A method and system for calibrating landslide monitoring sensor signals

Abstract

This invention relates to the field of landslide monitoring technology, specifically disclosing a method and system for calibrating landslide monitoring sensor signals. The method includes: continuously acquiring monitoring signals from multiple sensors; analyzing the monitoring signals to obtain the statistical characteristics of each sensor, the consistency of changes with physically associated sensors, and the degree of deviation from historical signals; calculating the reliability score of each sensor based on the statistical characteristics, consistency of changes, and degree of deviation; analyzing and obtaining error patterns when the reliability score is lower than a preset threshold; extracting calibration schemes from a preset calibration scheme library based on the error patterns; and calibrating the monitoring signals of the corresponding sensors according to the calibration schemes. This method achieves automated, data-driven calibration of landslide monitoring sensor signals, overcomes the limitations of traditional manual calibration in field environments, and improves the accuracy of monitoring data and the reliability of the system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of landslide monitoring, in particular to a landslide monitoring sensor signal calibration method and system. BACKGROUND

[0002] Landslide monitoring systems are widely deployed in potential landslide areas, continuously collecting geological and environmental parameter data through surface displacement, deep displacement, pore water pressure, and rainfall sensors, analyzing landslide stability and issuing warnings to protect downstream residents and infrastructure. With technological advancements, the system is moving towards higher automation, intelligence, and real-time, improving data collection efficiency and warning reliability, and adapting to long-term monitoring needs in complex field environments.

[0003] However, long-term exposure of sensors to harsh field environments leads to signal errors such as zero drift and sensitivity decay, causing sensor readings to deviate from actual physical quantities, which may cause false alarms or missed alarms. Traditional calibration methods rely on manual or external equipment, which is not practical and costly in field scenarios. In addition, error patterns are diverse, such as zero drift and sensitivity decay, which need to be accurately diagnosed to apply appropriate calibration solutions.

[0004] Currently, there is no effective technical solution to the above problems. SUMMARY

[0005] The purpose of the present application is to provide a landslide monitoring sensor signal calibration method and system to realize automatic and data-driven calibration of landslide monitoring sensor signals.

[0006] In a first aspect, the present application provides a landslide monitoring sensor signal calibration method applied to sensors of a landslide state monitoring system, comprising the following steps:

[0007] S1, continuously acquiring monitoring signals of multiple sensors;

[0008] S2, analyzing statistical characteristics of each sensor, consistency of changes of sensors associated with physics, and deviation degree from historical signals according to the monitoring signals;

[0009] S3, calculating a credibility score of each sensor according to the statistical characteristics, consistency of changes, and deviation degree;

[0010] S4, when the credibility score is lower than a preset score threshold, analyzing an error pattern according to one or more of the statistical characteristics, the consistency of changes, and the deviation degree;

[0011] S5, extracting a calibration scheme from a preset calibration scheme library according to the error pattern;

[0012] S6, calibrating the monitoring signal of the corresponding sensor according to the calibration scheme.

[0013] The method of the present application realizes the automation and data-driven calibration of landslide monitoring sensor signals, overcomes the limitations of traditional manual calibration in the field environment, improves the accuracy of monitoring data and the reliability of the system, reduces the demand and cost of manual calibration, and enhances the early warning capability of the landslide monitoring system.

[0014] The landslide monitoring sensor signal calibration method, wherein step S2 comprises:

[0015] S21, calculating the mean, variance and fluctuation range of each sensor as statistical characteristics according to the monitoring signal;

[0016] S22, calculating the Pearson correlation coefficient between the sensor pairs with physical correlation as the change consistency of the corresponding sensors;

[0017] S23, obtaining the deviation degree according to the difference between the historical signals of the same type of sensors under the same environmental conditions or the same time conditions and the monitoring signals.

[0018] The above steps provide a specific and operable method to quantify the statistical characteristics, change consistency and deviation degree of landslide monitoring sensor signals, improve the accuracy of information acquisition, and provide a reliable data basis for subsequent reliability evaluation and error mode identification.

[0019] The landslide monitoring sensor signal calibration method, wherein step S23 comprises:

[0020] S231, extracting the historical signals of the same type of sensors under the same environmental conditions or the same time conditions, and calculating the mean and variance of the historical signals;

[0021] S232, calculating the mean difference and variance difference between the historical signals and the monitoring signals based on the statistical characteristics;

[0022] S233, calculating the deviation degree according to the mean difference and the variance difference.

[0023] The landslide monitoring sensor signal calibration method, wherein step S3 comprises:

[0024] S31, obtaining the activity level of the landslide at the previous time;

[0025] S32, extracting a weight group from a pre-set weight database according to the activity level and the type of the sensor;

[0026] S33, calculate a credibility score according to the weight group, the statistical characteristics, the change consistency and the deviation degree.

[0027] The landslide monitoring sensor signal calibration method, wherein the determination process of the activity level of the landslide comprises:

[0028] A1, extract the displacement amount, rainfall and groundwater level change amplitude of the corresponding landslide in a preset time window;

[0029] A2, according to the displacement amount, rainfall and groundwater level change amplitude of the landslide, query the preset activity level discrimination rule library to obtain the activity level.

[0030] The landslide monitoring sensor signal calibration method, wherein step S4 comprises:

[0031] S41, when the credibility score is lower than the preset score threshold, according to the statistical characteristics, the change consistency and the deviation degree, respectively extract the first deviation state feature, the second deviation state feature and the third deviation state feature;

[0032] S42, splice the first deviation state feature, the second deviation state feature and the third deviation state feature to obtain a fusion deviation feature vector;

[0033] S43, input the fusion deviation feature vector into a preset error mode recognition model to generate the error mode.

