically includes: bandpass filtering the original infrasound signal to obtain a filtered infrasound signal; and removing environmental noise from the filtered infrasound signal by wavelet transform to obtain the current infrasound signal.

[0064] In another exemplary embodiment of the present application, step 102 specifically includes:

[0065] Performing Fourier transform on the current infrasound signal to obtain a frequency domain signal; estimating the power spectrum density of the current infrasound signal according to the frequency domain signal by using the Welch method; and calculating the power spectrum entropy value of the current infrasound signal according to the power spectrum density of the current infrasound signal.

[0066] In another exemplary embodiment of the present application, calculating the peak frequency of the current infrasound signal in step 104 specifically includes:

[0067] Performing short-time Fourier transform (STFT) or wavelet transform on the current infrasound signal to obtain a time-frequency diagram; obtaining a power spectrum diagram according to the time-frequency diagram; and determining the frequency corresponding to the maximum power value in the power spectrum diagram as the peak frequency of the current infrasound signal.

[0068] In another exemplary embodiment of the present application, the method for determining the warning threshold in step 103 is:

[0069] The infrasound signal is collected when no debris flow occurs on the ground surface; the power spectrum entropy value of the infrasound signal is calculated when no debris flow occurs on the ground surface to obtain a power spectrum entropy value range; and the lower limit of the power spectrum entropy value in the power spectrum entropy value range is determined as a warning threshold.

[0070] In another exemplary embodiment of the present application, the warning threshold in step 103 can also be optimized by statistical analysis or threshold selection algorithm in machine learning (such as Otsu algorithm, K-means clustering, etc.) to reduce the false alarm rate.

[0071] In another exemplary embodiment of the present application, in order to improve the accuracy of early warning, a machine learning algorithm can also be used to assist in analysis and optimization, and can adaptively optimize and determine whether the current infrasound signal is a debris flow precursor signal. This embodiment mainly involves steps such as signal processing, feature extraction, and model building, and combines machine learning algorithms and calculation methods to optimize the design of the model.

[0072] Establish a debris-rock infrasound signal recognition model.

[0073] The current infrasound signal is input into a debris flow infrasound signal recognition model to determine the infrasound source type of the current infrasound signal, that is, to determine whether the current infrasound signal is a debris flow precursor signal.

[0074] The method for establishing the debris infrasound signal recognition model specifically includes:

[0075] Obtain infrasound source training data S1; the infrasound source training data S1 is an infrasound signal obtained by detecting a sound source of known type.

[0076] The infrasound source training data S1 is subjected to low-pass filtering by using a Chebyshev II filter to obtain filtered infrasound source training data. The filtered infrasound source training data only retains the infrasound part, that is, the part with a frequency less than 20 Hz.

[0077] Determine whether the sound pressure of the filtered infrasound source training data is less than the set sound pressure. The set sound pressure can be set to 5Pa. When the sound pressure P1 of the filtered infrasound source training data is less than 5Pa, continue to distinguish, otherwise it is regarded as a non-mudslide infrasound signal.

[0078] If the sound pressure of the filtered infrasound source training data is less than the set sound pressure, a digital signal processing method is used to determine the short-time zero-crossing rate, peak frequency, center frequency and spectral line width of the filtered infrasound source training data, and the short-time zero-crossing rate, peak frequency, center frequency and spectral line width of the filtered infrasound source training data are used as the characteristic values ​​of the infrasound source training data; the digital signal processing method includes fast Fourier transform, first-order moment and second-order moment spectrum estimation methods, etc. The short-time zero-crossing rate, peak frequency, center frequency and spectral line width are respectively denoted as g1, f1, m1, and w1.

[0079] Establish the feature vector corresponding to the feature value of the infrasound source training data. That is: [g1, f1, m1, w1]; the infrasound source training data selected in this embodiment include: debris flow, lightning, single blasting, multiple blasting, engine and wind, and the corresponding feature vectors are: debris flow: [0.26, 8.6, 11.4, 9.3], lightning: [0.12, 2.8, 4.8, 3.8], single blasting: [0.06, 4.5, 8.3, 6.4], multiple blasting: [0.06, 4.5, 8.3, 6.4], engine: [0.22, 5.2, 10.4, 6.7], wind: [0.11, 2.7, 5.6, 3.9].

