Debris flow ground vibration signal identification method and system

Abstract

The invention relates to the field of signal identification, in particular to a debris flow ground vibration signal identification method and system, and the system comprises a precipitation monitoring module which is used for automatically sending early warning information when the daily precipitation reaches a precipitation threshold value, and a signal processing module which is used for collecting and preprocessing ground vibration signals corresponding to monitoring stations, the disaster early warning module is used for learning the ground vibration pre-data to generate a corresponding probability prediction map and issuing a cooperative instruction when a probability value is greater than a probability threshold value, and the cooperative identification module is used for capturing a secondary vibration signal of a monitoring station and judging whether the secondary vibration signal is abnormal or not according to a frequency threshold value. The debris flow early warning and protection method adopts corresponding protection measures, combines machine learning, recognizes the possibility of debris flow occurrence in advance, and automatically adopts protection measures such as road blocking, thereby reducing disaster loss, realizing early warning and protection of debris flow, and ensuring personnel and property safety.

Description

Technical Field

[0001] The present invention relates to the field of signal recognition, and in particular, to a method and system for recognizing debris flow ground vibration signals. Background Art

[0002] Traditional debris flow monitoring technologies mainly rely on means such as rain gauges, mud level gauges, video monitoring, etc. However, these methods have certain limitations. For example, rain gauges can only provide rainfall information and cannot directly monitor the occurrence of debris flows; video monitoring has poor effects at night or under bad weather conditions. In addition, although microwave radars based on the Doppler principle have the ability to monitor all-weather, they are prone to false alarms due to environmental factors such as the movement of wind and grass, falling rocks, and water level changes.

[0003] In recent years, with the development of artificial intelligence and deep learning technologies, researchers have begun to explore the use of machine learning models to improve the accuracy and reliability of debris flow monitoring.

[0004] Chinese Patent Application Publication No.: CN118606867A, discloses a debris flow prediction method based on hybrid machine learning, which relates to the field of debris flow early warning technology. By constructing an index system, determining the indexes for debris flow prediction, standardizing the index values in the index system, the standardization includes normalization and removing outliers, constructing a support vector classification model, using the AOVA algorithm to optimize hyperparameters, obtaining a hybrid machine learning model, constructing a training data set, and training the hybrid machine learning model to obtain a debris flow prediction model, and using the debris flow prediction model to predict the probability of debris flow occurrence, which solves the problem of difficult debris flow prediction in the prior art.

[0005] However, the prior art has the following problems: In the prior art, the signal analysis method has not been adaptively changed according to environmental factors, and the accuracy of abnormal determination of vibration signals is not high, and the reliability of early identification of debris flow occurrence is poor. Summary of the Invention

[0006] Therefore, the present invention provides a method and system for recognizing debris flow ground vibration signals to overcome the problems in the prior art that the signal analysis method has not been adaptively changed according to environmental factors, the accuracy of abnormal determination of vibration signals is not high, and the reliability of early identification of debris flow occurrence is poor.

[0007] To achieve the above object, the present invention provides a system for recognizing debris flow ground vibration signals, including: A precipitation monitoring module, which is used to judge whether the daily precipitation is abnormal according to a precipitation threshold, and automatically send a warning message when the daily precipitation reaches the precipitation threshold; A signal processing module, which is connected to the precipitation monitoring module, is used to collect the ground vibration signal corresponding to the monitoring station when receiving the early warning information, and preprocess the ground vibration signal to generate ground vibration pre-data; A disaster early warning module, which is connected to the signal processing module, is used to learn the ground vibration pre-data according to the occurrence probability prediction model to generate a corresponding probability prediction map, and determine whether to issue a collaboration instruction according to the comparison result between the probability value extracted from the probability prediction map and the probability threshold. Among them, The occurrence probability prediction model is generated by training a ground vibration training set; The ground vibration training set is formed by the ground vibration pre-data; The probability threshold is the critical value of the debris flow occurrence probability at the monitoring station; A collaboration recognition module, which is connected to the disaster early warning module, is used to receive the collaboration instruction, and determine whether to take corresponding protection measures according to the determination result of whether the secondary vibration signal captured by the monitoring station is abnormal according to the frequency threshold. Among them, The secondary vibration signal is a low-frequency vibration signal generated by the mixed movement of fluid and solid at the monitoring station.