[0034] The landslide monitoring sensor signal calibration method, wherein the error mode comprises: zero drift, sensitivity attenuation, intermittent failure and periodic interference.

[0035] The landslide monitoring sensor signal calibration method, wherein step S5 comprises:

[0036] S51, according to the activity level of the landslide at the last moment, the error mode and the type of the corresponding sensor, extract the corresponding calibration scheme from the preset calibration scheme library.

[0037] The landslide monitoring sensor signal calibration method, wherein the calibration scheme comprises a calibration means and a calibration parameter type, and step S6 comprises:

[0038] S61, based on the calibration parameter type, determine the calibration parameter according to the monitoring signal of the sensor to be calibrated and the historical signal of the same type sensor under the same environmental condition or the same time condition;

[0039] S62, based on the calibration means and the calibration parameter, update the configuration information of the corresponding sensor to calibrate the monitoring signal of the corresponding sensor.

[0040] In a second aspect, the application further provides a landslide monitoring sensor signal calibration system, which is applied to a sensor of a landslide state monitoring system, and the system comprises:

[0041] an acquisition module, configured to continuously acquire monitoring signals of a plurality of sensors;

[0042] a characteristic information acquisition module, configured to analyze and acquire statistical characteristics of each sensor, consistency of changes of physically associated sensors, and deviation degrees from historical signals according to the monitoring signals;

[0043] a scoring module, configured to calculate a credibility score of each sensor according to the statistical characteristics, the consistency of changes, and the deviation degrees;

[0044] an error analysis module, configured to analyze and acquire an error mode according to one or more of the statistical characteristics, the consistency of changes, and the deviation degrees when the credibility score is lower than a preset score threshold;

[0045] a calibration scheme determination module, configured to extract a calibration scheme from a preset calibration scheme library according to the error mode;

[0046] a calibration module, configured to calibrate the monitoring signals of the corresponding sensor according to the calibration scheme.

[0047] The system of the application realizes automatic and data-driven calibration of landslide monitoring sensor signals, overcomes the limitations of traditional manual calibration in a field environment, improves the accuracy of monitoring data and the reliability of the system, reduces the demand and cost of manual calibration, and enhances the early warning capability of the landslide monitoring system.

[0048] As can be seen from the above, the application provides a landslide monitoring sensor signal calibration method and system, wherein the method of the application analyzes the statistical characteristics of sensor signals, the consistency of changes of physically associated sensors, and the deviation degrees from historical signals, calculates the credibility score of the signals, and diagnoses a specific error mode based on these analysis results when the score is lower than a threshold, so as to extract and apply a corresponding calibration scheme from a preset calibration scheme library, thereby realizing automatic and data-driven calibration of landslide monitoring sensor signals, overcoming the limitations of traditional manual calibration in a field environment, improving the accuracy of monitoring data and the reliability of the system, reducing the demand and cost of manual calibration, and enhancing the early warning capability of the landslide monitoring system. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 A flowchart of a landslide monitoring sensor signal calibration method provided by an embodiment of the application.

[0050] Figure 2A structural schematic diagram of a landslide monitoring sensor signal calibration system provided by an embodiment of the present application.

[0051] The reference signs: 201, an acquisition module; 202, a characteristic information acquisition module; 203, a scoring module; 204, an error analysis module; 205, a calibration scheme determination module; 206, a calibration module. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0053] It should be noted that: similar reference signs and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms “first”, “second”, etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0054] In a first aspect, referring to Figure 1 Some embodiments of the present application provide a landslide monitoring sensor signal calibration method applied to sensors of a landslide state monitoring system, which comprises the following steps:

[0055] S1, continuously acquiring monitoring signals of multiple sensors;

[0056] S2, analyzing and acquiring statistical characteristics of each sensor, consistency of changes of sensors associated with physics, and deviation degree from historical signals according to the monitoring signals;

[0057] S3, calculating and acquiring a credibility score of each sensor according to the statistical characteristics, the consistency of changes, and the deviation degree;

[0058] S4, when the credibility score is lower than a preset score threshold, analyzing and acquiring an error mode according to one or more of the statistical characteristics, the consistency of changes, and the deviation degree;

[0059] S5, extracting a calibration scheme from a preset calibration scheme library according to the error mode;

[0060] S6. Calibrating the monitoring signal of the corresponding sensor according to the calibration scheme.

[0061] Specifically, continuously acquiring the monitoring signals of multiple sensors refers to uninterruptedly receiving real-time or quasi-real-time data streams from various monitoring devices such as ground displacement sensors, deep displacement sensors, pore water pressure sensors, and rainfall sensors, which can be achieved in a wired transmission, wireless transmission, or data logger regular collection manner, such as data transmission through an optical fiber network or a wireless network.

[0062] More specifically, the statistical property refers to the inherent numerical distribution characteristics of the sensor monitoring signal within a period of time, which can be achieved by calculating the mean, variance, standard deviation, fluctuation range, skewness, kurtosis, etc., such as calculating the average value of the signal and the dispersion degree of the signal value, which is mainly used to reflect the stability, dispersion degree, or distribution form of the signal itself.

[0063] More specifically, the consistency of change refers to the synchronization or correlation between the signal change trend of a sensor and the signal change trend of other sensors with physical correlation, which can be achieved by calculating the Pearson correlation coefficient, Spearman rank correlation coefficient, cross-correlation function, etc., such as calculating the correlation between the ground displacement and deep displacement signals, which is mainly used to determine whether the sensor signal is coordinated with other signals that change synchronously.