[0080] The feature vector of the infrasound source training data S1 is subjected to 2-norm regularization processing to obtain the regularized feature vector of the infrasound source training data. That is, each of the four types of features, namely, short-time zero-crossing rate, peak frequency, center frequency and spectral line width, is subjected to standardization processing separately.

[0081] According to the eigenvector of the regularized infrasound source training data and the corresponding infrasound source type, the KNN (K-nearest neighbor classification algorithm, k-NearestNeighbor) model (K=1, using Euclidean distance) is trained to obtain a debris infrasound signal recognition model. In this embodiment, the KNN model is used to find the sound source category of the nearest neighbor to the infrasound source training data S1, that is, the sound source type of the infrasound source training data S1, thereby realizing the category judgment of the current infrasound signal.

[0082] As shown in Table 1, the infrasound source training data S1 of this embodiment extracts the key features of debris flow infrasound and environmental interference infrasound. With the help of Matlab, Python and K-nearest neighbor classification algorithm (KNN), a debris flow infrasound signal recognition model is established to realize the discrimination of the current infrasound signal monitored, greatly reducing the false alarm caused by the influence of environmental interference noise. According to the debris flow infrasound signal recognition model and the working mode of signal reception and storage, 1000 sampling points can be obtained from the database every 10 seconds, and these sampling points can be used as a section of the signal to be identified (i.e., the current infrasound signal) for discrimination to confirm whether this section of the signal is a debris flow precursor signal.

[0083] Table 1 Key characteristics of infrasound signals from six sound sources

[0084]

[0085] Since Matlab has powerful digital signal processing functions and provides many ready-made signal processing function libraries for direct calling, Matlab can be used to implement filtering, denoising and key feature extraction of the current infrasound signal. However, Matlab cannot access the database and does not have rich machine learning libraries. Python can not only be well integrated with the database, but also provides a large number of function libraries that can be used for machine learning, and can also meet the requirements of graphical interface design. To this end, based on the above-established debris flow infrasound signal recognition model, an intelligent identification system for debris flow infrasound signals was developed using Python, Matlab and Oracle database. Matlab is used to implement the filtering, denoising and key feature extraction functions of the current infrasound signal, and it is compiled into an independent function module for Python to call. Python is responsible for taking the current infrasound signal data out of the database and passing it to Matlab for analysis. Matlab is responsible for passing the key features of the current infrasound signal to Python for it to determine whether the current infrasound signal is a debris flow precursor signal.

[0086] The model runs on the client and can access the monitoring data stored in the database on the server in real time. The model can identify the 1,000 sampling points obtained every 10 seconds to determine whether it is a debris flow precursor signal. If this signal is identified as a non-debris flow infrasound signal, the next signal will continue to be obtained for identification. Otherwise, the alarm link will be entered. After the model is running, it has been in a fully automatic operation state without manual intervention. At the same time, the operation efficiency of the recognition model is very high, and the identification process of each signal segment takes an extremely short time (about 200ms), which can well meet the real-time requirements. In addition, the model also provides a good data expansion interface, so that more categories of environmental interference infrasound can be accessed in the future, further improving the recognition accuracy of debris flow precursor signals.

[0087] In order to test the working performance of the debris flow infrasound signal recognition model established, the recognition accuracy of the recognition model was evaluated using the test set data consisting of the above debris flow infrasound and various types of environmental interference infrasound. In the test set data, each type of infrasound source has 100 valid samples (including 6 types of infrasound sources in total), so the test set has a total of 600 samples.

[0088] All the test set data are input into the recognition model, and the model can automatically perform recognition and classification. The discrimination results of these test set data are shown in Table 2. A total of 38 sampled data were misidentified, of which 11 debris flow infrasound sampling data were all identified as automobile infrasound signals, which were missed reports; 9 non-debris flow infrasound sampling data were identified as debris flow infrasound signals, which were false reports; and the remaining 18 sampled data were identified as non-self-environmental interference infrasound (this part has no substantial impact on the application effect of debris flow infrasound monitoring). Through the recognition results, it can be seen that the recognition false alarm rate and missed report rate of the model are very low.