[0008] Further, the precipitation monitoring module includes: A weighing sensor cluster, which is used to calculate the daily precipitation by measuring the weight of the collected rainwater; A precipitation comparator, which is connected to the weighing sensor, is used to compare the daily precipitation with the precipitation threshold; An early warning device, which is connected to the precipitation comparator, is used to send an early warning information to the signal processing module when the daily precipitation is greater than the precipitation threshold. Among them, the early warning information includes the daily precipitation value, the specific location of the monitoring station, and the sending time; The precipitation threshold is related to the rainfall condition and / or soil water content of the monitoring station.

[0009] Further, the signal processing module includes: A timer, which is used to collect the ground vibration signal corresponding to the monitoring station at preset time intervals and sort the ground vibration signals according to the collection time; A signal collector, which is connected to the timer, is used to collect the ground vibration signal of the monitoring station; A preprocessor, which is connected to the signal collector, is used to preprocess the ground vibration signal, select several index features, and generate corresponding ground vibration pre-data; A trainer, which is connected to the pre-processor and is used to train a corresponding occurrence probability prediction model by using the ground vibration training set; Among them, the index features include the periodicity, duration, and peak value of the ground vibration signal.

[0010] Further, the disaster warning module includes: A learner, which is used to learn the ground vibration pre-data according to the occurrence probability prediction model and generate a corresponding probability prediction map; A probability comparator, which is connected to the learner and is used to extract the probability value in the probability prediction map and compare the probability value with the probability threshold; An instruction issuing module, which is connected to the probability comparator and is used to issue a collaboration instruction to the collaborative recognition module when the probability value is greater than the probability threshold.

[0011] Further, the collaborative recognition module includes: An auxiliary signal collector, which is used to capture the secondary vibration signal of the monitoring site and obtain the corresponding secondary vibration frequency; A frequency comparator, which is connected to the auxiliary signal collector and is used to compare the secondary vibration frequency with the frequency threshold, form a corresponding comparison result, and take corresponding protective measures according to the comparison result.

[0012] Further, taking the geometric center of the monitoring site as a reference, load cells are arranged at random integer multiples of a preset spacing to form a test base station group; Among them, the preset spacing is related to the distance between the monitoring site and / or the rainfall condition of the monitoring site.

[0013] Further, the pre-processor pre-processes the ground vibration signal to generate the corresponding ground vibration pre-data, and the pre-processing includes: Based on several index features of the ground vibration signal, the ground vibration signal is standardized to form corresponding ground vibration pre-data; The standardization process is to segment the ground vibration signal at a standard sampling rate; The standard sampling rate is the sampling rate that the occurrence probability prediction model can recognize, and for a single sampling, its corresponding standard sampling rate is a single sampling rate.

[0014] Further, when the probability value is less than the probability threshold, the collaborative recognition module sends a re-identification instruction to the signal processing module, and the signal processing module re-collects the ground vibration signal corresponding to the monitoring site and performs disaster judgment.

[0015] Further, when the probability value is greater than the probability threshold, the collaborative recognition module sends a blocking instruction to block the road corresponding to the monitoring site.

[0016] On the other hand, the present invention provides a method for identifying debris flow ground vibration signals, including: The trainer iteratively optimizes the occurrence probability prediction model until the loss value of the occurrence probability prediction model converges to a preset loss value; Testing the current occurrence probability prediction model with the ground vibration training set; When the test result meets the preset test result, save the learning parameters of the occurrence probability prediction model.

[0017] Compared with the prior art, the present invention sets up a precipitation monitoring module to automatically send a warning message when the daily precipitation reaches the precipitation threshold, a signal processing module to collect and preprocess the ground vibration signal corresponding to the monitoring site to generate pre-ground vibration data, a disaster warning module to learn the pre-ground vibration data to generate a corresponding probability prediction map, and issue a collaborative instruction when the probability value is greater than the probability threshold, and a collaborative recognition module to capture the secondary vibration signal of the monitoring site, judge whether the secondary vibration signal is abnormal according to the frequency threshold, and take corresponding protection measures. Combining machine learning, it can identify the possibility of debris flow in advance, and reduce disaster losses by automatically taking protection measures such as blocking roads, realizing early warning and protection of debris flow, and ensuring the safety of people and property.

