Sensor-based seasonal frozen soil monitoring method and system

Abstract

The invention provides a sensor-based seasonal frozen soil monitoring method and system, and the method comprises the steps: dividing a monitored region into grids, deploying a sensor group in the monitored region, obtaining the geographic parameters of the monitored region, and obtaining the terrain impact factors of seasonal frozen soil according to the geographic parameters; acquiring heat flux between two points of the sensor monitoring point i and the sensor monitoring point j in the monitored area; calculating a final weight matrix, and calculating spatial autocorrelation among the sensor monitoring points based on the terrain influence factors, the heat flux and the final weight matrix; unnecessary monitoring points are screened out according to spatial autocorrelation, the monitoring points are sequenced according to spatial autocorrelation values, the monitoring points with larger absolute values are sequenced in the front, and the monitoring points with higher sequencing are preferentially monitored. According to the method and the corresponding system, redundant monitoring points in seasonal frozen soil monitoring are screened, so that the deployment number and the maintenance cost of the sensors can be reduced, and the operation cost of the whole monitoring system is reduced.

Description

Technical Field

[0001] The invention proposes a sensor-based seasonal frozen soil monitoring method and system, and relates to the technical field of frozen soil monitoring. Background Art

[0002] The frozen soil area in the world accounts for about 70% of the continental area, and the frozen soil area in my country accounts for about 53% of the land area. Frozen soil is a soil medium that is extremely sensitive to temperature. Its freeze-thaw effect is affected by the temperature change during the season change, and the volume change of frozen soil is inevitable. Under the freeze-thaw effect of frozen soil, the rocks in the frozen soil area are destroyed, causing the buildings to suffer deformation, collapse and other hazards. In particular, water conservancy projects such as dams will cause incalculable losses if they are destroyed. Seasonal frozen soil is like the active layer of permafrost, which is always in an alternating state of melting in summer and freezing in winter.

[0003] However, the scope of traditional seasonal frozen soil monitoring is usually very large, and point monitoring using a large number of sensors will generate massive amounts of data. The method of direct comparison with threshold size is inefficient and prone to false alarms. An innovative method is needed to quickly and effectively detect anomalies and reduce the amount of data processing. Summary of the invention

[0004] The present invention provides a sensor-based seasonal frozen soil monitoring method and system to solve the above-mentioned problems:

[0005] The present invention proposes a sensor-based seasonal frozen soil monitoring method, the method comprising:

[0006] Divide the monitored area into grids, deploy sensor groups in the monitored area, obtain geographic parameters of the monitored area, and obtain terrain influence factors of seasonal frozen soil based on the geographic parameters;

[0007] Obtain the heat flux between the sensor monitoring point i and the sensor monitoring point j in the monitored area;

[0008] Calculate the final weight matrix, and calculate the spatial autocorrelation between sensor monitoring points based on terrain influence factors, heat flux and the final weight matrix;

[0009] Unnecessary monitoring points are screened out based on spatial autocorrelation, and the monitoring points are sorted according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points with higher rankings are monitored first.

[0010] Furthermore, the monitored area is divided into grids, a sensor group is deployed in the monitored area, geographic parameters of the monitored area are obtained, and terrain influencing factors of seasonal frozen soil are obtained according to the geographic parameters, including:

[0011] The monitored area is divided into grids, and sensors are deployed at equal intervals in each grid, wherein the sensors include: temperature sensors, moisture sensors, soil moisture sensors, and strain sensors;

[0012] Get the altitude, slope and slope direction of the sensor monitoring point;

[0013] Calculate the terrain influence factors between sensor monitoring points based on the altitude, slope and slope aspect of the sensor monitoring points;

[0014]

[0015] Among them, T ij represents the terrain influence factor between sensor monitoring points i and j, Δh ij It represents the elevation difference between sensor monitoring points i and j, Δs ij represents the slope difference between sensor monitoring points i and j, Δp ij represents the slope difference between sensor monitoring points i and j, H0 represents the characteristic elevation, S0 represents the characteristic slope, P0 represents the characteristic slope, k h Represents the elevation weight coefficient, k s represents the slope weight coefficient, k p Represents the aspect weight factor.

[0016] Furthermore, obtaining the heat flux between the sensor monitoring point i and the sensor monitoring point j in the monitored area includes:

[0017] For each sensor monitoring point, the weighted average thermal conductivity is calculated according to its depth information and the measurement data of the soil thermal conductivity sensor, and the weighted average thermal conductivity of the sensor monitoring point i is obtained. The weighted average thermal conductivity reflects the thermal conductivity of the soil at this point.

[0018]

[0019] Among them, λ i represents the weighted average thermal conductivity of monitoring point i, represents the soil thermal conductivity of the kth soil thermal conductivity sensor at monitoring point i, n i represents the number of sensors measuring soil thermal conductivity at monitoring point i, represents the weight of the k-th soil thermal conductivity sensor at monitoring point i,

[0020]

[0021] in, represents the depth of the kth sensor at monitoring point i,

[0022] According to the weighted average thermal conductivity of adjacent monitoring points, the average thermal conductivity between two points is calculated:

[0023]

[0024] Among them, λ ij represents the average thermal conductivity between sensor monitoring points i and j, λ j represents the weighted average thermal conductivity of monitoring point j;

[0025] Based on the average thermal conductivity between the sensor monitoring point i and the sensor monitoring point j, the heat flux between the sensor monitoring point i and the sensor monitoring point j is obtained.

[0026]

[0027] Among them, q ij represents the heat flux between monitoring points i and j, It represents the average temperature gradient between monitoring points i and j and can be calculated using the central difference method.

[0028] Furthermore, the final weight matrix is ​​calculated, and the spatial autocorrelation between the sensor monitoring points is calculated based on the terrain influence factor, heat flux and the final weight matrix, including:

[0029] The final weight matrix W is obtained based on the terrain influence factor and the heat conduction factor. ij =T ij ·(1+c·|q ij |), where c represents the weight factor;

[0030] For each sensor monitoring point i, the correlation model is used to quantify its correlation with surrounding points. Specifically, the correlation model is:

[0031]

[0032] Among them, x i represents the observation value of the i-th sensor, represents the average value of all sensor observations, s 2 represents the variance of the observations, W ij Represents the weight between the i-th point and the j-th point in the weight matrix.