[0064] More specifically, the deviation degree refers to the difference between the sensor monitoring signal and the historical signal or expected pattern of the sensor under similar environmental or time conditions, which can be achieved by calculating the mean difference, variance difference, root mean square error, or pattern matching-based difference measure between the current signal and the historical signal, such as comparing the statistical distribution difference between the current rainfall signal and the historical same-period rainfall signal, which is mainly used to detect whether the signal deviates from its normal behavior range.

[0065] More specifically, the reliability score refers to a numerical value that quantitatively evaluates the reliability of the sensor monitoring signal, which can be achieved by weighted summation, fuzzy logic reasoning, or machine learning model output based on statistical properties, change consistency, and deviation degree, such as mapping the analysis results of each indicator to a score of 0 to 100 through a function, which is mainly used to comprehensively judge whether the sensor signal is reliable as a basis for whether it needs to be calibrated.

[0066] More specifically, the error mode refers to the specific error type or reason that causes the sensor signal to be unreliable, which can include zero drift, sensitivity decay, intermittent failure, periodic interference, data jump, data stickiness, etc., which is mainly used to diagnose the root cause of signal abnormalities in order to select a targeted calibration method.

[0067] More specifically, the calibration scheme library refers to a pre-stored set of signal correction methods designed for different error modes and sensor types, which can be implemented in the form of a database table, a lookup table, or a software structure containing different calibration algorithm modules, such as storing a bias addition algorithm for zero drift and a multiplication coefficient algorithm for sensitivity decay, which is mainly used to provide calibration methods matching the diagnosed error modes.

[0068] The method of the present application first continuously acquires a plurality of types of landslide monitoring sensor signals. Then, multi-dimensional analysis is performed on these signals to extract statistical characteristics reflecting the characteristics of the signals themselves, physical correlation consistency reflecting the mutual relationship between the signals, and deviation degree reflecting the difference between the signals and normal behavior. Based on these analysis results, the credibility score of each sensor signal is calculated, which comprehensively reflects the reliability of the signal. When the credibility score is lower than the preset threshold, it indicates that the signal may have errors, at which time the method further utilizes the statistical characteristics, change consistency, and deviation degree information obtained from previous analysis to identify the specific error mode of the signal, such as zero drift or sensitivity decay. Once the error mode is determined, the method of the present application will look up and extract the corresponding calibration scheme from the pre-set calibration scheme library according to the error mode. Finally, according to the extracted calibration scheme, the sensor with errors is calibrated to correct the signal value, so that the monitoring signal output by the sensor is closer to the true physical quantity. The entire process forms a closed loop, achieving automatic detection, diagnosis, and correction of sensor signal errors.

[0069] The method of the present application calculates the credibility score of the signal by continuously analyzing the statistical characteristics of the sensor signal, the change consistency of the sensor with physical correlation, and the deviation degree from the historical signal, and diagnoses the specific error mode based on these analysis results when the score is lower than the threshold, thereby extracting and applying the corresponding calibration scheme from the pre-set calibration scheme library, achieving automatic and data-driven calibration of landslide monitoring sensor signals, overcoming the limitations of traditional manual calibration in the field environment, improving the accuracy of monitoring data and the reliability of the system, reducing the demand and cost of manual calibration, and enhancing the early warning capability of the landslide monitoring system.

[0070] In some preferred embodiments, step S2 comprises:

[0071] S21, calculating the mean, variance, and fluctuation range of each sensor according to the monitoring signal as the statistical characteristics;

[0072] S22, calculating the Pearson correlation coefficient between the sensor pairs with physical correlation as the change consistency of the corresponding sensors;

[0073] S23, obtaining the deviation degree according to the difference between the historical signal and the monitoring signal of the same type sensor under the same environmental condition or the same time condition.

[0074] Specifically, the sensor pair with physical correlation refers to a combination of sensors whose monitored physical quantities have an inherent connection or mutual influence in the landslide monitoring scene, for example, displacement sensors at different depths, rainfall sensors, and pore water pressure sensors, etc., which can be identified by using pre-set sensor correlation rules or correlation thresholds determined based on historical data analysis.

[0075] More specifically, the same type sensor refers to a sensor that monitors the same type of physical quantity, for example, multiple ground surface displacement sensors or multiple pore water pressure sensors.

[0076] More specifically, the same environmental condition or the same time condition refers to the case where the current monitoring signal is collected under similar external environmental factors (such as rainfall, temperature, etc.) or in a similar time period (such as the same season, the same time point) in the historical data, which can be realized by establishing the association between environmental parameters or time index and historical monitoring data. The difference between the historical signal and the monitoring signal refers to the deviation degree of the current monitoring signal from the historical signal collected under similar conditions in terms of numerical value or statistical characteristics, which can be quantified by using various mathematical or statistical methods.

[0077] More specifically, step S21 quantifies the central tendency, dispersion, and variation amplitude of the signal by calculating the mean, variance, and fluctuation range of the monitoring signal of each sensor, thereby obtaining the statistical characteristics of the sensor. These statistical quantities provide a quantitative description of the basic behavior of the signal of a single sensor. Then, step S22 identifies the sensor pair with physical correlation and calculates the Pearson correlation coefficient between the monitoring signals of these sensor pairs. The Pearson correlation coefficient reflects the degree of linear correlation between two signals, thereby quantifying the consistency of changes between sensors, which helps to determine whether the sensor signal conforms to the expected physical linkage relationship. Finally, step S23 quantifies the deviation degree of the signal by comparing the current monitoring signal with the historical signal of the same type sensor under similar environmental or time conditions. This comparison takes into account the influence of external factors and time regularity, making the evaluation of the deviation degree more accurate. Through the analysis of these three dimensions, the scheme comprehensively and quantitatively obtains the core information required for evaluating the reliability of the sensor signal, providing a reliable data basis for subsequent reliability evaluation and error mode identification.