[0018] Further, by setting up a weighing sensor cluster to measure the weight of the collected rainwater, calculating the daily precipitation, comparing the daily precipitation with the precipitation threshold, and sending a warning message to the signal processing module, by monitoring the change of precipitation, it can detect abnormal rainfall conditions that may cause debris flow in advance, and timely notify the signal processing module to start subsequent ground vibration signal monitoring and analysis. This modular design enables the system to respond to debris flow risks hierarchically and stage by stage, improving the timeliness and accuracy of early warning.

[0019] Further, by setting up a signal processing module to collect, preprocess and analyze the ground vibration signal, it provides a scientific basis for the disaster warning module. This module can extract key features from complex ground vibration signals and learn the laws of debris flow occurrence through model training, thus providing technical support for the early warning of debris flow.

[0020] Further, by setting up a disaster warning module, through the collaborative work of a learner, a probability comparator and an instruction issuing module, it converts the ground vibration signal into a quantitative assessment of the probability of debris flow occurrence, and issues a collaborative instruction according to the risk level. This module not only improves the early warning ability of the system, but also realizes multi-dimensional debris flow risk monitoring and protection through linkage with other modules.

[0021] Furthermore, by setting up a collaborative recognition module to capture secondary vibration signals and analyze their frequency characteristics, further confirm the prediction results of the disaster warning module, and take practical protective measures, it can effectively distinguish debris flow activities from other interference signals, ensuring that the early warning and protection functions of the system can be accurately executed, greatly improving the reliability and practicality of the system, and providing strong support for the early warning and emergency response of debris flow disasters.

[0022] Furthermore, by converting the collected ground vibration signals into ground vibration pre-data suitable for subsequent analysis and model input, the preprocessing ensures the consistency and high quality of the data, providing a reliable data basis for the training and prediction of the occurrence probability prediction model.

[0023] Furthermore, by taking corresponding actions according to the judgment results of the disaster warning module, using the "re-recognition instruction" or "blocking instruction" to dynamically adjust the response strategy of the system, ensuring the most appropriate measures are taken at different risk levels, not only improving the flexibility and reliability of the system, but also providing strong support for the early warning and emergency response of debris flow disasters.

[0024] Furthermore, by optimizing the occurrence probability prediction model, it ensures that the system can accurately predict the occurrence probability of debris flow, not only improving the early warning ability of the system, but also ensuring the reliability and stability of the model in practical applications through a strict testing and parameter preservation mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic structural diagram of the debris flow ground vibration signal recognition system according to an embodiment of the present invention; Figure 2 It is a schematic structural diagram of the precipitation monitoring module according to an embodiment of the present invention; Figure 3 It is a schematic structural diagram of the signal processing module according to an embodiment of the present invention; Figure 4 It is a flowchart of the debris flow ground vibration signal recognition method according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] In order to make the objectives and advantages of the present invention clearer, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0027] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0028] It should be noted that in the description of the present invention, the terms indicating the direction or positional relationship such as "upper", "lower", "left", "right", "inner", "outer", etc. are based on the direction or positional relationship shown in the drawings. This is only for convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention.

[0029] In addition, it should also be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0030] Please refer to Figure 1 as shown, which is a schematic structural diagram of the debris flow ground vibration signal recognition system according to an embodiment of the present invention, including: A precipitation monitoring module, which is used to judge whether the daily precipitation is abnormal according to a precipitation threshold, and automatically send a warning message when the daily precipitation reaches the precipitation threshold; A signal processing module, which is connected to the precipitation monitoring module, and is used to collect the ground vibration signal corresponding to the monitoring station when receiving the warning message, and preprocess the ground vibration signal to generate ground vibration pre-data; A disaster warning module, which is connected to the signal processing module, and is used to learn the ground vibration pre-data according to a probability prediction model to generate a corresponding probability prediction map, and, determine whether to issue a cooperation instruction according to the comparison result between the probability value extracted from the probability prediction map and a probability threshold, where The probability prediction model is generated by training a ground vibration training set; The ground vibration training set is formed by ground vibration pre-data; The probability threshold is the critical value of the probability of debris flow occurring at the monitoring station; A cooperation recognition module, which is connected to the disaster warning module, and is used to receive the cooperation instruction, and, determine whether to take corresponding protection measures according to the judgment result of whether the secondary vibration signal captured by the monitoring station is abnormal according to a frequency threshold, where The secondary vibration signal is a low-frequency vibration signal generated by the mixed movement of fluid and solid at the monitoring station.