[0089] Table 2. Discrimination results of test set data

[0090]

[0091] The verification results of the test set data show that the debris flow infrasound signal recognition model has a high recognition accuracy. Based on the debris flow infrasound signal recognition model, the debris flow infrasound monitoring equipment is scientifically and rationally selected and deployed to minimize the impact of environmental interference noise, especially the impact of wind noise, which can provide accurate and reliable information support for on-site debris flow infrasound monitoring and early warning.

[0092] As an optional execution method, the model can also be established and trained in the following manner.

[0093] Supervised learning: Use labeled historical infrasound data to train classification models. For example, use classification algorithms such as support vector machines (SVM) and random forests to establish a classification model for whether a debris flow has occurred based on features such as power spectrum entropy and peak frequency. After training, the model can determine in real time whether the current infrasound signal is a precursor to a debris flow.

[0094] Unsupervised learning: Clustering algorithms (K-means, DBSCAN) can be used to automatically discover abnormal patterns from unlabeled infrasound data and further optimize warning thresholds.

[0095] Deep learning: Use convolutional neural networks (CNN) or long short-term memory networks (LSTM) to perform deep feature extraction on time series signals, analyze complex time-frequency features, and improve the model’s detection accuracy for debris flow infrasound signals.

[0096] The overall workflow of the debris flow early warning method of this application is as follows:

[0097] (1) Data collection and preprocessing:

[0098] Acquisition equipment: Use infrasound sensors to conduct real-time monitoring in debris flow prone areas and record surface infrasound signal data. The acquisition frequency is usually in the low frequency range (such as 0.1Hz to 20Hz), because debris flow infrasound signals are generally low frequency signals.

[0099] Data collection when no debris flow occurs: Infrasound signals are collected for a long time (e.g., one month) when no debris flow occurs to ensure the diversity and reliability of data samples.

[0100] Debris flow status data collection: When a debris flow occurs, infrasound signals are collected for comparison and analysis.

[0101] Preprocessing methods: including signal denoising and filtering. After the infrasound signal is collected, it is first band-pass filtered to retain the frequency range useful for debris flow monitoring (e.g. 0.1Hz to 20Hz) and filter out high-frequency or low-frequency noise. After that, denoising algorithms such as wavelet transform can be used to further remove environmental noise to ensure the accuracy of the signal.

[0102] (2) Data analysis:

[0103] Real-time signal processing: Use MATLAB program to process the current infrasound signal collected in real time, calculate the power spectrum entropy and time-frequency characteristics, and realize real-time warning.

[0104] Fourier transform: Perform fast Fourier transform (FFT) on the filtered infrasound signal to obtain a frequency signal.

[0105] Calculation of the power spectrum entropy range when debris flow does not occur: Analyze the power spectrum entropy of the infrasound signal when debris flow does not occur, obtain the entropy value range (e.g., 4.37-4.42) when debris flow does not occur, and then set the warning value: Set the lower limit of the power spectrum entropy (e.g., 4.37) under normal conditions as the warning value. When the power spectrum entropy is lower than 4.37, it is a potential abnormal state and proceed to the next step.

[0106] Power spectrum calculation: Perform Fourier transform on the collected infrasound signal, and use the FFT result to calculate the power spectral density (PSD) of the signal. The PSD can be estimated by the Welch method, etc. to reduce the variance in the spectrum estimation.

[0107] Calculation of power spectrum entropy: Information entropy is a commonly used basic concept in machine learning. It is an extension of the concept of information quantity. Information entropy is generally used to represent the chaos of events. Models such as logistic regression and decision trees involve relevant knowledge of information entropy. Power spectral entropy (PSE) is an entropy index used to measure the uniformity of signal frequency distribution. According to the power spectrum density, the power spectrum entropy value of each signal segment is calculated. The power spectrum entropy value is calculated using the following formula:

[0108]

[0109] Among them, P(f i ) is the normalized power spectrum density of the ith frequency component, and n represents the number of frequency components. H represents the power spectrum entropy value. The lower the entropy value, the stronger the concentration of the signal spectrum, indicating that the possibility of abnormality of the infrasound signal increases.