[0033] Furthermore, unnecessary monitoring points are screened out based on spatial autocorrelation, and the monitoring points are sorted according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points with higher rankings are monitored first, including:

[0034] If I i is greater than the first preset threshold and |x i -∑ j Wij ·x j | is less than the second preset threshold, the sensor monitoring point is considered to be a redundant point and is removed;

[0035] Otherwise, the point is considered not a redundant point. Press I i Sort by absolute value, check I first i Larger points to obtain outliers.

[0036] The present invention proposes a sensor-based seasonal frozen soil monitoring system, the system comprising:

[0037] A terrain influence factor acquisition module is used to divide the monitored area into grids, deploy a sensor group in the monitored area, acquire geographical parameters of the monitored area, and acquire the terrain influence factor of seasonal frozen soil according to the geographical parameters;

[0038] A heat flux acquisition module is used to acquire the heat flux between two points, a sensor monitoring point i and a sensor monitoring point j, in the monitored area;

[0039] Obtain autocorrelation module, which is used to calculate the final weight matrix, and calculate the spatial autocorrelation between sensor monitoring points based on terrain influence factors, heat flux and the final weight matrix;

[0040] The screening module is used to screen out unnecessary monitoring points based on spatial autocorrelation and sort the monitoring points according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points with higher rankings are monitored first.

[0041] Furthermore, the module for obtaining terrain influence factors includes:

[0042] A division module is used to divide the monitored area into grids and deploy sensors at equal intervals in each grid, wherein the sensors include: temperature sensors, moisture sensors, soil moisture sensors and strain sensors;

[0043] The module for obtaining geographic data is used to obtain the altitude, slope and slope direction of the sensor monitoring point;

[0044] A terrain influence factor calculation module is used to calculate the terrain influence factors between sensor monitoring points based on the altitude, slope and slope direction of the sensor monitoring points;

[0045]

[0046] Among them, T ij represents the terrain influence factor between sensor monitoring points i and j, Δh ij It represents the elevation difference between sensor monitoring points i and j, Δs ij represents the slope difference between sensor monitoring points i and j, Δpij represents the slope difference between sensor monitoring points i and j, H0 represents the characteristic elevation, S0 represents the characteristic slope, P0 represents the characteristic slope, k h Represents the elevation weight coefficient, k s represents the slope weight coefficient, k p Represents the aspect weight factor.

[0047] Furthermore, the heat flux acquisition module includes:

[0048] The module for calculating the weighted average thermal conductivity of the sensor monitoring point i is used to calculate the weighted average thermal conductivity for each sensor monitoring point according to its depth information and the measurement data of the soil thermal conductivity sensor, and obtain the weighted average thermal conductivity of the sensor monitoring point i, wherein the weighted average thermal conductivity reflects the thermal conductivity performance of the soil at the point.

[0049]

[0050] Among them, λ i represents the weighted average thermal conductivity of monitoring point i, represents the soil thermal conductivity of the kth soil thermal conductivity sensor at monitoring point i, n i represents the number of sensors measuring soil thermal conductivity at monitoring point i, represents the weight of the k-th soil thermal conductivity sensor at monitoring point i,

[0051]

[0052] in, represents the depth of the kth sensor at monitoring point i,

[0053] The module for calculating the average thermal conductivity between two points is used to calculate the average thermal conductivity between two points based on the weighted average thermal conductivity of adjacent monitoring points:

[0054]

[0055] Among them, λ ij represents the average thermal conductivity between sensor monitoring points i and j, λ j represents the weighted average thermal conductivity of monitoring point j;

[0056] The heat flux calculation module is used to obtain the heat flux between the sensor monitoring point i and the sensor monitoring point j based on the average thermal conductivity between the two points.

[0057]

[0058] Among them, q ijrepresents the heat flux between monitoring points i and j, It represents the average temperature gradient between monitoring points i and j and can be calculated using the central difference method.

[0059] Further, the module for obtaining autocorrelation includes:

[0060] The module for obtaining the final weight matrix is ​​used to obtain the final weight matrix based on the terrain influence factor and the heat conduction factor, W ij =T ij ·(1+c·|q ij |), where c represents the weight factor;

[0061] The quantification correlation module is used to quantify the correlation between each sensor monitoring point i and the surrounding points through the correlation model. Specifically, the correlation model is:

[0062]

[0063] Among them, x i represents the observation value of the i-th sensor, represents the average value of all sensor observations, s 2 represents the variance of the observations, W ij Represents the weight between the i-th point and the j-th point in the weight matrix.

[0064] Furthermore, the screening module includes:

[0065] Screening module, used if I i is greater than the first preset threshold and |x i -∑ j W ij ·x j | is less than the second preset threshold, the sensor monitoring point is considered to be a redundant point and is removed;

[0066] Prioritize monitoring module, otherwise it is considered that the point is not a redundant point, press I i Sort by absolute value, check I first i Larger points to obtain outliers.

[0067] The invention has the following beneficial effects: improving monitoring efficiency, by scientifically and reasonably dividing regions and deploying sensors, and by optimizing the layout of monitoring points using spatial autocorrelation analysis, the monitoring efficiency can be significantly improved and unnecessary waste of resources can be reduced; enhancing monitoring accuracy, by combining terrain influencing factors and heat flux data for spatial autocorrelation analysis, the spatial distribution and dynamic changes of the frozen soil state can be more accurately reflected, thereby improving the accuracy of monitoring; reducing monitoring costs, by screening redundant monitoring points, the number of sensors deployed and the maintenance cost can be reduced, and the operating cost of the entire monitoring system can be reduced; providing support for decision-making, real-time and accurate frozen soil monitoring data can provide a scientific basis for relevant departments and decision makers, helping them to better understand the frozen soil state, formulate effective management and protection measures, and reduce the occurrence and losses of frozen soil disasters; by comprehensively using geographic information technology, sensor technology and data analysis technology, efficient and accurate monitoring of seasonal frozen soil states is achieved, providing strong support for frozen soil research, disaster warning and ecological protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic diagram of a sensor-based seasonal frozen soil monitoring method and system according to the present invention; DETAILED DESCRIPTION

[0069] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0070] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. The embodiments described are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0071] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0072] One embodiment of the present invention provides a sensor-based seasonal frozen soil monitoring method, the method comprising:

[0073] Divide the monitored area into grids, deploy sensor groups in the monitored area, obtain geographic parameters of the monitored area, and obtain terrain influence factors of seasonal frozen soil based on the geographic parameters;

[0074] Obtain the heat flux between the sensor monitoring point i and the sensor monitoring point j in the monitored area;

[0075] Calculate the final weight matrix, and calculate the spatial autocorrelation between sensor monitoring points based on terrain influence factors, heat flux and the final weight matrix;

[0076] Unnecessary monitoring points are screened out based on spatial autocorrelation, and the monitoring points are sorted according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points with higher rankings are monitored first.

