Desert sand fixation monitoring and management system based on data analysis

Abstract

The present invention discloses a desert sand fixation monitoring and management system based on data analysis, which belongs to the technical field of desert sand fixation data analysis. The Internet of Things sends collection instructions to sensors deployed in the implementation area to obtain monitoring data, performs integrity check and preliminary outlier test on the monitoring data, and then stores it in a cache area. New monitoring indicator characteristic values ​​are obtained through missing value processing, outlier processing and regional division, and a new monitoring indicator value set is constructed; the monitoring indicator value set is analyzed for spatiotemporal characteristics to locate abnormal sand fixation areas; the collection frequency of sensors in the abnormal sand fixation areas is adjusted, and collection instructions are sent through the Internet of Things technology to monitor the changing trend of ecological benefits in the abnormal sand fixation areas. The present application can efficiently collect, process and analyze monitoring data of sand fixation areas, locate abnormal areas based on the analysis results, and intelligently adjust the sensor collection frequency to optimize the monitoring process.

Description

Technical Field

[0001] The present invention relates to the technical field of desert sand fixation data analysis, and more particularly to a desert sand fixation monitoring and management system based on data analysis. Background Art

[0002] Traditional desert environmental monitoring relies primarily on manual field surveys and a small number of fixed-point monitoring devices. Manual monitoring not only consumes significant manpower, material resources, and time, but also has a limited monitoring range, making it difficult to fully cover the vast desert areas. Furthermore, the low frequency of manual monitoring prevents timely acquisition of real-time information on changes in the desert environment. In complex desert environments, monitoring equipment is susceptible to wind, sand, high temperatures, and low temperatures, leading to inaccurate data collection or equipment failure. Furthermore, due to relatively backward data transmission and processing methods, monitoring data often suffers from significant delays, making it difficult to provide timely and effective support for sand fixation decision-making.

[0003] While existing sand barrier monitoring technologies can improve efficiency and reduce costs, they still have shortcomings. These include: Incomplete monitoring, focusing solely on wind accumulation and erosion, and failing to comprehensively capture multi-dimensional ecological benefit data, including soil, meteorological, and vegetation data; poor real-time performance, relying on periodic drone operations, making it difficult to track the dynamics of sand fixation areas; and a lack of intelligent decision-making capabilities, preventing automated adjustments to strategies based on monitoring data. Furthermore, these technologies are unable to provide comprehensive decision-making support for sand fixation management and timely early warning responses to abnormal situations. To overcome these limitations, the present invention proposes a desert sand fixation monitoring and management system based on data analysis. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a desert sand fixation monitoring and management system based on data analysis, which solves the technical problem of how to efficiently collect, process and analyze the monitoring data of the sand fixation area, accurately locate the abnormal area based on the analysis results, and intelligently adjust the sensor acquisition frequency to optimize the monitoring. Through the data acquisition module, the sensor acquisition frequency is set according to the characteristics of different monitoring indicators, and instructions are sent to various sensors in the implementation area of ​​the sand fixation project to collect monitoring data. The data is stored in a cache area after integrity and outlier checks. When new data is added, missing values ​​and outliers are processed to construct a new monitoring indicator value set; the data analysis module performs spatiotemporal feature analysis on the stored indicator value set, including trend and mutation point identification of time series and autocorrelation analysis of spatial distribution, and locates the abnormal sand fixation area based on the analysis results; the decision feedback module adjusts the abnormal sensor acquisition frequency after monitoring the abnormal area, and feeds the adjusted frequency back to the data acquisition module. At the same time, the characteristic values ​​of the monitoring indicators of the abnormal area, the real-time data of the abnormal sensor and the acquisition frequency adjustment record are displayed through a visual interface. It can also carry out early warning and on-site inspection according to the proportion of abnormal sensors, so as to achieve effective monitoring and management of the desert sand fixation area.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A desert sand fixation monitoring and management system based on data analysis, including data acquisition module, data analysis module and decision feedback module;

[0007] The data acquisition module is used to collect monitoring data in the implementation area of ​​the sand fixation project. It sends collection instructions to sensors deployed in the implementation area through the Internet of Things to obtain monitoring data. The monitoring data is then stored in the cache area after integrity checks and preliminary outlier tests. When new monitoring data is added to the cache area, the characteristic values ​​of the new monitoring indicators are obtained through missing value processing, outlier processing and regional division. The new monitoring indicator value set is constructed and stored in the storage area of ​​the central management platform.