[0110] Entropy value range when debris flow does not occur: Based on historical infrasound signal data, the power spectrum entropy value range when debris flow does not occur is calculated. For example, by long-term monitoring of infrasound signals when debris flow does not occur, a normal power spectrum entropy value range (such as 4.37-4.42) is statistically obtained. This entropy value range is usually further refined and layered to adapt to changes in different environments.

[0111] (3) Real-time signal analysis and entropy monitoring:

[0112] Real-time entropy calculation: With the real-time collection of infrasound signals, the early warning system will continuously calculate the power spectrum entropy of the current signal. If the current entropy value is lower than the early warning threshold, it means that the signal characteristics have changed significantly, which may indicate the occurrence of debris flow, and the time-frequency characteristics will be further analyzed.

[0113] (4) Time-frequency feature analysis and peak frequency calculation:

[0114] Time-frequency analysis: To further determine the abnormal characteristics of the signal, short-time Fourier transform (STFT) or wavelet transform is used to analyze the signal in the time-frequency domain. The time-frequency diagram can reflect the frequency changes of the signal at different time points.

[0115] Peak frequency calculation: Extract the peak frequency of the signal from the real-time monitoring signal. The peak frequency refers to the frequency corresponding to the maximum value in the signal power spectrum. Find the peak frequency of the signal in the time-frequency diagram. If this peak frequency is between 2Hz and 20Hz, and the power spectrum entropy is also in the abnormal range, it can be preliminarily determined as a debris flow precursor, and a debris flow warning is triggered.

[0116] Compared with the traditional single feature analysis method, this application combines the dual criteria of power spectrum entropy and peak frequency. When the infrasound signal collected in real time meets two conditions: 1) the power spectrum entropy is lower than the set warning value, 2) the peak frequency is between 2Hz and 20Hz, the system will automatically trigger the debris flow warning and send out an alarm signal. This multiple verification mechanism can reduce false alarms and provide more reliable technical support for actual debris flow monitoring, enhancing the sensitivity and accuracy of debris flow precursor signals. In addition, the introduction of machine learning algorithms can adaptively optimize the warning model and further reduce the false alarm rate and missed alarm rate. This improvement effectively overcomes the defects of traditional methods that are sensitive to noise and lack of early warning capabilities.

[0117] This application combines the time-frequency analysis of debris flow infrasound signals with power spectrum entropy to further improve the accuracy and reliability of early warning. The specific idea is that in addition to calculating the power spectrum entropy, MATLAB is used to analyze the time-frequency characteristics of the infrasound signal in real time, draw the time-frequency diagram of the infrasound signal, and analyze whether the peak frequency of the infrasound signal is between 2Hz and 20Hz when the power spectrum entropy is lower than the early warning value.

[0118] During the movement of debris flows, they will generate infrasound waves with a frequency below 20 Hz. These infrasound signals can penetrate the noise interference in the atmosphere, have a long propagation distance and are less affected by weather conditions. Through the introduction of infrasound monitoring technology, compared with traditional surface vibration monitoring, infrasound signals are more stable and have strong long-distance transmission capabilities, which can provide early warning. It can also reduce the false alarm rate: infrasound monitoring can effectively distinguish debris flow signals from other types of vibration or noise sources, and further optimize the identification model through multi-frequency infrasound analysis to reduce the false alarm problem in traditional earthquake monitoring.

[0119] Compare the infrasound signals of the state without debris flow with those of the state with debris flow. Figure 2 Part (a) shows the time sequence diagram of the debris flow infrasound signal collected by the equipment when the debris flow does not occur. Figure 2 Part (b) shows the time-frequency diagram of the debris flow infrasound signal collected by the equipment when the debris flow does not occur. Figure 2 Part (c) shows the power spectrum of the debris flow infrasound signal collected by the equipment when there is no debris flow. It can be seen from the timing diagram that under normal conditions, the amplitude of the debris flow infrasound signal collected is maintained between -20mv and +20mV, and there is no sudden increase in amplitude. The time-frequency diagram shows that the frequency range corresponding to the signal is relatively large, approximately between 5Hz and 50Hz. Further analysis of the power spectrum shows that the signal frequency range is between 1.5Hz and 50Hz, while the debris flow infrasound signal frequency is in the range of 2-20Hz. It is thus judged that this section of the infrasound signal is the infrasound signal generated by the debris flow ditch background noise, rather than the infrasound signal generated by the debris flow.