[0078] The above steps provide a specific and operable method to quantitatively obtain the statistical characteristics, change consistency, and deviation degree of the landslide monitoring sensor signal, improving the accuracy of information acquisition and providing a reliable data basis for subsequent reliability evaluation and error mode identification.

[0079] In some preferred embodiments, step S23 comprises:

[0080] S231, extract the historical signals of the same type of sensor under the same environmental conditions or the same time conditions, and calculate the mean and variance of the historical signals;

[0081] S232, based on the statistical characteristics, calculate the mean difference and variance difference between the historical signals and the monitoring signals;

[0082] S233, calculate the deviation degree according to the mean difference and the variance difference.

[0083] Specifically, the mean difference refers to the difference between the mean of the monitoring signal and the mean of the historical signal. The variance difference refers to the difference between the variance of the monitoring signal and the variance of the historical signal.

[0084] More specifically, in step S233, the deviation degree is positively correlated with the mean difference and the variance difference.

[0085] More specifically, the mean difference reflects the shift of the overall level of the signal, and the variance difference reflects the change of the volatility of the signal. By calculating these differences, the abstract "difference" is converted into a specific numerical value, laying a foundation for subsequent analysis. In this way, the difference in statistical characteristics is integrated into a quantitative deviation degree index, providing a more reliable input for the sensor reliability score, which helps to improve the accuracy of sensor signal calibration.

[0086] In some preferred embodiments, in step S233, the deviation degree is calculated in the following manner:

[0087] Deviation degree = a · |(μ c -μ h、 ) / μ c | + β · |(σ h ²-σ h ²) / σ h ²| (1)

[0088] Wherein, a and β are preset weight coefficients, which can be set according to the use requirements, and are preferably 0.5; μ , μ

[0001] σ ² and σ ² are the mean of the monitoring signal, the mean of the historical signal, the variance of the monitoring signal and the variance of the historical signal, respectively; wherein, μ

[0002] and σ ² are not 0.

[0089] In some preferred embodiments, step S3 comprises:

[0090] S31, obtain the activity level of the landslide at the previous time.

[0091] S32, extracting a weight set from a preset weight database according to the activity level and the type of the sensor;

[0092] S33, calculating a confidence score according to the weight set, the statistical characteristic, the change consistency and the deviation degree.

[0093] Specifically, the activity level of the landslide refers to a classification index reflecting the current activity state of the landslide. The preset weight database refers to a set of weight parameters used to calculate the confidence score, and the database pre-configures corresponding weight values according to different landslide activity levels and different sensor types. The weight set refers to a set of weight values extracted from the weight database and applicable to a specific landslide activity level and a specific sensor type, and the set contains weights for weighting the statistical characteristic, the change consistency and the deviation degree.

[0094] More specifically, the above steps provide a more fine and adaptive sensor signal confidence score calculation method. By considering the activity level of the landslide and the type of the sensor, the weights of the evaluation indexes are dynamically adjusted, thereby improving the accuracy of the confidence evaluation and providing a more reliable basis for subsequent error mode identification and calibration. First, the activity level of the landslide at the last time is obtained, which is to introduce the state of the landslide at the last time as a state reference benchmark. Under different activity levels, the normal fluctuation range and abnormal performance characteristics of the sensor signal are different, so the evaluation standard needs to be adjusted according to the current state of the landslide. Then, according to the obtained activity level and the type of the sensor to be evaluated, the corresponding weight set is extracted from the preset weight database. This weight database pre-stores the relative importance (i.e. weight) of the three indexes of statistical characteristic, change consistency and deviation degree in evaluating confidence for different types of sensors under different landslide activity levels. By extracting the weight set according to the activity level and the type of the sensor, it can be ensured that the weight used to calculate the confidence score is the most suitable in the current situation. For example, in the active period of the landslide, the change consistency of the displacement sensor may be more reflective of its confidence than the degree of deviation from the historical signal, and the weight of the change consistency will be higher. The introduction of this dynamic weight makes the confidence evaluation no longer static, but can adapt to the changes in the state of the landslide and the type of the sensor. Finally, according to the extracted weight set, the three indexes of statistical characteristic, change consistency and deviation degree obtained previously are weighted and calculated, thereby obtaining the final confidence score. Since the weight is dynamically determined according to the activity level of the landslide and the type of the sensor, the calculated confidence score can more accurately reflect the true reliability of the current sensor signal, effectively distinguish between real geological changes and sensor errors, and provide a more reliable basis for subsequent calibration decisions.

[0095] The method of the present application can dynamically adjust the credibility score calculation process according to the current activity level of the landslide and the type of sensor, so that the credibility evaluation process can fully utilize multi-dimensional signal features and adapt to complex monitoring environments and changes in landslide state by dynamically adjusting weights, thereby improving the accuracy and adaptability of credibility evaluation, and more effectively identifying errors in sensor signals, providing more reliable input for subsequent error pattern recognition and calibration.