[0077] The working principle and effect of the above technical solution are as follows: First, the monitored area is scientifically and reasonably divided into multiple small areas or grids to ensure that the geographical and climatic conditions in each area or grid are relatively consistent, which is convenient for subsequent data collection and analysis; a sensor group is deployed in each area or grid, and these sensors can collect geographical parameters (such as altitude, slope, slope aspect, etc.) and related parameters of frozen soil (such as temperature, humidity, deformation, etc.) in real time; according to the collected geographical parameters, a specific mathematical model or algorithm is used to calculate the terrain influence factor. This factor reflects the degree of influence of terrain differences on the state of frozen soil (such as freezing depth, melting rate, etc.); the heat flux between two points is calculated through the temperature difference between the sensor monitoring points and the soil thermal conductivity and other data. Heat flux is a physical quantity that describes the rate and direction of heat transfer in the soil, which is crucial to understanding the heat balance and dynamic changes of frozen soil; the final weight matrix is ​​constructed by combining the terrain influence factor and heat flux data. This matrix reflects the degree of spatial mutual influence between different monitoring points. The spatial autocorrelation model is used to calculate the spatial autocorrelation between sensor monitoring points. Spatial autocorrelation is an indicator that describes the similarity or correlation between a point in geographic space and its surrounding points. Monitoring point optimization and sorting, based on the spatial autocorrelation results, screen out redundant or unnecessary monitoring points to reduce monitoring costs and data redundancy, and sort the remaining monitoring points according to the size of the spatial autocorrelation value. Points with larger absolute values ​​indicate that they deviate greatly from the average value, that is, they are prone to abnormal situations. Therefore, these points are monitored first to more accurately capture changes in the permafrost state. Improve monitoring efficiency. By scientifically and rationally dividing areas and deploying sensors, and using spatial autocorrelation analysis to optimize the layout of monitoring points, the monitoring efficiency can be significantly improved and unnecessary waste of resources can be reduced. Enhance monitoring accuracy. Combining terrain influencing factors and heat flux data for spatial autocorrelation analysis can more accurately reflect the spatial distribution and dynamic changes of permafrost conditions, thereby improving monitoring accuracy. Reduce monitoring costs. By screening redundant monitoring points, the number of sensors deployed and maintenance costs can be reduced, thereby reducing the operating costs of the entire monitoring system. Provide support for decision-making. Real-time and accurate permafrost monitoring data can provide a scientific basis for relevant departments and decision makers, helping them to better understand the permafrost status, formulate effective management and protection measures, and reduce the occurrence and losses of permafrost disasters. Through the comprehensive use of geographic information technology, sensor technology and data analysis technology, efficient and accurate monitoring of seasonal permafrost conditions is achieved, providing strong support for permafrost research, disaster warning and ecological protection.

[0078] In one embodiment of the present invention, the monitored area is divided into grids, a sensor group is deployed in the monitored area, geographic parameters of the monitored area are obtained, and terrain influence factors of seasonal frozen soil are obtained according to the geographic parameters, including:

[0079] The monitored area is divided into grids, and sensors are deployed at equal intervals in each grid, wherein the sensors include: temperature sensors, moisture sensors, soil moisture sensors, and strain sensors;

[0080] Get the altitude, slope and slope direction of the sensor monitoring point;

[0081] Calculate the terrain influence factors between sensor monitoring points based on the altitude, slope and slope aspect of the sensor monitoring points;

[0082]

[0083] Among them, T ij represents the terrain influence factor between sensor monitoring points i and j, Δh ij It represents the elevation difference between sensor monitoring points i and j, Δs ij represents the slope difference between sensor monitoring points i and j, Δp ij represents the slope difference between sensor monitoring points i and j, H0 represents the characteristic elevation, S0 represents the characteristic slope, P0 represents the characteristic slope, k h Represents the elevation weight coefficient, k s represents the slope weight coefficient, k p Represents the aspect weight factor.

[0084] The working principle and effect of the above technical scheme are as follows: it is based on the influence of terrain factors (altitude, slope and aspect) on the state of seasonal frozen soil, and these influences are evaluated by calculating the terrain influence factors between sensor monitoring points. Specifically, the altitude, slope and aspect data of each sensor monitoring point are first obtained, and then the terrain influence factor between any two monitoring points is calculated using the given formula. The terrain influence factor takes into account the elevation difference, slope difference and aspect difference, and is adjusted by the characteristic value H and the corresponding weight coefficient to reflect the relative importance of the influence of different terrain factors on the frozen soil state. Accurately evaluate the impact of terrain: by comprehensively considering the three terrain factors of altitude, slope and aspect, as well as the relative differences between them, the scheme can more accurately evaluate the impact of terrain on the frozen soil state, and provide a more reliable data basis for subsequent frozen soil monitoring; improve monitoring efficiency, by calculating the terrain influence factor, it is possible to identify monitoring points with relatively consistent terrain influence factors, thereby optimizing the layout of monitoring points, reducing unnecessary monitoring points, and improving monitoring efficiency; support scientific research, the calculation of terrain influence factors provides quantitative data support for scientific research, and helps to deeply understand the influence mechanism of terrain on the distribution and dynamic changes of frozen soil.

[0085] In one embodiment of the present invention, obtaining the heat flux between a sensor monitoring point i and a sensor monitoring point j in a monitored area includes:

[0086] For each sensor monitoring point, the weighted average thermal conductivity is calculated according to its depth information and the measurement data of the soil thermal conductivity sensor, and the weighted average thermal conductivity of the sensor monitoring point i is obtained. The weighted average thermal conductivity reflects the thermal conductivity of the soil at this point.

[0087]

[0088] Among them, λ i represents the weighted average thermal conductivity of monitoring point i, represents the soil thermal conductivity of the kth soil thermal conductivity sensor at monitoring point i, n i represents the number of sensors measuring soil thermal conductivity at monitoring point i, represents the weight of the k-th soil thermal conductivity sensor at monitoring point i,

[0089]

[0090] in, represents the depth of the kth sensor at monitoring point i, represents the depth of the k+1th sensor at monitoring point i;

[0091] According to the weighted average thermal conductivity of adjacent monitoring points, the average thermal conductivity between two points is calculated:

[0092]

[0093] Among them, λ ij represents the average thermal conductivity between sensor monitoring points i and j, λ j represents the weighted average thermal conductivity of monitoring point j;

[0094] Based on the average thermal conductivity between the sensor monitoring point i and the sensor monitoring point j, the heat flux between the sensor monitoring point i and the sensor monitoring point j is obtained.