[0120] like Figure 3 The following is the infrasound signal collected by the equipment when a debris flow geological disaster occurs. Figure 3 Part (a) shows the time sequence diagram of the infrasound signal collected by the equipment when a debris flow geological disaster occurs. Figure 3 Part (b) shows the time-frequency diagram of the infrasound signal collected by the equipment when a debris flow geological disaster occurs. Figure 3 Part (c) shows the power spectrum of the infrasound signal collected by the equipment when a debris flow geological disaster occurs. It can be seen that Figure 3 and Figure 2 Shows a clear difference. Figure 3 The signal timing diagram shows that the signal amplitude has changed, and the amplitude has increased significantly between -200mV and 350mv. The time-frequency diagram has also changed, showing that the signal energy is concentrated between 35s and 42s, and the peak frequency is distributed below 8Hz. Further analysis of the power spectrum density diagram shows that when a debris flow occurs, higher energy appears, and the peak frequency is mostly between 2Hz and 7.5Hz. It can be judged that this section of the signal is an infrasound signal generated when a debris flow disaster occurs, and an early warning should be issued.

[0121] Based on the same inventive concept, the embodiment of the present application also provides a debris flow warning system for implementing the debris flow warning method involved above. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more debris flow warning system embodiments provided below can refer to the limitations of the debris flow warning method above, and will not be repeated here.

[0122] In an exemplary embodiment, Figure 4 As shown, a debris flow early warning system is provided, comprising:

[0123] The acquisition module is used to obtain the current infrasound signal on the surface.

[0124] The power spectrum entropy value calculation module is used to calculate the power spectrum entropy value of the current infrasound signal.

[0125] The power spectrum entropy value comparison module is used to compare the power spectrum entropy value of the current infrasound signal with the warning threshold.

[0126] The determination module is used to calculate the peak frequency of the current infrasound signal when the power spectrum entropy value of the current infrasound signal is less than the warning threshold, and determine that the current infrasound signal is a debris flow precursor signal when the peak frequency is within a set frequency range.

[0127] The warning signal output module is used to output a warning signal according to the debris flow precursor signal.

[0128] Specifically, the acquisition module can use infrasound measurement equipment, such as a high-sensitivity infrasound sensor, to receive debris flow infrasound signals in real time, and the data acquisition system: connect the infrasound sensor to the data acquisition system to realize real-time signal acquisition. Then use the signal processing software, through the MATLAB program, to calculate the power spectrum and power spectrum entropy of the infrasound signal in real time, and display it graphically. Because the larger the information entropy, the greater the uncertainty, which means the less information contained in the signal. On the contrary, the smaller the information entropy, the greater the certainty of the signal, which means the more information contained in the signal. Based on this, assuming that there is no debris flow geological disaster at ordinary times, the signal collected by the debris flow infrasound signal measurement equipment is defined as a signal under normal conditions. At this time, the value of the power spectrum entropy of the infrasound signal will change within a certain range, such as a range of 4.37-4.42, which is the power spectrum entropy value range of the infrasound signal under normal conditions. For this reason, this value range can be used as an early warning value (assuming that the early warning value K is 4.37). When the power spectrum entropy calculation value of the infrasound signal collected by the infrasound measurement equipment is not within this range, based on the principle that the smaller the information entropy, the greater the amount of information contained, then when the calculated power spectrum entropy L value is less than the warning value of 4.37, it is considered that the infrasound signal collected at this time may be the infrasound signal generated when the debris flow occurs. In this case, the time-frequency characteristics of the infrasound signal are further analyzed by MATLAB, and the time-frequency diagram of the infrasound signal is drawn. When the power spectrum entropy is calculated below the warning value, the peak frequency is calculated, and then it is determined whether the peak frequency is between 2Hz and 20Hz, because the peak frequency of the infrasound signal when the debris flow occurs is between 2Hz and 20Hz, that is, a debris flow has occurred, and an early warning is issued at the same time.