[0008] The data analysis module is used to analyze the spatiotemporal characteristics of the monitoring indicator value set in the storage area, identify abnormal sand consolidation areas, and through time series analysis, mine the temporal variation characteristics of each monitoring indicator characteristic value in the monitoring indicator value set to identify the temporal variation trend. Moreover, through spatial distribution analysis, the spatial distribution characteristics of the monitoring indicator characteristic values ​​are obtained to identify spatial abnormal trends. Based on the temporal variation trend and spatial abnormal trends, the abnormal sand consolidation areas are located.

[0009] The decision feedback module is used to adjust the collection frequency of sensors in the abnormal sand consolidation area, send collection instructions through the Internet of Things technology to obtain the characteristic values ​​of monitoring indicators in the abnormal sand consolidation area, and monitor the changing trend of ecological benefits in the abnormal sand consolidation area.

[0010] Specifically, the steps of spatiotemporal feature analysis include:

[0011] Setting an abnormality analysis interval for each monitoring indicator characteristic value, and filtering the monitoring indicator characteristic values ​​from the monitoring indicator value set in the storage area after each abnormality analysis interval, to construct an abnormality analysis set of the monitoring indicator characteristic values;

[0012] The anomaly analysis set is divided according to the time series to obtain the time series anomaly analysis set of the monitoring indicator characteristic values ​​of each sand consolidation area;

[0013] For the time series anomaly analysis set after missing value processing, identify the time series change trend, including trend identification and mutation point identification, and calculate the time series anomaly score of the time series anomaly analysis set;

[0014] The anomaly analysis set is divided according to the spatial sequence to obtain the spatial anomaly analysis set of the monitoring indicator characteristic values ​​at each moment;

[0015] For the spatial anomaly analysis set after missing value processing, perform spatial change trend identification, including autocorrelation analysis and mutation point identification, and calculate the spatial anomaly score of the spatial anomaly analysis set;

[0016] Configure the spatial anomaly threshold and temporal anomaly threshold. If the spatial anomaly score of the spatial anomaly analysis set is greater than the spatial anomaly threshold, the moment of the spatial anomaly analysis set is marked as a temporal anomaly moment. Otherwise, no processing is performed.

[0017] If the time series anomaly score of the time series anomaly analysis set is greater than the time series anomaly threshold, the sand-fixing area where the time series anomaly analysis set is located is marked as a spatial anomaly area;

[0018] In the implementation area of ​​the sand fixation project, the sand fixation areas with both temporal abnormal moment marks and spatial abnormal area marks are screened and marked as abnormal sand fixation areas.

[0019] Specifically, the steps for identifying time series change trends include:

[0020] Set a smoothing time window, perform moving average on the time series anomaly analysis set, obtain the moving average of each monitoring indicator characteristic value in the time series anomaly analysis set, calculate the moving deviation between each monitoring indicator characteristic value and its moving average, and calculate the standard deviation of the moving deviation;

[0021] Configure the deviation change threshold. If the standard deviation of the moving deviation of the time series anomaly analysis set is greater than the deviation change threshold, the time series anomaly analysis set is identified as a mutation point. Otherwise, no processing is performed.

[0022] Specifically, the steps of identifying mutation points in the time series anomaly analysis set include:

[0023] Configure the deviation anomaly threshold. If the moving deviation of the monitoring indicator characteristic value is greater than the deviation anomaly threshold, the monitoring indicator characteristic value is marked as a potential mutation point. Otherwise, no action is taken.

[0024] The relative change rate of each potential mutation point is calculated by the relative size of the monitoring indicator characteristic value at the potential mutation point relative to the monitoring indicator characteristic value at the adjacent time;

[0025] Configure a change rate threshold. If the relative change rate of a potential mutation point is greater than the change rate threshold, the point is marked as a mutation point, and the timestamp and time series anomaly analysis set of the mutation point are marked. Otherwise, no processing is performed.

[0026] The time series anomaly score of the time series anomaly analysis set is calculated by combining the standard deviation of the moving deviation of the characteristic values ​​of the monitoring indicators in the time series anomaly analysis set and the number of mutation points.

[0027] Specifically, the steps for identifying spatial change trends include:

[0028] Perform global spatial autocorrelation analysis on the spatial anomaly analysis set. Combined with the spatial location of the sand fixation area, obtain the difference between the monitoring indicator characteristic value and its mean value for each sand fixation area in the spatial anomaly analysis set, and calculate the overall spatial autocorrelation score of the monitoring indicator characteristic value of the spatial anomaly analysis set within the sand fixation area.