[0095]

[0096] Among them, q ij represents the heat flux between monitoring points i and j, It represents the average temperature gradient between monitoring points i and j and can be calculated using the central difference method.

[0097]

[0098] in, represents the temperature measured by the kth sensor at monitoring point i, represents the temperature measured by the k+1th sensor at monitoring point i, n j Represents the number of sensors measuring soil thermal conductivity at monitoring point j.

[0099] The working principle and effect of the above technical scheme are as follows: the above technical scheme is based on the measurement and calculation of soil thermal conductivity (i.e. thermal conductivity) to evaluate the heat flux between different monitoring points; first, for each sensor monitoring point, the weighted average thermal conductivity is calculated according to its depth information and the measurement data of the soil thermal conductivity sensor, and this value reflects the thermal conductivity of the soil at that point; then, the average thermal conductivity between the two points is calculated according to the weighted average thermal conductivity of the adjacent monitoring points. Finally, the heat flux between the two points is calculated based on this average thermal conductivity and the average temperature gradient between the two points. Accurately evaluate the thermal conductivity, by calculating the weighted average thermal conductivity, it can more accurately reflect the thermal conductivity of soil at different depths, and provide accurate basic data for subsequent heat flux calculations; improve the accuracy of heat flux calculations: the average thermal conductivity and average temperature gradient of adjacent monitoring points are used to calculate the heat flux, taking into account the changes in soil thermal conductivity and the influence of temperature gradient, and improving the accuracy of heat flux calculations; support the study of soil thermal conditions, the scheme provides quantitative data support for the study of soil thermal conditions, and helps to deeply understand the soil heat conduction mechanism and its impact on environmental changes. The weighted average thermal conductivity formula calculates the weighted average thermal conductivity of the monitoring point by considering the thermal conductivity measurement values ​​of sensors at different depths and their corresponding weights (proportional to the depth difference). This design can more accurately reflect the thermal conductivity of the entire soil profile; it can comprehensively consider the thermal conductivity of soil at different depths, avoid the errors that may be caused by a single measurement value, and improve the accuracy and representativeness of thermal conductivity calculation; the average thermal conductivity formula between two points simply takes the average of the weighted average thermal conductivities of two adjacent monitoring points as the average thermal conductivity between the two points. This design is easy to calculate and can reflect the average level of soil thermal conductivity between the two points. The calculation is simple and intuitive, and the average thermal conductivity between the two points can be quickly obtained, which facilitates the subsequent heat flux calculation; the heat flux formula is based on the law of heat conduction and uses the average thermal conductivity and the average temperature gradient to calculate the heat flux. This design can accurately reflect the heat transfer in the soil, has a clear physical meaning, and is accurate in calculation. It can be directly used to evaluate the heat transfer in the soil, provide strong support for the study of soil thermal conditions, and deeply obtain the similarities and differences between the thermal conductivity properties of seasonally frozen soils.

[0100] In one embodiment of the present invention, a final weight matrix is ​​calculated, and the spatial autocorrelation between sensor monitoring points is calculated based on the terrain influence factor, the heat flux and the final weight matrix, including:

[0101] The final weight matrix W is obtained based on the terrain influence factor and the heat conduction factor. ij =T ij ·(1+c·|q ij |), where c represents the weight factor;

[0102] For each sensor monitoring point i, the correlation model is used to quantify its correlation with surrounding points. Specifically, the correlation model is:

[0103]

[0104] Among them, x i represents the observation value of the i-th sensor, represents the average value of all sensor observations, s 2 represents the variance of the observations, W ij Represents the weight between the i-th point and the j-th point in the weight matrix.

[0105] The working principle and effect of the above technical solution are as follows: first, the final weight matrix between the sensor monitoring points is calculated based on the terrain influence factor and the heat flux. The terrain influence factor takes into account the influence of factors such as altitude, slope, and slope aspect on the relationship between the monitoring points; the heat flux reflects the heat exchange between the monitoring points due to heat conduction, and c is the weight factor, which is used to adjust the relative importance of the heat flux in the weight calculation. The weight matrix Wij obtained in this way takes into account both the terrain factors and the influence of heat conduction, so that it can more comprehensively describe the relationship between the monitoring points; for each sensor monitoring point i, its correlation with the surrounding points is quantified through the correlation model. The correlation model measures the correlation between the monitoring point and the surrounding points by calculating the difference between the weighted average of the observation values ​​of each monitoring point and all other monitoring points. The weight matrix plays a key role here. It adjusts the correlation strength between different monitoring points according to the terrain and heat conduction factors. Taking into account a variety of factors, this technical solution can more comprehensively describe the relationship between monitoring points by integrating terrain influence factors, heat conduction factors and sensor observation data, thereby improving the accuracy and reliability of data analysis; enhancing data interpretability, by quantifying the correlation between monitoring points, this technical solution provides a powerful tool for data analysis, helps to reveal the potential interactions and associations between monitoring points, and enhances the interpretability of data; supports decision making, based on the calculated weight matrix and correlation quantification results, decision makers can more scientifically evaluate the importance of different monitoring points and their relationships, thereby formulating more reasonable monitoring strategies and decisions. The weight matrix formula combines terrain influence factors and heat conduction factors, and calculates the weights between monitoring points by multiplication and weighted summation. This method not only takes into account the basic influence of terrain on the relationship between monitoring points, but also incorporates the adjustment effect of heat conduction, a dynamic factor, on the relationship strength; this design enables the weight matrix to more accurately reflect the actual relationship strength between monitoring points, thereby improving the accuracy and practicality of data analysis. At the same time, by adjusting the weight factor c, the relative importance of terrain and heat conduction factors in the weight calculation can be flexibly balanced; the correlation quantification model is based on the concept of Z score (standardized score) in statistics, and quantifies the correlation by calculating the difference between each monitoring point and the weighted average of all other monitoring point observations. The weight matrix here plays a role in adjusting the contribution of observations from different monitoring points to the correlation calculation; this design enables the correlation quantification model to more accurately reflect the actual correlation strength between monitoring points, avoiding the misleading nature of a single observation. At the same time, by introducing the weight matrix, the model can more flexibly adapt to the complex relationship between different monitoring points, improving the flexibility and applicability of data analysis.