[0031] By setting up a precipitation monitoring module to automatically send warning messages when the daily precipitation reaches the precipitation threshold, a signal processing module to collect and preprocess the ground vibration signals corresponding to the monitoring stations to generate pre-ground vibration data, a disaster warning module to learn the pre-ground vibration data to generate corresponding probability prediction maps and issue collaborative instructions when the probability value is greater than the probability threshold, and a collaborative identification module to capture the secondary vibration signals of the monitoring stations, judge whether the secondary vibration signals are abnormal according to the frequency threshold, and take corresponding protective measures. Combining machine learning, it can identify the possibility of debris flow in advance and reduce disaster losses by automatically taking protective measures such as blocking roads, thus realizing the early warning and protection of debris flow and ensuring the safety of personnel and property.

[0032] Please refer to Figure 2 as shown, which is a schematic structural diagram of the precipitation monitoring module of the embodiment of the present invention, including: A weighing sensor cluster, which is used to calculate the daily precipitation by measuring the weight of the collected rainwater. A precipitation comparator, which is connected to the weighing sensor and used to compare the daily precipitation with the precipitation threshold. An alarm, which is connected to the precipitation comparator and used to send a warning message to the signal processing module when the daily precipitation is greater than the precipitation threshold. Among them, the warning message includes the daily precipitation value, the specific location of the monitoring station, and the sending time. The precipitation threshold is related to the rainfall conditions and / or soil moisture content of the monitoring station.

[0033] In specific implementation, the weighing sensor cluster is used to measure the weight of rainwater in real time. The rainwater flows into the container above the weighing sensor through a collection device (such as a funnel or a water tank). The weighing sensor converts the rainwater weight into an electrical signal and calculates the daily precipitation through a preset conversion relationship (such as the corresponding relationship between weight and precipitation).

[0034] The setting of the precipitation threshold can be flexibly adjusted according to the climate conditions and geological environment of different regions.

[0035] When the precipitation comparator judges that the daily precipitation exceeds the precipitation threshold, the alarm is triggered to generate a corresponding warning message, and the warning message is sent to the signal processing module through wireless network or wired transmission.

[0036] Content of the warning message: Daily precipitation value: Provide specific rainfall data for subsequent analysis.

[0037] Specific location of the monitoring station: Quickly locate the risk area through GPS coordinates or geographical identifiers.

[0038] Issuing time: Records the time when the early warning information is generated for easy traceability and response.

[0039] By setting up a cluster of weighing sensors to measure the weight of the rainwater collected, the daily precipitation is calculated, and the daily precipitation is compared with the precipitation threshold to send early warning information to the signal processing module. By monitoring the change of precipitation, the abnormal rainfall conditions that may lead to debris flow can be detected in advance, and the signal processing module is notified in time to start the subsequent monitoring and analysis of ground vibration signals. This modular design enables the system to respond to debris flow risks in a hierarchical and phased manner, improving the timeliness and accuracy of early warning.

[0040] Please refer to Figure 3 as shown, which is a schematic structural diagram of the signal processing module of the embodiment of the present invention, including: A timer for collecting ground vibration signals corresponding to the monitoring site at preset time intervals and sorting the ground vibration signals according to the collection time; A signal collector connected to the timer for collecting the ground vibration signals of the monitoring site; A pre-processor connected to the signal collector for pre-processing the ground vibration signals, selecting several index features, and generating corresponding ground vibration pre-data; A trainer connected to the pre-processor for training a corresponding occurrence probability prediction model using a ground vibration training set; Among them, the index features include the periodicity, duration, and peak value of the ground vibration signal.

[0041] In a specific implementation, the timer triggers the signal collector according to a preset time interval (such as every minute, every hour, etc.). Each collected ground vibration signal will be marked with the specific collection time and stored in chronological order for subsequent analysis and processing.

[0042] The signal collector is connected to the timer and starts collection according to the trigger signal of the timer. It is usually equipped with high-precision sensors (such as seismographs, accelerometers, or strain gauges) for detecting ground vibration signals. The collected ground vibration signals are output in digital or analog form and transmitted to the pre-processor.

[0043] Selection of index features: Periodicity: The periodicity feature of the ground vibration signal reflects the regularity of the vibration. The periodic change is related to the flow characteristics of the debris flow, such as the pulsed movement of the debris flow.