[0029] Configure the autocorrelation score threshold. If the overall spatial autocorrelation score of the monitoring indicator characteristic values ​​of the spatial anomaly analysis set in the sand fixation area is greater than the autocorrelation score threshold, no processing is performed. Otherwise, the spatial anomaly analysis set is identified as a sudden change point.

[0030] For each sand-fixing area, the local spatial autocorrelation coefficient is calculated and a correlation mutation threshold is configured. If the local spatial autocorrelation coefficient of the sand-fixing area is less than the correlation mutation threshold, the sand-fixing area is marked as a mutation area. Otherwise, no processing is performed.

[0031] The overall spatial autocorrelation score of the monitoring indicator eigenvalues ​​combined with the spatial anomaly analysis set , the local spatial autocorrelation coefficient of the sand-fixing area and the number of mutation areas in the spatial anomaly analysis set are used to calculate the spatial anomaly score of the spatial anomaly analysis set.

[0032] Specifically, the steps for collecting monitoring data include:

[0033] Set the sensor data collection frequency, and based on the data collection frequency, send collection instructions to sensors in the sand fixation project implementation area through the Internet of Things;

[0034] When the sensor receives the collection instruction, it measures the monitoring data, and the monitoring data is returned to the central management platform through the Internet of Things in the form of digital or analog signals;

[0035] After the central management platform marks the received monitoring data with sensor tags and timestamps, it checks the integrity of the monitoring data and determines whether there are missing values. If there are missing values ​​in the received monitoring data, the data acquisition module resends the acquisition instruction based on the sensor tags to obtain replacement monitoring data to fill the missing values.

[0036] Determine whether the substitute monitoring data is a missing value. If the substitute monitoring data is a missing value, issue a sensor warning; otherwise, do not perform any operation;

[0037] Perform a preliminary outlier check on the monitoring data after the integrity check. Set the sensor data range according to the sensor type. Filter the monitoring data outside the sensor data range based on the sensor tag and sensor data range of the monitoring data and mark it as abnormal monitoring data.

[0038] The monitoring data that have undergone integrity checks and preliminary outlier tests are classified according to sensor tags and stored in the cache area of ​​the central management platform.

[0039] Specifically, the monitoring indicators include soil monitoring indicators, meteorological monitoring indicators, vegetation monitoring indicators and sand fixation equipment monitoring indicators;

[0040] Specifically, the steps for obtaining the characteristic value of the monitoring indicator include:

[0041] When new monitoring data is added to the cache area, the new monitoring data of the corresponding sensor is selected according to the monitoring indicator type, a new monitoring data set is constructed, and the location coordinates of each sensor in the new monitoring data set are obtained;

[0042] Perform missing value processing on the newly added monitoring data set, filter out missing values ​​in the newly added monitoring data set, count the number of missing values, and calculate the missing value frequency of the newly added monitoring data set;

[0043] Configure a missing frequency threshold. If the missing value frequency of the newly added monitoring dataset exceeds the missing frequency threshold, the sensor fault troubleshooting process is initiated and the newly added monitoring dataset is set to an empty set. Otherwise, the missing values ​​of the newly added monitoring dataset are filled using spatial interpolation.

[0044] Perform outlier processing on the newly added monitoring data sets, filter the data points marked as abnormal monitoring data in each newly added monitoring data set, count the number of data points marked as abnormal monitoring data, and calculate the outlier frequency of the newly added monitoring data sets;

[0045] Configure an abnormal frequency threshold. If the frequency of abnormal values ​​in the newly added monitoring dataset is greater than the abnormal frequency threshold, cluster analysis is performed on the abnormal monitoring data to locate the abnormal area. Otherwise, the abnormal monitoring data in the newly added monitoring dataset is filled in through spatial interpolation.

[0046] Specifically, the step of obtaining the characteristic value of the monitoring indicator also includes:

[0047] The sand fixation area of ​​the sand fixation project implementation area is divided into sand fixation areas. According to the sensor location coordinates within the sand fixation area, the newly added monitoring data set after missing value processing and outlier processing is divided to obtain a new monitoring indicator set for each sand fixation area.

[0048] The newly added monitoring indicator set is divided according to the sensor type. A new monitoring indicator subset for each type of sensor is obtained. The mean of the monitoring data in each new monitoring indicator subset is calculated as the monitoring indicator characteristic value of the new monitoring indicator subset in the sand fixation area.