[0129] like Figure 5 As shown in the figure, the working process of the debris flow warning system based on power spectrum entropy and peak frequency is described. The specific working steps are as follows:

[0130] 1. Infrasound measurement equipment collects infrasound data in real time:

[0131] Infrasound sensors are installed in areas prone to debris flow to collect infrasound signals in real time. These signals provide raw data for subsequent early warning analysis.

[0132] 2.MATLAB extracts data in real time for analysis:

[0133] The collected infrasound signals are input into the MATLAB platform to extract data in real time for processing. MATLAB acts as a data analysis tool in this link, quickly extracting and analyzing the features of the signal.

[0134] 3.MATLAB real-time signal processing:

[0135] In this step, the signal is further processed, which may involve filtering, denoising, Fourier transform, etc., in order to prepare for the subsequent power spectrum entropy and peak frequency calculation.

[0136] 4. Calculate the power spectrum entropy, give the power spectrum graph and calculate the power spectrum entropy value:

[0137] The power spectrum entropy value is calculated using Fourier transform and power spectrum density (PSD). The power spectrum entropy value measures the uniformity of the frequency distribution of the signal and can reflect the complexity of the infrasound signal. When the entropy value decreases, it usually means that the signal is concentrated on certain specific frequencies, which may be a precursor to a debris flow.

[0138] 5. Comparative analysis of power spectrum entropy value and warning threshold K:

[0139] The calculated power spectrum entropy value is compared with the preset warning threshold value K. The K value is usually the lower limit of the power spectrum entropy value of the infrasound signal when there is no debris flow. If the current power spectrum entropy value is less than K (that is, the signal becomes abnormal), it is considered that the signal has entered an abnormal state and there is a possibility of debris flow, and the next step of analysis is continued; if it is not less than K, it returns to collect more data.

[0140] 6. Calculate the peak frequency f k And determine whether it is between 2Hz and 20Hz:

[0141] In the case of abnormal power spectrum entropy value, the peak frequency of the infrasound signal is further calculated. The peak frequency of the debris flow infrasound signal usually falls between 2Hz and 20Hz. If the calculated peak frequency is within this range, it is considered that the possibility of debris flow is very high.

[0142] 7. Issue warnings:

[0143] When the power spectrum entropy value is less than the warning threshold K and the peak frequency f k In the range of 2Hz to 20Hz, the system issues a debris flow warning signal to remind relevant departments to take countermeasures.

[0144] The entire system can effectively identify debris flow precursors through real-time monitoring of infrasound signals, combined with the dual criteria of power spectrum entropy and peak frequency. When the power spectrum entropy of the infrasound signal decreases and the peak frequency of the signal is within a specific range, the system will trigger an early warning. The flowchart clarifies the entire process from signal acquisition to early warning issuance, using MATLAB as a data processing and analysis platform to achieve real-time monitoring and early warning functions.

[0145] This application combines power spectrum entropy with time-frequency analysis to improve the accuracy of debris flow infrasound warning. On the basis of real-time calculation of power spectrum entropy, further analysis of the time-frequency characteristics of the signal is performed to determine whether the peak frequency of the signal is between 2Hz and 20Hz, which can effectively reduce the false alarm rate.

[0146] This application realizes debris flow warning by combining power spectrum entropy and peak frequency, which is significantly different from the existing technology in many technical aspects and has the following key technical advantages:

[0147] 1. Anomaly recognition based on power spectrum entropy.

[0148] Difference: Traditional debris flow warning systems rely on vibration sensors, mud and water level monitoring, rainfall monitoring, etc., focusing on changes in environmental physical parameters rather than the spectral characteristics of signals. Power spectrum entropy, as an indicator of signal complexity, is rarely used in existing debris flow warning technologies.