[0044] Duration: The duration of the ground vibration signal reflects the persistence of the vibration. A longer vibration duration implies a larger scale or longer movement time of the debris flow.

[0045] Peak: The peak of the ground vibration signal reflects the intensity of the vibration. A higher peak means a greater impact force of the debris flow, which is an important indicator for judging the danger of the debris flow.

[0046] By setting up a signal processing module, the ground vibration signal is collected, preprocessed and analyzed to provide a scientific basis for the disaster warning module. This module can extract key features from complex ground vibration signals and learn the laws of debris flow occurrence through model training, so as to provide technical support for the early warning of debris flow.

[0047] Specifically, the disaster warning module includes: A learner, which is used to learn the ground vibration pre-data according to the occurrence probability prediction model and generate a corresponding probability prediction map; A probability comparator, which is connected to the learner, used to extract the probability value in the probability prediction map and compare the probability value with the probability threshold; An instruction issuing module, which is connected to the probability comparator, used to issue a collaboration instruction to the collaborative recognition module when the probability value is greater than the probability threshold.

[0048] In a specific implementation, the learner receives the ground vibration pre-data generated by the signal processing module, analyzes the ground vibration pre-data using a pre-trained occurrence probability prediction model (such as a classifier or regression model based on machine learning), calculates the probability of debris flow occurrence according to the input ground vibration characteristics (such as periodicity, duration, peak, etc.), and visualizes the calculated probability result in the form of a map for subsequent analysis and comparison.

[0049] The probability threshold is a critical value set according to the historical data and geological conditions of the monitoring site, used to judge the danger level of debris flow occurrence. By comparing the probability value with the probability threshold, it can quickly judge whether the current ground vibration signal indicates a high risk of debris flow, provide a decision-making basis for the instruction issuing module, and decide whether further warning or protection measures need to be taken. Among them, the probability prediction map converts complex ground vibration signals into intuitive probability values.

[0050] By setting up the disaster warning module, through the collaborative work of the learner, probability comparator and instruction issuing module, the ground vibration signal is converted into a quantitative assessment of the probability of debris flow occurrence, and a collaboration instruction is issued according to the risk level. This module not only improves the warning ability of the system, but also realizes multi-dimensional debris flow risk monitoring and protection through linkage with other modules.

[0051] Specifically, the collaborative recognition module includes: An auxiliary signal collector, which is used to capture the secondary vibration signal of the monitoring site and obtain the corresponding secondary vibration frequency; A frequency comparator, which is connected to an auxiliary signal collector, is used to compare the secondary vibration frequency with a frequency threshold, form a corresponding comparison result, and take corresponding protection measures according to the comparison result.

[0052] In specific implementation, the secondary vibration signal refers to the low-frequency vibration signal generated by the mixed movement of the fluid (mud) and solid (stones) in the debris flow.

[0053] The auxiliary signal collector captures the secondary vibration signal through highly sensitive sensors (such as low-frequency accelerometers or seismographs), and the collected secondary vibration signal undergoes Fourier transform or other spectrum analysis methods to extract the secondary vibration frequency.

[0054] The preset frequency threshold is set according to the typical frequency characteristics of the debris flow, usually in the low-frequency range (such as 1 - 5 Hz).

[0055] If the secondary vibration frequency is higher than the frequency threshold, the frequency comparator triggers protection measures, such as: sending an alarm message to relevant departments or personnel, activating protection devices (such as debris flow interception nets, sound and light alarms, etc.), and automatically notifying downstream areas to take evacuation measures.

[0056] By setting a collaborative recognition module, capturing the secondary vibration signal and analyzing its frequency characteristics, further confirming the prediction result of the disaster warning module, and taking actual protection measures, it can effectively distinguish debris flow activities from other interference signals, ensure that the warning and protection functions of the system can be accurately executed, greatly improve the reliability and practicality of the system, and provide strong support for the early warning and emergency response of debris flow disasters.

[0057] Specifically, taking the geometric center of the monitoring site as a reference, load cells are arranged at random integer multiples of a preset spacing to form a test base station group; Among them, the preset spacing is related to the distance between the monitoring sites and / or the rainfall conditions of the monitoring sites.