[0106] In one embodiment of the present invention, unnecessary monitoring points are screened out according to spatial autocorrelation, and the monitoring points are sorted according to the size of the spatial autocorrelation value, with the ones with larger absolute values ​​being ranked first, and the monitoring points ranked first are monitored first, including:

[0107] If I i is greater than the first preset threshold and |x i -∑ j W ij ·x j | is less than the second preset threshold, the sensor monitoring point is considered to be a redundant point and is removed;

[0108] Otherwise, the point is considered not a redundant point. Press I i Sort by absolute value, check I first i Larger points to obtain outliers.

[0109] The working principle and effect of the above technical solution are as follows: based on two conditions, determine whether the sensor monitoring point is a redundant point, and further process the abnormal point; for redundant point determination, first, calculate the correlation quantization value I_i of each sensor monitoring point, which reflects the correlation strength between the monitoring point and the surrounding points, and then determine I_i. i Is it greater than the first preset threshold? If so, it means that the correlation between the monitoring point and the surrounding points is very strong and may be redundant. At the same time, calculate the difference between the observation value of the monitoring point and the weighted average observation value of the surrounding points |x i -∑ j W ij ·x j |, and determine whether it is less than the second preset threshold. If it is less than, it means that the observation value of the monitoring point is very close to the observation value of the surrounding points, which further supports the judgment that it is a redundant point. If the above two conditions are met at the same time, the sensor monitoring point is considered to be a redundant point and is removed. For monitoring points that do not meet the redundant point judgment conditions, they are sorted according to the absolute value of. Prioritize I iLarger points, because these points are not highly correlated with surrounding points, but the observed values ​​may differ greatly, and are more likely to be outliers. Further inspection and processing of these points, such as manual review, data cleaning, or marking as outliers, can be performed. Improve data quality. By removing redundant points, duplicate information in the data set can be reduced, and the quality and representativeness of the data can be improved. This helps with subsequent data analysis and modeling, making the results more accurate and reliable; optimize the monitoring network. After removing redundant points, the layout of the sensor monitoring network can be optimized, and unnecessary monitoring points can be reduced, thereby saving resources and costs. At the same time, the retained monitoring points are more critical and effective, and can more comprehensively reflect the situation in the monitoring area; enhance the ability to detect abnormal points: by giving priority to checking points with larger I_i, abnormal points can be more effectively identified and processed, which helps to timely detect the melting of seasonal frozen soil and improve the accuracy and credibility of the data; by setting thresholds and sorting mechanisms, rapid screening and classification of monitoring points are achieved, which greatly improves the efficiency of data processing and reduces the time and cost of manual intervention; by comprehensively considering the correlation of monitoring points and the difference in observed values, redundant points and abnormal points are effectively identified and processed, and the quality and processing efficiency of data are improved. At the same time, it also optimizes the layout of the monitoring network and provides strong support for the subsequent analysis and application of seasonal frozen soil monitoring data.

[0110] One embodiment of the present invention is a sensor-based seasonal frozen soil monitoring system, the system comprising:

[0111] A terrain influence factor acquisition module is used to divide the monitored area into grids, deploy a sensor group in the monitored area, acquire geographical parameters of the monitored area, and acquire the terrain influence factor of seasonal frozen soil according to the geographical parameters;

[0112] A heat flux acquisition module is used to acquire the heat flux between two points, a sensor monitoring point i and a sensor monitoring point j, in the monitored area;

[0113] Obtain autocorrelation module, which is used to calculate the final weight matrix, and calculate the spatial autocorrelation between sensor monitoring points based on terrain influence factors, heat flux and the final weight matrix;

[0114] The screening module is used to screen out unnecessary monitoring points based on spatial autocorrelation and sort the monitoring points according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points with higher rankings are monitored first.

[0115] The working principle and effect of the above technical solution are as follows: First, the monitored area is scientifically and reasonably divided into multiple small areas or grids to ensure that the geographical and climatic conditions in each area or grid are relatively consistent, which is convenient for subsequent data collection and analysis; a sensor group is deployed in each area or grid, and these sensors can collect geographical parameters (such as altitude, slope, slope aspect, etc.) and related parameters of frozen soil (such as temperature, humidity, deformation, etc.) in real time; according to the collected geographical parameters, a specific mathematical model or algorithm is used to calculate the terrain influence factor. This factor reflects the degree of influence of terrain differences on the state of frozen soil (such as freezing depth, melting rate, etc.); the heat flux between two points is calculated through the temperature difference between the sensor monitoring points and the soil thermal conductivity and other data. Heat flux is a physical quantity that describes the rate and direction of heat transfer in the soil, which is crucial to understanding the heat balance and dynamic changes of frozen soil; the final weight matrix is ​​constructed by combining the terrain influence factor and heat flux data. This matrix reflects the degree of spatial mutual influence between different monitoring points. The spatial autocorrelation model is used to calculate the spatial autocorrelation between sensor monitoring points. Spatial autocorrelation is an indicator that describes the similarity or correlation between a point in geographic space and its surrounding points. Monitoring point optimization and sorting, based on the spatial autocorrelation results, screen out redundant or unnecessary monitoring points to reduce monitoring costs and data redundancy, and sort the remaining monitoring points according to the size of the spatial autocorrelation value. Points with larger absolute values ​​indicate that they deviate greatly from the average value, that is, they are prone to abnormal situations. Therefore, these points are monitored first to more accurately capture changes in the permafrost state. Improve monitoring efficiency. By scientifically and rationally dividing areas and deploying sensors, and using spatial autocorrelation analysis to optimize the layout of monitoring points, the monitoring efficiency can be significantly improved and unnecessary waste of resources can be reduced. Enhance monitoring accuracy. Combining terrain influencing factors and heat flux data for spatial autocorrelation analysis can more accurately reflect the spatial distribution and dynamic changes of permafrost conditions, thereby improving monitoring accuracy. Reduce monitoring costs. By screening redundant monitoring points, the number of sensors deployed and maintenance costs can be reduced, thereby reducing the operating costs of the entire monitoring system. Provide support for decision-making. Real-time and accurate permafrost monitoring data can provide a scientific basis for relevant departments and decision makers, helping them to better understand the permafrost status, formulate effective management and protection measures, and reduce the occurrence and losses of permafrost disasters. Through the comprehensive use of geographic information technology, sensor technology and data analysis technology, efficient and accurate monitoring of seasonal permafrost conditions is achieved, providing strong support for permafrost research, disaster warning and ecological protection.