[0149] Advantages: Power spectrum entropy can reflect the uniformity of signal distribution in the frequency domain. Before a debris flow occurs, infrasound signals usually concentrate from scattered frequency distribution to a specific frequency range, and the entropy value decreases significantly. Therefore, by monitoring the power spectrum entropy, signal anomalies can be detected in advance, which provides a more sensitive and reliable criterion for debris flow warning, making up for the shortcomings of traditional physical parameter monitoring, which is delayed and has a high false alarm rate.

[0150] 2. Double criteria: power spectrum entropy combined with peak frequency.

[0151] Difference: Traditional technologies usually use a single physical parameter or signal feature for early warning (such as vibration signals or water level changes), which makes it easy to be interfered by external factors and has a high false alarm rate. In contrast, the present invention combines the two features of power spectrum entropy and peak frequency to form a dual criterion.

[0152] Advantages: This dual criterion can effectively reduce false alarms. The reduction of power spectrum entropy can be used as a preliminary judgment of the abnormality of debris flow infrasound signals, while the peak frequency further verifies whether the source of the abnormal signal meets the characteristics of debris flow. Through such dual verification, the early warning system not only improves the accuracy of debris flow warnings, but also better distinguishes debris flow signals from other noise signals, reduces false alarm rates, and increases the reliability of early warnings.

[0153] 3. Real-time signal analysis and MATLAB processing platform.

[0154] Real-time processing enables the early warning system to capture changes in infrasound signals in a timely manner and achieve a faster response. This is particularly important in rapid disasters such as debris flows, which can provide decision makers with valuable reaction time. In addition, as a mature data processing platform, MATLAB has powerful signal processing and algorithm implementation capabilities, and can efficiently handle complex time-frequency analysis tasks to ensure the stability and operability of the early warning system.

[0155] 4. Flexibility of time-frequency domain analysis.

[0156] Difference: Traditional early warning systems focus on signal processing in the time domain, such as monitoring changes in signal amplitude, while ignoring the potential of frequency domain analysis. This application introduces short-time Fourier transform (STFT) or wavelet transform to perform joint analysis in the time and frequency domains, mining the spectral characteristics of the signal in different time periods.

[0157] Advantages: Time-frequency domain analysis provides more detailed signal features and can detect the gradual and concentrated changes in signal frequency before a debris flow occurs. Compared with traditional single time or frequency analysis, this dynamic feature extraction method is more comprehensive and can more accurately identify precursor signals of debris flows, further improving the sensitivity of early warning.

[0158] 5. Adaptive warning value setting.

[0159] Difference: Many existing systems use fixed thresholds to determine whether to issue an early warning. However, due to the large differences in environmental conditions in different regions and at different times, fixed thresholds are difficult to adapt to changing environments.

[0160] Advantages: In this application, the warning value of the power spectrum entropy can be adaptively adjusted according to historical data. By counting the entropy value range of the infrasound signal when no debris flow occurs, this application can dynamically set the warning threshold to adapt to changes in different terrains, climates, and background noise. This adaptive capability enables the system to always maintain high sensitivity and accuracy in different environments.

[0161] 6. Machine learning and signal feature extraction.

[0162] Difference: Many existing debris flow warning systems rely on thresholds set by rules or experience, and rarely introduce machine learning algorithms for automatic extraction and classification of signal features.

[0163] Advantages: This application can train a large amount of historical infrasound data through machine learning algorithms (such as support vector machines, random forests, etc.), learn the abnormal characteristics of infrasound signals, and classify and predict them. Compared with fixed threshold settings, machine learning can improve the prediction accuracy of the model through self-learning, especially in complex environments, and can better adapt to noise and background changes.

[0164] 7. Enhanced anti-interference capability.

[0165] Difference: Traditional debris flow warning technology relies on a single signal feature (such as surface vibration), which is easily interfered by other natural phenomena or human activities, resulting in frequent false alarms. For example, traffic vibrations and ordinary sediment movement caused by rainfall will trigger traditional warning systems.