[0096] In some preferred embodiments, the determination process of the activity level of the landslide comprises:

[0097] A1, extracting the displacement amount, rainfall and groundwater level change amplitude corresponding to the landslide in a preset time window;

[0098] A2, querying a preset activity level discrimination rule base according to the displacement amount, rainfall and groundwater level change amplitude of the landslide to obtain the activity level.

[0099] Specifically, the preset time window refers to a continuous time period for analyzing the recent activity state of the landslide, and its length can be set according to actual monitoring needs and landslide characteristics. The displacement amount refers to the overall or local movement distance of the landslide body within the preset time window, which can be obtained using data from surface displacement sensors or deep displacement sensors. Rainfall refers to the total amount of rainfall accumulated in the landslide area within the preset time window, which can be obtained using data from rain sensors. The groundwater level change amplitude refers to the maximum change range of the groundwater level in the landslide area within the preset time window, which can be obtained using data from pore water pressure sensors or groundwater level gauges. The activity level discrimination rule base refers to a series of rules, thresholds or models stored for determining the activity level of the landslide, which can be implemented using a rule set based on expert experience, a decision tree model trained based on historical monitoring data, or a lookup table, etc. The activity level refers to a classification representation of the current activity intensity of the landslide, which is preferably divided into three levels of stable, slow and active, representing the states of no activity, slow deformation or rapid deformation of the landslide body.

[0100] Specifically, the determination process first extracts, in step A1, displacement, rainfall and groundwater level change amplitude data related to the landslide to be evaluated within a preset time window from the monitoring system. These physical quantities are key factors affecting the stability and activity of landslides, and can directly or indirectly reflect the current state and activity trend of the landslide. By limiting the extraction within a preset time window, it is ensured that the analyzed data can reflect the recent behavior of the landslide, so that the determined activity level is timely and relevant. Then, in step A2, the displacement, rainfall and groundwater level change amplitude obtained in step A1 are taken as inputs to query the preset activity level discrimination rule library. The rules or models preset in the rule library determine the current activity level of the landslide based on these input parameters.

[0101] More specifically, the method of the present application achieves objective classification of the state of the landslide by querying the rule library, providing a clear basis for subsequent adjustment of sensor reliability evaluation weights according to the activity level of the landslide. This process provides the necessary input for weight selection based on the activity level, enabling the sensor reliability evaluation to better adapt to the characteristics of the landslide in different active states, thereby improving the accuracy of the evaluation.

[0102] In some preferred embodiments, step S4 comprises:

[0103] S41, when the reliability score is lower than the preset score threshold, extracting first deviation state features, second deviation state features and third deviation state features according to the statistical properties, change consistency and deviation degree, respectively;

[0104] S42, concatenating the first deviation state features, the second deviation state features and the third deviation state features to obtain a fusion deviation feature vector;

[0105] S43, inputting the fusion deviation feature vector into a preset error pattern recognition model to generate an error pattern.

[0106] Specifically, the first deviation state features refer to quantitative indicators that can reflect the abnormal state of the signal itself, extracted according to the statistical properties (such as mean, variance, fluctuation range, etc.) of the sensor monitoring signal. The second deviation state features refer to quantitative indicators that can reflect the abnormality of the cooperative relationship between sensors, extracted according to the change consistency (such as Pearson correlation coefficient) between the signal to be diagnosed and the physically related sensor signal. The third deviation state features refer to quantitative indicators that can reflect the abnormality of the signal relative to its normal historical behavior, extracted according to the deviation degree between the signal to be diagnosed and the historical signal of the same type of sensor. The fusion deviation feature vector is used to comprehensively describe the abnormal state of the sensor.

[0107] More specifically, the error mode recognition model is preferably a classification model built based on a support vector machine multi-classifier, which can receive the fusion deviation feature vector as input and output the corresponding sensor error mode category. It can be built with One-vs-Rest or One-vs-One strategy, and the model can be built and trained in the same way as the classifier, which will not be described here.

[0108] Specifically, when the credibility score of the sensor is lower than the preset threshold, indicating that the signal may have errors, the method enters the error mode diagnosis stage. First, according to the statistical characteristics, the consistency of changes and the degree of deviation calculated before, the first deviation state feature, the second deviation state feature and the third deviation state feature are extracted respectively. These deviation state features are further refined and quantified from the original characteristics, aiming to capture the specific patterns or abnormal signals of different error modes in these characteristics. By extracting respectively, it can ensure that the abnormal state of the signal is quantified comprehensively from different dimensions (statistical behavior of the signal itself, relationship with other related signals, comparison with historical behavior). Then, the first, second and third deviation state features extracted are spliced to form a fusion deviation feature vector. This fusion vector integrates abnormal information from different dimensions, providing a more comprehensive and discriminative feature representation, which helps to more accurately describe the error state of the current signal. Finally, the fusion deviation feature vector is input into a preset error mode recognition model built based on a support vector machine multi-classifier. This model is pre-trained and can learn and recognize the feature vector patterns corresponding to different error modes. By inputting the fusion feature vector into this model, the model can automatically classify and diagnose the error type of the current signal, thereby generating a specific error mode. This machine learning-based classification method utilizes the rules learned by the model from a large amount of historical data, enabling automatic, objective and high-precision recognition of error modes, providing a reliable basis for subsequent targeted calibration.

[0109] The above diagnosis process is executed after the credibility score calculation and before the calibration scheme extraction, which connects the judgment result of signal anomaly and provides specific error type information for the subsequent calibration step, making the entire calibration process more complete and effective, thereby improving the data reliability and early warning accuracy of the landslide monitoring system.

[0110] In some preferred embodiments, the error mode includes:

[0111] Zero drift, sensitivity decay, intermittent failure and periodic interference.