[0116] In one embodiment of the present invention, the module for obtaining terrain influence factors includes:

[0117] A division module is used to divide the monitored area into grids and deploy sensors at equal intervals in each grid, wherein the sensors include: temperature sensors, moisture sensors, soil moisture sensors and strain sensors;

[0118] The module for obtaining geographic data is used to obtain the altitude, slope and slope direction of the sensor monitoring point;

[0119] A terrain influence factor calculation module is used to calculate the terrain influence factors between sensor monitoring points based on the altitude, slope and slope direction of the sensor monitoring points;

[0120]

[0121] Among them, T ij represents the terrain influence factor between sensor monitoring points i and j, Δh ij It represents the elevation difference between sensor monitoring points i and j, Δs ij represents the slope difference between sensor monitoring points i and j, Δp ij represents the slope difference between sensor monitoring points i and j, H0 represents the characteristic elevation, S0 represents the characteristic slope, P0 represents the characteristic slope, k h Represents the elevation weight coefficient, k s represents the slope weight coefficient, k p Represents the aspect weight factor.

[0122] The working principle and effect of the above technical scheme are as follows: it is based on the influence of terrain factors (altitude, slope and aspect) on the state of seasonal frozen soil, and these influences are evaluated by calculating the terrain influence factors between sensor monitoring points. Specifically, the altitude, slope and aspect data of each sensor monitoring point are first obtained, and then the terrain influence factor between any two monitoring points is calculated using the given formula. The terrain influence factor takes into account the elevation difference, slope difference and aspect difference, and is adjusted by the characteristic value H and the corresponding weight coefficient to reflect the relative importance of the influence of different terrain factors on the frozen soil state. Accurately evaluate the impact of terrain: by comprehensively considering the three terrain factors of altitude, slope and aspect, as well as the relative differences between them, the scheme can more accurately evaluate the impact of terrain on the frozen soil state, and provide a more reliable data basis for subsequent frozen soil monitoring; improve monitoring efficiency, by calculating the terrain influence factor, it is possible to identify monitoring points with relatively consistent terrain influence factors, thereby optimizing the layout of monitoring points, reducing unnecessary monitoring points, and improving monitoring efficiency; support scientific research, the calculation of terrain influence factors provides quantitative data support for scientific research, and helps to deeply understand the influence mechanism of terrain on the distribution and dynamic changes of frozen soil.

[0123] In one embodiment of the present invention, the heat flux acquisition module comprises:

[0124] The module for calculating the weighted average thermal conductivity of the sensor monitoring point i is used to calculate the weighted average thermal conductivity for each sensor monitoring point according to its depth information and the measurement data of the soil thermal conductivity sensor, and obtain the weighted average thermal conductivity of the sensor monitoring point i, wherein the weighted average thermal conductivity reflects the thermal conductivity performance of the soil at the point.

[0125]

[0126] Among them, λ i represents the weighted average thermal conductivity of monitoring point i, represents the soil thermal conductivity of the kth soil thermal conductivity sensor at monitoring point i, n i represents the number of sensors measuring soil thermal conductivity at monitoring point i, represents the weight of the k-th soil thermal conductivity sensor at monitoring point i,

[0127]

[0128] in, represents the depth of the kth sensor at monitoring point i,

[0129] The module for calculating the average thermal conductivity between two points is used to calculate the average thermal conductivity between two points based on the weighted average thermal conductivity of adjacent monitoring points:

[0130]

[0131] Among them, λ ij represents the average thermal conductivity between sensor monitoring points i and j, λ j represents the weighted average thermal conductivity of monitoring point j;

[0132] The heat flux calculation module is used to obtain the heat flux between the sensor monitoring point i and the sensor monitoring point j based on the average thermal conductivity between the two points.

[0133]

[0134] Among them, q ij represents the heat flux between monitoring points i and j, It represents the average temperature gradient between monitoring points i and j and can be calculated using the central difference method.

[0135] In one embodiment of the present invention, the module for obtaining autocorrelation includes:

[0136] The module for obtaining the final weight matrix is ​​used to obtain the final weight matrix based on the terrain influence factor and the heat conduction factor, W ij =T ij·(1+c·|q ij |), where c represents the weight factor;

[0137] The quantification correlation module is used to quantify the correlation between each sensor monitoring point i and the surrounding points through the correlation model. Specifically, the correlation model is:

[0138]

[0139] Among them, x i represents the observation value of the i-th sensor, represents the average value of all sensor observations, s 2 represents the variance of the observations, W ij Represents the weight between the i-th point and the j-th point in the weight matrix.

[0140] The observation values ​​of the i-th sensor include: temperature of seasonally frozen soil, soil moisture, soil humidity sensor and strain data.

[0141] The working principle and effect of the above technical scheme are as follows: the above technical scheme is based on the measurement and calculation of soil thermal conductivity (i.e. thermal conductivity) to evaluate the heat flux between different monitoring points; first, for each sensor monitoring point, the weighted average thermal conductivity is calculated according to its depth information and the measurement data of the soil thermal conductivity sensor, and this value reflects the thermal conductivity of the soil at that point; then, the average thermal conductivity between the two points is calculated according to the weighted average thermal conductivity of the adjacent monitoring points. Finally, the heat flux between the two points is calculated based on this average thermal conductivity and the average temperature gradient between the two points. Accurately evaluate the thermal conductivity, by calculating the weighted average thermal conductivity, it can more accurately reflect the thermal conductivity of soil at different depths, and provide accurate basic data for subsequent heat flux calculations; improve the accuracy of heat flux calculations: the average thermal conductivity and average temperature gradient of adjacent monitoring points are used to calculate the heat flux, taking into account the changes in soil thermal conductivity and the influence of temperature gradient, and improving the accuracy of heat flux calculations; support the study of soil thermal conditions, the scheme provides quantitative data support for the study of soil thermal conditions, and helps to deeply understand the soil heat conduction mechanism and its impact on environmental changes. The weighted average thermal conductivity formula calculates the weighted average thermal conductivity of the monitoring point by considering the thermal conductivity measurement values ​​of sensors at different depths and their corresponding weights (proportional to the depth difference). This design can more accurately reflect the thermal conductivity of the entire soil profile; it can comprehensively consider the thermal conductivity of soil at different depths, avoid the errors that may be caused by a single measurement value, and improve the accuracy and representativeness of thermal conductivity calculation; the average thermal conductivity formula between two points simply takes the average of the weighted average thermal conductivities of two adjacent monitoring points as the average thermal conductivity between the two points. This design is easy to calculate and can reflect the average level of soil thermal conductivity between the two points. The calculation is simple and intuitive, and the average thermal conductivity between the two points can be quickly obtained, which facilitates the subsequent heat flux calculation; the heat flux formula is based on the law of heat conduction and uses the average thermal conductivity and the average temperature gradient to calculate the heat flux. This design can accurately reflect the heat transfer in the soil, has a clear physical meaning, and is accurate in calculation. It can be directly used to evaluate the heat transfer in the soil, provide strong support for the study of soil thermal conditions, and deeply obtain the similarities and differences between the thermal conductivity properties of seasonally frozen soils.