[0049] Obtain the monitoring indicator characteristic values ​​of each monitoring indicator in all sand fixation areas under each sensor type, build a new monitoring indicator value set for the sand fixation area, mark it with a timestamp, and store it in the storage area of ​​the central management platform.

[0050] Specifically, the steps for filling missing values ​​in the newly added monitoring dataset include:

[0051] Divide the newly added monitoring data set according to the sensor type to obtain the newly added monitoring data subset;

[0052] Get the location coordinates of the sensor where the missing value in the newly added monitoring data subset is located , configure the compensation radius, and divide the compensation area of ​​missing values ​​according to the compensation radius , filter the monitoring data of non-missing values ​​in the compensation area, and calculate the compensation monitoring data of missing values ​​based on inverse distance weighted interpolation.

[0053] Specifically, the steps of cluster analysis of abnormal monitoring data include:

[0054] Obtain the location coordinates of the sensor where the abnormal monitoring data is located in the newly added monitoring data set, set the number of clusters, and divide the sensor where the abnormal monitoring data is located into multiple clusters based on the cluster number using a clustering algorithm;

[0055] Count the number of abnormal monitoring data in each cluster, and calculate the abnormal concentration of the sensor where the abnormal monitoring data is located in each cluster based on the area of ​​the cluster;

[0056] Configure the concentration threshold. If the abnormal concentration of a cluster is greater than the concentration threshold, the area where the cluster is located is divided into an abnormal area, an abnormal warning is issued for the cluster in the abnormal area, and the abnormal monitoring data in the cluster in the abnormal area is set to null; otherwise, the abnormal monitoring data of the cluster is filled through spatial interpolation.

[0057] Specifically, the steps for adjusting the acquisition frequency of the sensor in the abnormal sand consolidation area include:

[0058] When an abnormal sand consolidation area is detected, the characteristic values ​​of the monitoring indicators at the abnormal moments in the abnormal sand consolidation area are screened, and the sensors where the characteristic values ​​of the monitoring indicators are located are obtained, marked as abnormal sensors, and the acquisition frequency of the abnormal sensors is obtained;

[0059] Configure an abnormal sensor threshold, count the number of abnormal sensors in the abnormal sand-fixing area, and calculate the abnormal sensor ratio. If the abnormal sensor ratio in the sand-fixing area is greater than the abnormal sensor threshold, issue a sand-fixing area warning; otherwise, no action is taken.

[0060] A set of adjustment coefficients is set based on the proportion of abnormal sensors in the abnormal sand-fixing area, and each adjustment coefficient is applied to the proportion of abnormal sensors in different abnormal sand-fixing areas;

[0061] Based on the adjustment coefficient, the collection frequency of the abnormal sensors in the abnormal sand consolidation area is adjusted according to the current collection frequency of the abnormal sensors in the abnormal sand consolidation area.

[0062] Beneficial effects of the present invention:

[0063] 1. A multi-layered data quality assurance system has been established from the very beginning of data collection. During data collection, sensor data is checked for integrity. If missing values ​​are discovered, a process for re-collecting replacement data is immediately initiated to ensure that no data is missing. Furthermore, preliminary outlier testing is performed to screen out abnormal monitoring data that falls outside the sensor data range, preventing erroneous data from entering subsequent processing steps. During data processing, missing and outliers in newly added monitoring data sets are addressed using scientific methods such as spatial interpolation and cluster analysis to ensure the accuracy and reliability of the final stored data. This enhances data reliability during system operation and provides a solid foundation for stable system operation.

[0064] 2. Through spatiotemporal feature analysis methods, we deeply explore the potential information contained in the concentration of monitoring indicator values. In terms of time series analysis, by setting smoothing time windows and identifying mutation points, we accurately capture the changing trends of the characteristic values ​​of monitoring indicators over time and calculate the time series anomaly score. In spatial distribution analysis, we use techniques such as global and local spatial autocorrelation analysis to accurately obtain the spatial distribution characteristics of the characteristic values ​​of monitoring indicators and calculate the spatial anomaly score. Based on these two anomaly scores, the system can accurately locate areas with abnormal sand fixation, providing a key basis for subsequent decision-making.