[0112] Specifically, zero-point drift refers to the phenomenon that the output signal of the sensor deviates from its theoretical zero-point value when the input physical quantity is zero, which can be manifested as an overall increase or decrease of the baseline of the output signal. Sensitivity decay refers to the phenomenon that the proportional relationship between the output signal of the sensor and the change in the input physical quantity decreases, which can be manifested as a decrease in the gain of the output signal of the sensor. Intermittent failure refers to the phenomenon that the sensor outputs abnormally or has no output for a period of time and then returns to normal, which can be manifested as an interruption or an outlier in the data stream. Among them, periodic interference refers to the phenomenon that a noise signal with a fixed or approximately fixed period is superimposed on the output signal of the sensor, which can be manifested as regular fluctuations in the output signal.

[0113] More specifically, specifically, the method of the present application provides a clear classification target for the error mode recognition model constructed by the multi-classifier based on support vector machine by clearly defining the type of error mode. The fusion feature vector is input into the preset multi-classifier based on support vector machine, and the classifier maps the feature vector to one of the four error modes defined in the present scheme according to its training result. For example, if the mode of the feature vector is similar to the feature mode labeled as "zero-point drift" in the training sample, the classifier outputs "zero-point drift" as the recognized error mode.

[0114] The method of the present application clearly defines the type of error mode, provides a clear classification target for the error mode recognition model, enables the model to be trained and applied, and improves the accuracy of error mode recognition. The specific error mode is identified, so that the corresponding calibration scheme can be extracted from the calibration scheme library subsequently. Different error modes require different calibration methods and parameters, and the error mode type is clearly defined, so that the calibration process can be calibrated according to the recognized error type. The accuracy and effectiveness of calibration are improved, and the reliability of data acquisition of the landslide monitoring system is ultimately improved, thereby enhancing the accuracy of early warning.

[0115] In some preferred embodiments, the first deviation state feature includes a mean drift amplitude and a variance fluctuation frequency; the second deviation state feature includes a reduction amplitude of the Pearson correlation coefficient; and the third deviation state feature includes a mean difference and a variance difference between the monitoring signal and the historical signal.

[0116] In particular, the mean shift amplitude refers to the degree of change of the mean value of the sensor signal over time, which can be achieved by calculating the amount of change or the rate of change of the mean value of the signal within a time window. The variance fluctuation frequency refers to the speed or periodicity of the variance of the sensor signal fluctuating within a period of time, which can be achieved by performing spectral analysis on the variance sequence of the signal or calculating the number of times the variance crosses a certain threshold within a unit of time. The reduction amplitude of the Pearson correlation coefficient refers to the degree of decline of the Pearson correlation coefficient between the monitoring signal and the signal of the physically associated sensor relative to the normal state or historical state, which can be achieved by calculating the difference between the current correlation coefficient and the historical average correlation coefficient or the normal correlation coefficient. The mean difference between the monitoring signal and the historical signal refers to the difference between the mean value of the current monitoring signal and the mean value of the historical signal under similar conditions, which can be achieved by directly calculating the difference between the mean value of the current signal and the mean value of the historical signal.

[0117] More specifically, the method of the present application specifically defines the specific content of the three deviation state features for constructing the fusion deviation feature vector, thereby making the error mode recognition process more specific and operable. The first deviation state feature is defined to include the mean shift amplitude and the variance fluctuation frequency, which means that the features extracted from the statistical characteristics of the sensor are the size of the change in the signal mean value and the speed of the variance fluctuation frequency, which can reflect the drift or noise characteristics of the signal, providing basic data for identifying error modes such as zero drift or periodic interference. The second deviation state feature is defined to include the reduction amplitude of the Pearson correlation coefficient, which means that the feature extracted from the consistency of the change between the sensor and the physically associated sensor is the degree of decline of the correlation between them, which helps to identify the relative abnormality between the sensor and other normally working sensors, which may indicate problems such as sensitivity decay or intermittent failure. The third deviation state feature is defined to include the mean difference and variance difference between the monitoring signal and the historical signal, which means that the feature extracted from the deviation degree between the monitoring signal and the historical signal is the difference between the mean value and the variance of the current signal and its historical performance under similar conditions, which can reflect whether the signal deviates from its normal behavior pattern, providing historical comparison basis for identifying various types of errors.

[0118] More specifically, the method of the present application provides a clear and specific input feature vector for the error mode recognition model based on the fusion deviation feature vector by explicitly defining the specific quantitative indicators of the three deviation state features, improving the accuracy and operability of the error mode recognition, thereby more reliably diagnosing the type of sensor error and providing a basis for subsequent selection of targeted calibration solutions.

[0119] In some preferred embodiments, step S5 comprises:

[0120] S51, extracting a corresponding calibration scheme from a preset calibration scheme library according to the landslide activity level, the error mode and the type of the corresponding sensor of the last time.

[0121] Specifically, the type of the sensor refers to the specific category of the sensor that occurs the error, which determines the measurement principle, physical characteristics and sensitivity to environmental factors of the sensor; the preset calibration scheme library refers to a set that stores a variety of sensor signal calibration strategies, methods and parameters, which structurally organizes the calibration schemes for different landslide states, error types and sensor types.