[0142] In one embodiment of the present invention, the screening module comprises:

[0143] Screening module, used if I i is greater than the first preset threshold and |x i -∑ j W ij ·x j | is less than the second preset threshold, the sensor monitoring point is considered to be a redundant point and is removed;

[0144] Prioritize monitoring module, otherwise it is considered that the point is not a redundant point, press Ii Sort by absolute value, check I first i Larger points to obtain outliers.

[0145] The working principle and effect of the above technical solution are as follows: first, the final weight matrix between the sensor monitoring points is calculated based on the terrain influence factor and the heat flux. The terrain influence factor takes into account the influence of factors such as altitude, slope, and slope aspect on the relationship between the monitoring points; the heat flux reflects the heat exchange between the monitoring points due to heat conduction, and c is the weight factor, which is used to adjust the relative importance of the heat flux in the weight calculation. The weight matrix Wij obtained in this way takes into account both the terrain factors and the influence of heat conduction, so that it can more comprehensively describe the relationship between the monitoring points; for each sensor monitoring point i, its correlation with the surrounding points is quantified through the correlation model. The correlation model measures the correlation between the monitoring point and the surrounding points by calculating the difference between the weighted average of the observation values ​​of each monitoring point and all other monitoring points. The weight matrix plays a key role here. It adjusts the correlation strength between different monitoring points according to the terrain and heat conduction factors. Taking into account a variety of factors, this technical solution can more comprehensively describe the relationship between monitoring points by integrating terrain influence factors, heat conduction factors and sensor observation data, thereby improving the accuracy and reliability of data analysis; enhancing data interpretability, by quantifying the correlation between monitoring points, this technical solution provides a powerful tool for data analysis, helps to reveal the potential interactions and associations between monitoring points, and enhances the interpretability of data; supports decision making, based on the calculated weight matrix and correlation quantification results, decision makers can more scientifically evaluate the importance of different monitoring points and their relationships, thereby formulating more reasonable monitoring strategies and decisions. The weight matrix formula combines terrain influence factors and heat conduction factors, and calculates the weights between monitoring points by multiplication and weighted summation. This method not only takes into account the basic influence of terrain on the relationship between monitoring points, but also incorporates the adjustment effect of heat conduction, a dynamic factor, on the relationship strength; this design enables the weight matrix to more accurately reflect the actual relationship strength between monitoring points, thereby improving the accuracy and practicality of data analysis. At the same time, by adjusting the weight factor c, the relative importance of terrain and heat conduction factors in the weight calculation can be flexibly balanced; the correlation quantification model is based on the concept of Z score (standardized score) in statistics, and quantifies the correlation by calculating the difference between each monitoring point and the weighted average of all other monitoring point observations. The weight matrix here plays a role in adjusting the contribution of observations from different monitoring points to the correlation calculation; this design enables the correlation quantification model to more accurately reflect the actual correlation strength between monitoring points, avoiding the misleading nature of a single observation. At the same time, by introducing the weight matrix, the model can more flexibly adapt to the complex relationship between different monitoring points, improving the flexibility and applicability of data analysis.

[0146] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A sensor-based seasonal frozen soil monitoring method, characterized in that: The method comprises: Divide the monitored area into grids, deploy sensor groups in the monitored area, obtain geographic parameters of the monitored area, and obtain terrain influence factors of seasonal frozen soil based on the geographic parameters; Obtain the heat flux between the sensor monitoring point i and the sensor monitoring point j in the monitored area; Calculate the final weight matrix, and calculate the spatial autocorrelation between sensor monitoring points based on terrain influence factors, heat flux and the final weight matrix; Unnecessary monitoring points are screened out based on spatial autocorrelation, and the monitoring points are sorted according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points with higher rankings are monitored first.

2. A sensor-based seasonal frozen soil monitoring method according to claim 1, characterized in that: The monitored area is divided into grids, a sensor group is deployed in the monitored area, the geographical parameters of the monitored area are obtained, and the terrain influencing factors of seasonal frozen soil are obtained according to the geographical parameters, including: The monitored area is divided into grids, and sensors are deployed at equal intervals in each grid, wherein the sensors include: temperature sensors, moisture sensors, soil moisture sensors, and strain sensors; Get the altitude, slope and slope direction of the sensor monitoring point; Calculate the terrain influence factors between sensor monitoring points based on the altitude, slope and slope aspect of the sensor monitoring points; Among them, T ij represents the terrain influence factor between sensor monitoring points i and j, Δh ij It represents the elevation difference between sensor monitoring points i and j, Δs ij represents the slope difference between sensor monitoring points i and j, Δp ij represents the slope difference between sensor monitoring points i and j, H0 represents the characteristic elevation, S0 represents the characteristic slope, P0 represents the characteristic slope, k h Represents the elevation weight coefficient, k s represents the slope weight coefficient, k p Represents the aspect weight factor.