[0065] 3. Based on the proportion of abnormal sensors within the abnormal sand-fixing area, an adjustment coefficient is automatically set, intelligently adjusting the collection frequency of abnormal sensors. Taking into account the severity and scope of the anomaly, the sensor collection frequency is increased in areas with more severe anomalies to obtain more intensive monitoring data and timely understand the changing trends in ecological benefits. Furthermore, the adjusted collection frequency is fed back to the data acquisition module in real time, achieving a closed-loop intelligent decision-making process from anomaly monitoring to data collection strategy adjustment, enhancing the system's decision-making capabilities in complex situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a schematic structural diagram of a desert sand fixation monitoring and management system based on data analysis according to the present invention;

[0067] Figure 2 A flowchart of the specific steps of monitoring data collection of the present invention;

[0068] Figure 3 A flowchart of the specific steps of filling missing values ​​in a newly added monitoring data set according to the present invention;

[0069] Figure 4 Flowchart of the specific steps of cluster analysis of abnormal monitoring data in the present invention;

[0070] Figure 5 Flowchart of the specific steps of spatiotemporal feature analysis of the present invention;

[0071] Figure 6 A flowchart of the specific steps of identifying the temporal change trend of the present invention;

[0072] Figure 7 A flowchart of the specific steps of identifying spatial variation trends of the present invention;

[0073] Figure 8 The figure is a flow chart of the specific steps of adjusting the acquisition frequency of the present invention. DETAILED DESCRIPTION

[0074] See also Figure 1 ,This embodiment introduces a desert sand fixation monitoring and management system based on data analysis,,including a data acquisition module, a data analysis module and a decision feedback module;

[0075] The data acquisition module is used to collect monitoring data in the implementation area of ​​the sand fixation project. It sends collection instructions to sensors deployed in the implementation area through the Internet of Things to obtain monitoring data, and stores the monitoring data in the cache area after performing integrity checks and preliminary outlier tests. When new monitoring data is added to the cache area, the characteristic values ​​of the new monitoring indicators are obtained through missing value processing, outlier processing and regional division, and a new monitoring indicator value set is constructed and stored in the storage area of ​​the central management platform.

[0076] Sensors include: soil moisture sensor, soil nutrient sensor, temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, AI vision sensor and sand barrier displacement sensor. Monitoring data include: soil moisture data, soil nutrient data, temperature data, humidity data, wind speed data, wind direction data, vegetation coverage, vegetation growth rate and sand barrier displacement data; monitoring indicators include soil monitoring indicators, meteorological monitoring indicators, vegetation monitoring indicators and sand fixation equipment monitoring indicators. Soil monitoring indicators include soil moisture characteristics and soil nutrient characteristics; meteorological monitoring indicators include meteorological temperature characteristics, meteorological humidity characteristics, meteorological wind speed characteristics and meteorological wind direction characteristics; vegetation monitoring indicators include vegetation coverage characteristics and vegetation growth rate characteristics; sand fixation equipment monitoring indicators include sand barrier displacement characteristics;

[0077] See also Figure 2 Preferably, the specific steps of monitoring data collection include:

[0078] Set the sensor data collection frequency. The data collection module sends collection instructions to the sensor through the Internet of Things based on the data collection frequency. Set the sensor data collection frequency according to the characteristics and monitoring requirements of different monitoring indicators. For example: for sensors related to meteorological monitoring indicators, since they are greatly affected by weather and time factors and change more frequently, the data collection frequency can be set to once per minute to capture the rapid changes in meteorological conditions. For sensors related to soil monitoring indicators, their changes are relatively slow. They can be set to once an hour or once a day based on the evaporation and infiltration laws of soil moisture, as well as the consumption and replenishment cycle of soil nutrients. For sand barrier displacement sensors, considering the uncertainty of sand barriers being affected by wind and sand erosion and external forces, the collection frequency can be set to once every half hour so as to keep abreast of the stability of the sand barriers. For sensors related to vegetation monitoring indicators, such as AI visual sensors, monitor changes in the growth status of vegetation, set the collection frequency according to the growth rate of vegetation, and collect video data for several key time periods to obtain vegetation coverage and vegetation growth rate;

[0079] When the sensor receives the collection instruction, it measures the monitoring data, and the monitoring data is returned to the central management platform through the Internet of Things in the form of digital or analog signals;

[0080] After the central management platform marks the received monitoring data, including the sensor tag and the timestamp tag, it checks the integrity of the monitoring data and determines whether there are any missing values. If the received monitoring data has missing values, the data acquisition module resends the acquisition instruction based on the sensor tag to obtain alternative monitoring data to fill the missing values.