[0122] Specifically, the method of the present application, after the credibility score of the sensor signal is lower than the preset threshold and the error mode is identified, no longer selects the calibration scheme only according to the error mode, but further combines the landslide activity level of the last time and the type of the sensor that occurs the error. By taking these three information as the basis, query or matching is carried out in the preset calibration scheme library. The calibration scheme library pre-constructs the optimized calibration strategy under different combination conditions. For example, for the landslide in the active period, even the same zero drift error, different calibration algorithms or more stringent calibration parameters may need to be used for displacement sensors and pore water pressure sensors. By comprehensively considering these three factors, the system can extract a calibration scheme that is more targeted and more in line with the current actual monitoring scene and the characteristics of the sensor from the library. This multi-dimensional matching method makes the selected calibration scheme more effectively solve the specific error problem, while taking into account the demand of landslide state for data accuracy and reliability and the inherent characteristics of different sensor types. In this way, the calibration process is more intelligent and refined, improving the accuracy and reliability of the landslide monitoring data, and avoiding the mismatching problem that may be caused by using a general calibration scheme.

[0123] In some preferred embodiments, the calibration scheme includes a calibration means and a calibration parameter type, and step S6 includes:

[0124] S61, determining the calibration parameter based on the type of the parameter to be calibrated, according to the monitoring signal of the sensor to be calibrated and the historical signal of the same type of sensor under the same environmental condition or the same time condition;

[0125] S62, updating the configuration information of the corresponding sensor based on the calibration means and the calibration parameter to calibrate the monitoring signal of the corresponding sensor.

[0126] Specifically, the calibration scheme in the calibration scheme library includes a zero drift calibration scheme, a sensitivity attenuation calibration scheme and a periodic interference calibration scheme.

[0127] More specifically, the calibration means refers to a method or algorithm used to correct the sensor monitoring signal, which can be implemented in the form of adding a bias, multiplying a proportional coefficient, applying a filtering algorithm, or performing waveform fitting, etc.; the calibration parameter type refers to the kind of parameters that need to be determined when the calibration means is executed; the calibration parameter refers to the specific value calculated according to the calibration parameter type and actual data, which is used to quantify the sensor error and guide the calibration operation. Updating the configuration information of the corresponding sensor refers to applying the determined calibration parameter to the sensor data processing flow, which can be implemented in the form of modifying the parameters of the sensor driver, updating the configuration registers of the data acquisition module, or adjusting the algorithm parameters of the backend data processing software, etc.

[0128] Specifically, the present scheme solves the problem of how to effectively perform signal calibration after identifying the sensor error mode by explicitly defining the composition of the calibration scheme and the calibration steps. First, the system extracts the corresponding calibration scheme from the calibration scheme library according to the identified error mode (such as zero drift, sensitivity attenuation, or periodic interference). Each extracted calibration scheme provides the method (calibration means) required for calibration and the kind of parameters (calibration parameter type) that need to be determined. In step S61, the system calculates the specific calibration parameter value using the calibration parameter type specified in the calibration scheme, combined with the current monitoring signal of the sensor to be calibrated and the historical signal of the same type of sensor under similar conditions. For example, for zero drift, the mean difference between the current signal and the historical signal can be calculated as the bias value; for sensitivity attenuation, the amplitude ratio of the current signal to the historical signal can be calculated as the proportional coefficient. This way of dynamically determining the calibration parameter based on actual monitoring data and historical reference data improves the relevance and accuracy of calibration. Subsequently, in step S62, the system updates the configuration information of the corresponding sensor according to the specific calibration parameter value calculated in S61 and the calibration means obtained from the calibration scheme. This means that the calculated calibration parameter is applied to the subsequent signal processing process, for example, the original monitoring signal is modified in real time through a software algorithm, thereby realizing the calibration of the sensor monitoring signal. The whole process combines error identification, scheme selection, parameter determination, and signal correction organically to form an automated sensor signal calibration process. More specifically,

[0129] The method of the present application, combined with error mode identification and calibration scheme selection (as implemented in S4 and S5), provides the ability to automatically select the appropriate calibration method and determine the calibration parameter according to the actual deviation state of the sensor signal and the identified error mode, and then modify the sensor signal, which can effectively and accurately determine the required parameters for calibration and perform signal calibration operations. Thus, the problem of difficult implementation or poor effect in the calibration process is solved, and the reliability of landslide monitoring data is improved.

[0130] Second aspect, please refer toFigure 2 Some embodiments of the present application also provide a landslide monitoring sensor signal calibration system applied to sensors of a landslide state monitoring system, the system comprising:

[0131] The acquisition module 201 is configured to continuously acquire monitoring signals of the plurality of sensors.

[0132] The characteristic information acquisition module 202 is configured to analyze and acquire statistical characteristics of each sensor, consistency of changes of physically associated sensors, and deviation degree from historical signals according to the monitoring signals.

[0133] The scoring module 203 is configured to calculate a credibility score of each sensor according to the statistical characteristics, the consistency of changes, and the deviation degree.

[0134] The error analysis module 204 is configured to analyze and acquire an error mode according to one or more of the statistical characteristics, the consistency of changes, and the deviation degree when the credibility score is lower than a preset score threshold.

[0135] The calibration scheme determination module 205 is configured to extract a calibration scheme from a preset calibration scheme library according to the error mode.

[0136] The calibration module 206 is configured to calibrate the monitoring signal of the corresponding sensor according to the calibration scheme.

[0137] The system of the present application continuously analyzes the statistical characteristics of the sensor signals, the consistency of changes of the physically associated sensors, and the deviation degree from the historical signals, calculates the credibility score of the signals, and when the score is lower than the threshold, diagnoses the specific error mode based on the analysis results, extracts and applies the corresponding calibration scheme from the preset calibration scheme library, realizes the automatic and data-driven calibration of the landslide monitoring sensor signals, overcomes the limitations of the traditional manual calibration in the field environment, improves the accuracy of the monitoring data and the reliability of the system, reduces the demand and cost of manual calibration, and enhances the early warning capability of the landslide monitoring system.