[0058] Specifically, the preprocessor preprocesses the ground vibration signal to generate corresponding ground vibration pre-data. The preprocessing includes: Based on several index characteristics of the ground vibration signal, standardizing the ground vibration signal to form corresponding ground vibration pre-data; The standardization process is to segment the ground vibration signal at the standard sampling rate; The standard sampling rate is the sampling rate that the occurrence probability prediction model can recognize, and for a single sampling, its corresponding standard sampling rate is a single sampling rate.

[0059] Preferably, setting the sampling rate to 170 per second yields good learning results for the ground vibration pre-data, and for the system described in this application, it has the optimal analysis effect. An appropriate sampling rate can ensure the accuracy and real-time nature of the data, which is crucial for the training and prediction of the occurrence probability prediction model. An overly high sampling rate may lead to data redundancy and waste of computing resources, while an overly low sampling rate may result in information loss.

[0060] By converting the collected ground vibration signals into ground vibration pre-data suitable for subsequent analysis and model input, preprocessing ensures the consistency and high quality of the data, providing a reliable data basis for the training and prediction of the occurrence probability prediction model.

[0061] Specifically, when the probability value is less than the probability threshold, the collaborative recognition module sends a re-identification instruction to the signal processing module, and the signal processing module collects the ground vibration signals corresponding to the monitoring site again and conducts disaster judgment.

[0062] Specifically, when the probability value is greater than the probability threshold, the collaborative recognition module sends a blocking instruction to block the road corresponding to the monitoring site.

[0063] By taking corresponding actions according to the judgment results of the disaster warning module, using the "re-identification instruction" or "blocking instruction", dynamically adjusting the response strategy of the system, ensuring the most appropriate measures are taken under different risk levels, not only improving the flexibility and reliability of the system, but also providing strong support for the early warning and emergency response of debris flow disasters.

[0064] Please refer to Figure 4 as shown, which is a flowchart of the method for identifying ground vibration signals of debris flow in an embodiment of the present invention, including: Step S1, the trainer iteratively optimizes the occurrence probability prediction model until the loss value of the occurrence probability prediction model converges to a preset loss value; Step S2, testing the current occurrence probability prediction model with the ground vibration training set; Step S3, when the test result meets the preset test result, saving the learning parameters of the occurrence probability prediction model.

[0065] In a specific implementation, the trainer uses machine learning algorithms (such as gradient descent, stochastic gradient descent, etc.) to train the occurrence probability prediction model.

[0066] The loss value (Loss) is an index to measure the difference between the model prediction result and the real data. Common loss functions include mean square error (MSE), cross-entropy loss, etc.

[0067] Convergence condition: During the training process, the loss value of the model gradually decreases as the number of iterations increases. When the loss value converges to a preset loss value, it indicates that the model has achieved a good fitting effect, and the training process can be stopped.

[0068] The preset test result is the minimum requirement for the model performance. For example, the accuracy rate needs to reach over 90%.

[0069] By optimizing the occurrence probability prediction model, it is ensured that the system can accurately predict the occurrence probability of debris flow. This not only improves the early warning ability of the system but also ensures the reliability and stability of the model in practical applications through a strict testing and parameter preservation mechanism.

[0070] So far, the technical solution of the present invention has been described in combination with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or replacements to the relevant technical features, and the technical solutions after these changes or replacements will all fall within the protection scope of the present invention.

[0071] The above are only the preferred embodiments of the present invention and are not used to limit the present invention; for those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A debris flow ground vibration signal recognition system, characterized in that Including: A precipitation monitoring module, which is used to judge whether the daily precipitation is abnormal according to a precipitation threshold, and automatically send a warning message when the daily precipitation reaches the precipitation threshold; A signal processing module, which is connected to the precipitation monitoring module, and is used to collect the ground vibration signal corresponding to the monitoring station when receiving the warning message, and preprocess the ground vibration signal to generate preprocessed ground vibration data; A disaster warning module, which is connected to the signal processing module, and is used to learn the preprocessed ground vibration data according to a probability prediction model to generate a corresponding probability prediction map, and, determine whether to issue a cooperation instruction according to the comparison result between the probability value extracted from the probability prediction map and a probability threshold, where The probability prediction model is generated by training a ground vibration training set; The ground vibration training set is formed by the preprocessed ground vibration data; The probability threshold is the critical value of the probability of debris flow occurring at the monitoring station; A cooperation recognition module, which is connected to the disaster warning module, and is used to receive the cooperation instruction, and, determine whether to take corresponding protection measures according to the determination result of whether the secondary vibration signal captured by the monitoring station is abnormal according to a frequency threshold, where The secondary vibration signal is a low-frequency vibration signal generated by the mixed movement of fluid and solid at the monitoring station.