[0081] Determine whether the substitute monitoring data is a missing value. If the substitute monitoring data is a missing value, issue a sensor warning to check whether the sensor has a fault or communication interference, and retain the missing value; otherwise, do nothing.

[0082] Perform a preliminary outlier check on the monitoring data after the integrity check. Set the sensor data range according to the sensor type. Filter the monitoring data outside the sensor data range based on the sensor tag and sensor data range of the monitoring data and mark it as abnormal monitoring data.

[0083] The monitoring data that has undergone integrity check and preliminary outlier inspection will be classified and stored in the cache area of ​​the central management platform according to sensor tags.

[0084] Preferably, the specific steps of obtaining the characteristic value of the monitoring indicator include:

[0085] When new monitoring data is added to the cache area, the system selects the new monitoring data from relevant sensors based on the monitoring indicator type, constructs a new monitoring dataset, and obtains the location coordinates of each sensor in the new monitoring dataset. For example, for soil monitoring indicators, the system selects monitoring data from soil moisture sensors and soil nutrient sensors; for meteorological monitoring indicators, the system selects monitoring data from temperature, humidity, wind speed, and wind direction sensors. Simultaneously, the system quickly obtains the precise location coordinates of each sensor from the sensor information database to ensure that the data is closely associated with the geographic location.

[0086] Perform missing value processing on the newly added monitoring data set, filter out missing values ​​in the newly added monitoring data set, count the number of missing values, and calculate the missing value frequency of the newly added monitoring data set, that is:

[0087] ;

[0088] in, It is The frequency of missing values ​​in the newly added monitoring datasets, It is The number of missing values ​​in the newly added monitoring datasets, It is The number of monitoring data in the newly added monitoring data sets;

[0089] Configuring the Missing Frequency Threshold , if the missing value frequency of the new monitoring data set is Greater than the missing frequency threshold , indicating that there are large-scale anomalies in the sensor monitoring data of the newly added monitoring dataset, the sensor fault troubleshooting process is started and the newly added monitoring dataset is set to an empty set; otherwise, the missing values ​​of the newly added monitoring dataset are filled by spatial interpolation;

[0090] See also Figure 3 Specifically, the steps to fill the missing values ​​of the newly added monitoring dataset through spatial interpolation include:

[0091] Divide the newly added monitoring data set according to the sensor type to obtain the newly added monitoring data subset ,in, It is The first of the new monitoring datasets New monitoring data subsets, It is The first of the new monitoring datasets The first in the newly added monitoring data subset Monitoring data collected by sensors; It is The first of the new monitoring datasets The number of sensors in the newly added monitoring data subset is given in the figure. Taking the newly added soil monitoring data set as an example, it can be divided into the newly added soil moisture monitoring data subset and the newly added soil nutrient monitoring data subset.

[0092] Get the location coordinates of the sensor where the missing value in the newly added monitoring data subset is located , configure the compensation radius, and divide the compensation area of ​​missing values ​​according to the compensation radius , filter the monitoring data of non-missing values ​​in the compensation area, and calculate the compensation monitoring data of missing values ​​according to the inverse distance weighted interpolation, that is:

[0093] ;

[0094] ;

[0095] in, yes The first of the new monitoring datasets Compensatory monitoring data for missing values ​​in the newly added monitoring data subset, is the first Monitoring data collected by sensors The location coordinates of is the first Monitoring data collected by sensors The weighting coefficient of is the position coordinate and distance, It is the power of distance, and its value range is [1, 2].

[0096] Perform outlier processing on the newly added monitoring data set, filter the data points marked as abnormal monitoring data in each newly added monitoring data set, count the number of data points marked as abnormal monitoring data, and calculate the outlier frequency of the newly added monitoring data set, that is:

[0097] ;

[0098] in, It is The frequency of outliers in the newly added monitoring datasets, It is The number of abnormal monitoring data in the newly added monitoring data set;

[0099] Configuring the Abnormal Frequency Threshold , if the frequency of abnormal values ​​in the new monitoring data set is Greater than the abnormal frequency threshold , then cluster analysis is performed on the abnormal monitoring data to locate the abnormal area, otherwise the abnormal monitoring data of the newly added monitoring data set are filled by spatial interpolation;

[0100] Specifically, the specific steps of filling the abnormal monitoring data of the newly added monitoring dataset through spatial interpolation include:

[0101] Divide the newly added monitoring data set according to the sensor type to obtain the newly added monitoring data subset;

[0102] Obtain the location coordinates of the sensor where the abnormal monitoring data in the newly added monitoring data subset is located, configure the filling radius, divide the filling area of ​​the abnormal monitoring data according to the filling radius, filter the monitoring data of non-abnormal monitoring data in the filling area, and calculate the filling monitoring data of the abnormal monitoring data based on inverse distance weighted interpolation.