[0138] In addition, the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one place or distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment.

[0139] Furthermore, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0140] In this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions.

[0141] The above description is merely illustrative of the application and not in limitation of the principles of the application. Numerous modifications and adaptations thereof will be readily apparent to those skilled in the art without departing from the spirit and scope of the application as defined in the following claims.

Claims

1. A landslide monitoring sensor signal calibration method applied to a sensor of a landslide state monitoring system, characterized in that, The method comprises the following steps: S1, continuously acquiring monitoring signals of multiple sensors; S2, analyzing statistical characteristics of each sensor, consistency of changes of physically related sensors, and deviation degree from historical signals according to the monitoring signals; S3, calculating a credibility score of each sensor according to the statistical characteristics, the consistency of changes, and the deviation degree; S4, when the credibility score is lower than a preset score threshold, analyzing an error mode according to one or more of the statistical characteristics, the consistency of changes, and the deviation degree; S5, extracting a calibration scheme from a preset calibration scheme library according to the error mode; S6, calibrating the monitoring signals of the corresponding sensor according to the calibration scheme; Step S3 comprises: S31, acquiring an activity level of the landslide at the last time; S32, extracting a weight set from a preset weight database according to the activity level and the type of the sensor; S33, calculating the credibility score according to the weight set, the statistical characteristics, the consistency of changes, and the deviation degree.

2. The landslide monitoring sensor signal calibration method of claim 1, wherein, Step S2 comprises: S21, calculating the mean, variance, and fluctuation range of each sensor as the statistical characteristics according to the monitoring signals; S22, calculating the Pearson correlation coefficient between a sensor pair with physical correlation as the consistency of changes of the corresponding sensors according to the sensor pair; S23, acquiring the deviation degree according to the difference between the historical signals and the monitoring signals of the same type of sensors under the same environmental conditions or the same time conditions.

3. The landslide monitoring sensor signal calibration method of claim 2, wherein, Step S23 comprises: S231, extracting the historical signals of the same type of sensors under the same environmental conditions or the same time conditions, and calculating the mean and variance of the historical signals; S232, calculating the mean difference and variance difference between the historical signals and the monitoring signals based on the statistical characteristics; S233, calculating the deviation degree according to the mean difference and the variance difference.

4. The landslide monitoring sensor signal calibration method of claim 1, wherein, The determination process of the activity level of the landslide comprises: A1, extracting the displacement amount of the corresponding landslide, the rainfall amount, and the groundwater level change amplitude within a preset time window; A2, acquiring the activity level according to the displacement amount of the landslide, the rainfall amount, and the groundwater level change amplitude by querying a preset activity level discrimination rule library.

5. The landslide monitoring sensor signal calibration method of claim 1, wherein, Step S4 comprises: S41, when the credibility score is lower than a preset score threshold, extracting a first deviation state feature, a second deviation state feature, and a third deviation state feature according to the statistical characteristics, the consistency of changes, and the deviation degree, respectively; S42, concatenating the first deviation state feature, the second deviation state feature, and the third deviation state feature to obtain a fused deviation feature vector; S43, inputting the fused deviation feature vector into a preset error mode recognition model to generate the error mode.

6. The landslide monitoring sensor signal calibration method of claim 5, wherein, The error mode comprises zero drift, sensitivity attenuation, intermittent failure, and periodic interference.

7. The landslide monitoring sensor signal calibration method of claim 1, wherein, Step S5 comprises: S51, extracting a corresponding calibration scheme from a preset calibration scheme library according to the activity level of the landslide at the last time, the error mode, and the type of the corresponding sensor.

8. The landslide monitoring sensor signal calibration method of claim 1, wherein, The calibration scheme comprises a calibration means and a calibration parameter type, and step S6 comprises: S61, determining the calibration parameter according to the monitoring signal of the sensor to be calibrated and the historical signal of the same type of sensor under the same environmental condition or the same time condition based on the calibration parameter type; S62, updating the configuration information of the corresponding sensor based on the calibration means and the calibration parameter to calibrate the monitoring signal of the corresponding sensor.

9. A landslide monitoring sensor signal calibration system applied to a sensor of a landslide state monitoring system, characterized by, The system comprises: an acquisition module configured to continuously acquire monitoring signals of a plurality of sensors; a characteristic information acquisition module configured to analyze and acquire statistical characteristics of each sensor, consistency of changes of the sensor associated with physics, and deviation degree from historical signals according to the monitoring signals; a scoring module configured to calculate a credibility score of each sensor according to the statistical characteristics, the consistency of changes, and the deviation degree; an error analysis module configured to analyze and acquire an error mode according to one or more of the statistical characteristics, the consistency of changes, and the deviation degree when the credibility score is lower than a preset score threshold; a calibration scheme determination module configured to extract a calibration scheme from a preset calibration scheme library according to the error mode; a calibration module configured to calibrate the monitoring signal of the corresponding sensor according to the calibration scheme; the step of calculating the credibility score of each sensor according to the statistical characteristics, the consistency of changes, and the deviation degree comprises: S31, acquiring an activity level of a landslide at a previous time; S32, extracting a weight set from a preset weight database according to the activity level and the type of the sensor; S33, calculating the credibility score according to the statistical characteristics, the consistency of changes, and the deviation degree based on the weight set.

Images