[0166] Advantages: This application further verifies the occurrence of debris flow by analyzing the entropy value of infrasound signals and combining the dual criteria of peak frequency, and can effectively filter out signals that do not meet the characteristics of debris flow. In particular, the power spectrum entropy of infrasound signals is relatively stable under environmental noise, while the occurrence of debris flow will cause obvious spectrum concentration, so this method can effectively resist interference and reduce false alarms.

[0167] The main difference between this application and the existing debris flow warning technology is that the power spectrum entropy is introduced as an abnormality detection indicator, and the peak frequency dual criterion is combined to enhance the recognition ability of debris flow precursor signals. In addition, based on the introduction of real-time signal processing, adaptive threshold setting, time-frequency domain analysis and machine learning technology, the warning system of this application has higher sensitivity, accuracy and anti-interference ability. Compared with the existing technology, this method can not only provide earlier and more accurate warnings, but also reduce false alarms and enhance the reliability of the warning system in complex environments.

[0168] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device 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 internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store debris flow warning data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication 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, a debris flow warning method is implemented.

[0169] Those skilled in the art will understand that Figure 6The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0170] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0171] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0172] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0173] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0174] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0175] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0176] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A debris flow early warning method, characterized in that: The debris flow early warning method comprises: Obtain the current infrasound signal on the surface; Calculating a power spectrum entropy value of the current infrasound signal; Comparing the power spectrum entropy value of the current infrasound signal with the warning threshold; When the power spectrum entropy value of the current infrasound signal is less than the warning threshold, calculating the peak frequency of the current infrasound signal, and determining that the current infrasound signal is a debris flow precursor signal when the peak frequency is within a set frequency range; Outputting a warning signal according to the debris flow precursor signal.

2. The debris flow early warning method according to claim 1, characterized in that: Obtain the current infrasound signal on the surface, including: Collect original infrasound signals from the surface; The original infrasound signal is filtered and denoised to obtain a current infrasound signal.

3. The debris flow early warning method according to claim 2, characterized in that: Filtering and denoising the original infrasound signal to obtain a current infrasound signal specifically includes: Performing bandpass filtering on the original infrasound signal to obtain a filtered infrasound signal; The environmental noise in the filtered infrasound signal is removed by wavelet transform to obtain the current infrasound signal.

4. The debris flow early warning method according to claim 1, characterized in that: Calculating the power spectrum entropy value of the current infrasound signal specifically includes: Performing Fourier transform on the current infrasound signal to obtain a frequency domain signal; estimating the power spectral density of the current infrasound signal according to the frequency domain signal by using the Welch method; A power spectrum entropy value of the current infrasound signal is calculated according to the power spectrum density of the current infrasound signal.

5. The debris flow early warning method according to claim 1, characterized in that: Calculating the peak frequency of the current infrasound signal specifically includes: Performing short-time Fourier transform or wavelet transform on the current infrasound signal to obtain a time-frequency graph; Obtain a power spectrum diagram according to the time-frequency diagram; The frequency corresponding to the maximum power value in the power spectrum is determined as the peak frequency of the current infrasound signal.

6. The debris flow early warning method according to claim 1, characterized in that: The method for determining the warning threshold is: Collect infrasound signals when there is no debris flow on the surface; Calculate the power spectrum entropy value of the infrasound signal when there is no debris flow on the surface, and obtain the power spectrum entropy value range; The lower limit of the power spectrum entropy value in the power spectrum entropy value range is determined as the warning threshold.

7. A debris flow early warning system, characterized in that: The debris flow early warning system comprises: A collection module, used to obtain the current infrasound signal on the surface; A power spectrum entropy value calculation module, used to calculate the power spectrum entropy value of the current infrasound signal; A power spectrum entropy value comparison module, used to compare the power spectrum entropy value of the current infrasound signal with the warning threshold; A determination module, configured to calculate the peak frequency of the current infrasound signal when the power spectrum entropy value of the current infrasound signal is less than the warning threshold, and determine that the current infrasound signal is a debris flow precursor signal when the peak frequency is within a set frequency range; The warning signal output module is used to output a warning signal according to the debris flow precursor signal.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the debris flow warning method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the debris flow warning method described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the debris flow warning method described in any one of claims 1 to 6 is implemented.

Images