[0103] See also Figure 4 ,Specifically, the specific steps of cluster analysis of abnormal ,monitoring data include:

[0104] Get the location coordinates of the sensor where the abnormal monitoring data is located in the newly added monitoring data set, and set the number of clusters , the sensors where the abnormal monitoring data are located are divided into Clusters and obtain the coordinates of the cluster center;

[0105] Calculate the distance between the sensor where each abnormal monitoring data is located and the center of the cluster, and select the sensor distance farthest from the center of the cluster as the radius to calculate the area of ​​the cluster, that is:

[0106] ;

[0107] in, It is a cluster The area of ​​the region, It is a cluster The location coordinates of the sensor where the abnormal monitoring data is located, It is a cluster Center position coordinates, is the position coordinate and distance;

[0108] Count the number of abnormal monitoring data in each cluster, and calculate the abnormal concentration of the sensor where the abnormal monitoring data is located in each cluster according to the area of ​​the cluster, that is:

[0109] ;

[0110] in, It is a cluster Abnormal concentrations in It is a cluster The number of abnormal monitoring data in

[0111] Configure a concentration threshold. If the abnormal concentration of a cluster is greater than the concentration threshold, it means that the abnormal monitoring data in the cluster is clustered and there is a clustered anomaly. The area where the cluster is located is divided into an abnormal area, and an abnormal warning is issued for the cluster in the abnormal area. The abnormal monitoring data in the cluster in the abnormal area is set to null values. Otherwise, the abnormal monitoring data of the cluster is filled by spatial interpolation.

[0112] The sand fixation area of ​​the sand fixation project implementation area is divided into sand fixation areas. According to the location coordinates of the sensors in the sand fixation areas, the newly added monitoring data set after missing value processing and outlier processing is divided to obtain the newly added monitoring indicator set for each sand fixation area.

[0113] The newly added monitoring indicator set is divided according to the sensor type. A new monitoring indicator subset for each type of sensor is obtained. The mean of the monitoring data of each new monitoring indicator subset is calculated as the monitoring indicator characteristic value of the new monitoring indicator subset in the sand fixation area.

[0114] Obtain the characteristic values ​​of each monitoring indicator for each sensor type in all sand fixation areas, construct a new monitoring indicator value set for the sand fixation area, and store it in the storage area of ​​the central management platform after timestamping. The new monitoring indicator value set includes the monitoring indicator characteristic values ​​of each type of monitoring indicator in each sand fixation area for the newly added monitoring data in the cache area. The new monitoring indicator value set is timestamped according to the timestamp of the monitoring data in the new monitoring indicator set.

[0115] The data analysis module is used to analyze the spatiotemporal characteristics of the monitoring indicator value set in the storage area, identify abnormal sand consolidation areas, and through time series analysis, mine the temporal variation characteristics of each monitoring indicator characteristic value in the monitoring indicator value set to identify the temporal variation trend. Moreover, through spatial distribution analysis, the spatial distribution characteristics of the monitoring indicator characteristic values ​​are obtained to identify spatial abnormal trends. Based on the temporal variation trend and spatial abnormal trends, the abnormal sand consolidation areas are located.

[0116] See also Figure 5 Preferably, the specific steps of spatiotemporal feature analysis include:

[0117] Set the abnormal analysis interval for each monitoring indicator characteristic value. After each abnormal analysis interval, filter the monitoring indicator characteristic value from the monitoring indicator value set in the storage area to build an abnormal analysis set of monitoring indicator characteristic values. ,Right now:

[0118] ;

[0119] in, It is in Sand fixation area moments The characteristic value of the monitoring indicator at time is the abnormal analysis interval, which is used to measure the number of characteristic values ​​of the screening monitoring indicators. is the total number of sand-fixing areas, It is the final collection time of the characteristic values ​​of monitoring indicators when constructing the anomaly analysis set; according to the characteristics and actual needs of different monitoring indicators, a suitable anomaly analysis interval is set for each characteristic value of monitoring indicators to determine the collection time span of the screening characteristic values ​​of monitoring indicators.