3. A sensor-based seasonal frozen soil monitoring method according to claim 1, characterized in that: Obtaining the heat flux between the sensor monitoring point i and the sensor monitoring point j in the monitored area includes: For each sensor monitoring point, the weighted average thermal conductivity is calculated according to its depth information and the measurement data of the soil thermal conductivity sensor, and the weighted average thermal conductivity of the sensor monitoring point i is obtained. The weighted average thermal conductivity reflects the thermal conductivity of the soil at this point. Among them, λ i represents the weighted average thermal conductivity of monitoring point i, represents the soil thermal conductivity of the kth soil thermal conductivity sensor at monitoring point i, n i represents the number of sensors measuring soil thermal conductivity at monitoring point i, represents the weight of the k-th soil thermal conductivity sensor at monitoring point i, in, represents the depth of the kth sensor at monitoring point i, According to the weighted average thermal conductivity of adjacent monitoring points, the average thermal conductivity between two points is calculated: Among them, λ ij represents the average thermal conductivity between sensor monitoring points i and j, λ j represents the weighted average thermal conductivity of monitoring point j; Based on the average thermal conductivity between the sensor monitoring point i and the sensor monitoring point j, the heat flux between the sensor monitoring point i and the sensor monitoring point j is obtained. Among them, q ij represents the heat flux between monitoring points i and j, represents the average temperature gradient between monitoring points i and j, which can be calculated using the central difference method.

4. A sensor-based seasonal frozen soil monitoring method according to claim 1, characterized in that: Calculate the final weight matrix, and calculate the spatial autocorrelation between sensor monitoring points based on terrain influence factors, heat flux and the final weight matrix, including: The final weight matrix W is obtained based on the terrain influence factor and the heat conduction factor. ij =T ij ·(1+c·|q ij |), where c represents the weight factor; For each sensor monitoring point i, the correlation model is used to quantify its correlation with surrounding points. Specifically, the correlation model is: Among them, x i represents the observation value of the i-th sensor, represents the average value of all sensor observations, s 2 represents the variance of the observations, W ij Represents the weight between the i-th point and the j-th point in the weight matrix.

5. A sensor-based seasonal frozen soil monitoring method according to claim 1, characterized in that: Unnecessary monitoring points are screened out based on spatial autocorrelation, and the monitoring points are sorted according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points with higher rankings are prioritized, including: If I i is greater than the first preset threshold and |x i -∑ j W ij ·x j | is less than the second preset threshold, the sensor monitoring point is considered to be a redundant point and is removed; Otherwise, the point is considered not a redundant point. Press I i Sort by absolute value, check I first i Larger points to obtain outliers.

6. A sensor-based seasonal frozen soil monitoring system, characterized in that: The system comprises: A terrain influence factor acquisition module is used to divide the monitored area into grids, deploy a sensor group in the monitored area, acquire geographical parameters of the monitored area, and acquire the terrain influence factor of seasonal frozen soil according to the geographical parameters; A heat flux acquisition module is used to acquire the heat flux between two points, a sensor monitoring point i and a sensor monitoring point j, in the monitored area; Obtain autocorrelation module, which is used to calculate the final weight matrix, and calculate the spatial autocorrelation between sensor monitoring points based on terrain influence factors, heat flux and the final weight matrix; The screening module is used to screen out unnecessary monitoring points based on spatial autocorrelation and sort the monitoring points according to the size of the spatial autocorrelation value. The monitoring points with larger absolute values ​​are ranked first, and the monitoring points ranked first are monitored first.

7. A sensor-based seasonal frozen soil monitoring system according to claim 6, characterized in that: The module for obtaining terrain influence factors includes: A division module is used to divide the monitored area into grids and deploy sensors at equal intervals in each grid, wherein the sensors include: temperature sensors, moisture sensors, soil moisture sensors and strain sensors; The module for obtaining geographic data is used to obtain the altitude, slope and slope direction of the sensor monitoring point; A terrain influence factor calculation module is used to calculate the terrain influence factors between sensor monitoring points based on the altitude, slope and slope direction of the sensor monitoring points; Among them, T ij represents the terrain influence factor between sensor monitoring points i and j, Δh ij It represents the elevation difference between sensor monitoring points i and j, Δs ij represents the slope difference between sensor monitoring points i and j, Δp ij represents the slope difference between sensor monitoring points i and j, H0 represents the characteristic elevation, S0 represents the characteristic slope, P0 represents the characteristic slope, k h Represents the elevation weight coefficient, k s represents the slope weight coefficient, k p Represents the aspect weight factor.

8. A sensor-based seasonal frozen soil monitoring system according to claim 6, characterized in that: The heat flux acquisition module comprises: The module for calculating the weighted average thermal conductivity of the sensor monitoring point i is used to calculate the weighted average thermal conductivity for each sensor monitoring point according to its depth information and the measurement data of the soil thermal conductivity sensor, and obtain the weighted average thermal conductivity of the sensor monitoring point i, wherein the weighted average thermal conductivity reflects the thermal conductivity performance of the soil at the point. Among them, λ i represents the weighted average thermal conductivity of monitoring point i, represents the soil thermal conductivity of the kth soil thermal conductivity sensor at monitoring point i, n i represents the number of sensors measuring soil thermal conductivity at monitoring point i, represents the weight of the k-th soil thermal conductivity sensor at monitoring point i, in, represents the depth of the kth sensor at monitoring point i, The module for calculating the average thermal conductivity between two points is used to calculate the average thermal conductivity between two points based on the weighted average thermal conductivity of adjacent monitoring points: Among them, λ ij represents the average thermal conductivity between sensor monitoring points i and j, λ j represents the weighted average thermal conductivity of monitoring point j; The heat flux calculation module is used to obtain the heat flux between the sensor monitoring point i and the sensor monitoring point j based on the average thermal conductivity between the two points. Among them, q ij represents the heat flux between monitoring points i and j, represents the average temperature gradient between monitoring points i and j, which can be calculated using the central difference method.

9. A sensor-based seasonal frozen soil monitoring system according to claim 6, characterized in that: The module for obtaining autocorrelation comprises: The module for obtaining the final weight matrix is ​​used to obtain the final weight matrix based on the terrain influence factor and the heat conduction factor, W ij =T ij ·(1+c·|q ij |), where c represents the weight factor; The quantification correlation module is used to quantify the correlation between each sensor monitoring point i and the surrounding points through the correlation model. Specifically, the correlation model is: Among them, x i represents the observation value of the i-th sensor, represents the average value of all sensor observations, s 2 represents the variance of the observations, W ij Represents the weight between the i-th point and the j-th point in the weight matrix.

10. A sensor-based seasonal frozen soil monitoring system according to claim 6, characterized in that: The screening module comprises: Screening module, used if I i is greater than the first preset threshold and |x i -∑ j W ij ·x j | is less than the second preset threshold, the sensor monitoring point is considered to be a redundant point and is removed; Prioritize monitoring module, otherwise it is considered that the point is not a redundant point, press I i Sort by absolute value, check I first i Larger points to obtain outliers.

Images