2. The debris flow ground vibration signal recognition system according to claim 1, wherein The precipitation monitoring module includes: A weighing sensor cluster, which is used to calculate the daily precipitation by measuring the weight of the collected rainwater; A precipitation comparator, which is connected to the weighing sensor, and is used to compare the daily precipitation with the precipitation threshold; A warning device, which is connected to the precipitation comparator, and is used to send a warning message to the signal processing module when the daily precipitation is greater than the precipitation threshold, wherein, the warning message includes the daily precipitation value, the specific location of the monitoring station, and the sending time; The precipitation threshold is related to the rainfall condition and / or soil moisture content of the monitoring station.

3. The debris flow ground vibration signal recognition system according to claim 2, wherein The signal processing module includes: A timer, which is used to collect the ground vibration signal corresponding to the monitoring station at preset time intervals, and sort the ground vibration signals according to the collection time; A signal collector, which is connected to the timer, and is used to collect the ground vibration signal of the monitoring station; A preprocessor, which is connected to the signal collector, and is used to preprocess the ground vibration signal, select several index features, and generate corresponding preprocessed ground vibration data; A trainer, which is connected to the preprocessor, and is used to train a corresponding probability prediction model by using the ground vibration training set; wherein, the index features include the periodicity, duration and peak value of the ground vibration signal.

4. The debris flow ground vibration signal recognition system according to claim 3, wherein The disaster warning module includes: A learner, which is used to learn the preprocessed ground vibration data according to the probability prediction model, and generate a corresponding probability prediction map; A probability comparator, which is connected to the learner, and is used to extract the probability value from the probability prediction map, and compare the probability value with the probability threshold; The instruction issuing module, which is connected to the probability comparator, is used to issue a collaboration instruction to the collaborative recognition module when the probability value is greater than the probability threshold.

5. The debris flow ground vibration signal recognition system according to claim 4, wherein The collaborative recognition module includes: An auxiliary signal collector, which is used to capture the secondary vibration signal of the monitoring site and obtain the corresponding secondary vibration frequency; A frequency comparator, which is connected to the auxiliary signal collector, is used to compare the secondary vibration frequency with the frequency threshold to form a corresponding comparison result, and take corresponding protection measures according to the comparison result.

6. The debris flow ground vibration signal recognition system according to claim 5, characterized in that, Taking the geometric center of the monitoring site as a reference, load cells are arranged at random integer multiples of a preset spacing to form a test base station group; Wherein, the preset spacing is related to the distance between the monitoring sites and / or the rainfall conditions of the monitoring sites.

7. The debris flow ground vibration signal recognition system according to claim 6, wherein The pre-processor pre-processes the ground vibration signal to generate the corresponding ground vibration pre-data, and the pre-processing includes: Based on several index features of the ground vibration signal, the ground vibration signal is normalized to form the corresponding ground vibration pre-data; The normalization process is to divide the ground vibration signal at the standard sampling rate; The standard sampling rate is the sampling rate that the occurrence probability prediction model can recognize, and for a single sampling, its corresponding standard sampling rate is a single sampling rate.

8. The debris flow ground vibration signal recognition system according to claim 7, characterized in that, When the probability value is less than the probability threshold, the collaborative recognition module sends a re-identification instruction to the signal processing module, and the signal processing module re-collects the ground vibration signal corresponding to the monitoring site and performs disaster judgment.

9. The debris flow ground vibration signal recognition system according to claim 8, characterized in that, When the probability value is greater than the probability threshold, the collaborative recognition module sends a blockade instruction to blockade the road corresponding to the monitoring site.

10. A method for a debris flow ground vibration signal recognition system according to any one of claims 1-9, characterized in that, Including: The trainer iteratively optimizes the occurrence probability prediction model until the loss value of the occurrence probability prediction model converges to a preset loss value; Use the ground vibration training set to test the current occurrence probability prediction model; When the test result meets the preset test result, save the learning parameters of the occurrence probability prediction model.

Images