[0120] The abnormal analysis set is divided according to the time series, and the time series abnormal analysis set of the monitoring indicator characteristic values ​​of each sand consolidation area is obtained. The time series anomaly analysis set of the sand consolidation area is ; Perform missing value interpolation on the time series anomaly analysis set. The missing value interpolation filling methods include linear interpolation, polynomial interpolation, and statistical feature filling;

[0121] For the time series anomaly analysis set after missing value processing, identify the time series change trend, including trend identification and mutation point identification, and calculate the time series anomaly score of the time series anomaly analysis set;

[0122] See also Figure 6 Specifically, the steps for identifying time series change trends include:

[0123] Set a smoothing time window, perform moving average on the time series anomaly analysis set, smooth the data fluctuations of the time series anomaly analysis set, obtain the moving average of the characteristic value of each monitoring indicator in the time series anomaly analysis set, calculate the moving deviation between the characteristic value of each monitoring indicator and its moving average, and calculate the standard deviation of the moving deviation to evaluate the trend stability of the time series anomaly analysis set;

[0124] Configure a deviation change threshold. If the standard deviation of the moving deviation of the time series anomaly analysis set is greater than the deviation change threshold, it indicates that the characteristic value of the monitoring indicator of the time series anomaly analysis set fluctuates greatly and there is an abnormal trend. In this case, mutation points are identified for the time series anomaly analysis set. Otherwise, no processing is performed.

[0125] Identify mutation points in the time series anomaly analysis set, including configuring a deviation anomaly threshold. If the moving deviation of the monitoring indicator characteristic value is greater than the deviation anomaly threshold, the monitoring indicator characteristic value is marked as a potential mutation point. Otherwise, no action is taken.

[0126] The relative change rate of each potential mutation point is calculated by the relative size of the monitoring indicator characteristic value at the potential mutation point relative to the monitoring indicator characteristic value at the adjacent time, that is:

[0127] ;

[0128] in, It is Sand fixation area moments The characteristic value of the monitoring indicator at Relative rate of change, is the maximum function;

[0129] Configure a change rate threshold. If the relative change rate of a potential mutation point is greater than the change rate threshold, the point is marked as a mutation point, and the timestamp and time series anomaly analysis set of the mutation point are marked. Otherwise, no processing is performed.

[0130] Combined with the standard deviation of the moving deviation of the characteristic value of the monitoring indicator in the time series anomaly analysis set and the number of mutation points, the time series anomaly score of the time series anomaly analysis set is calculated, that is:

[0131] ;

[0132] in, Is a time series anomaly analysis set The time series anomaly score of Is a time series anomaly analysis set The standard deviation of the moving deviation of the monitoring indicator characteristic value, Is a time series anomaly analysis set The number of mutation points, 、 They are all non-negative weighted coefficients, ranging from (0,1);

[0133] The abnormal analysis set is divided according to the spatial sequence, and the spatial abnormal analysis set of the monitoring indicator characteristic value at each moment is obtained. The spatial anomaly analysis set is ; Perform spatial interpolation filling on the spatial anomaly analysis set;

[0134] For the spatial anomaly analysis set after missing value processing, perform spatial change trend identification, including autocorrelation analysis and mutation point identification, and calculate the spatial anomaly score of the spatial anomaly analysis set;

[0135] See also Figure 7 ,Specifically, the specific steps for identifying spatial ,change trends include:

[0136] Perform global spatial autocorrelation analysis on the spatial anomaly analysis set, and calculate the overall spatial autocorrelation score of the monitoring indicator characteristic values ​​of the spatial anomaly analysis set within the sand fixation area based on the spatial location of the sand fixation area, namely:

[0137] ;

[0138] in, Is the spatial anomaly analysis set The overall spatial autocorrelation score of the monitoring indicator characteristic value, Is the spatial anomaly analysis set The mean of the characteristic values ​​of the monitoring indicators, Is the spatial anomaly analysis set Middle The first sand fixation area and the The non-negative weighted coefficient of the sand consolidation area is in the range of (0,1). It is The sand fixation area is at The characteristic value of the monitoring indicator at time The value range is , It is The sand fixation area is at The characteristic value of the monitoring indicator at time The value range is ;

[0139] Configure the autocorrelation score threshold. If the overall spatial autocorrelation score of the monitoring indicator characteristic values ​​of the spatial anomaly analysis set within the sand fixation area is greater than the autocorrelation score threshold, it indicates that the monitoring indicator characteristic values ​​are spatially clustered, there is strong spatial autocorrelation in the entire sand fixation area, and there is no mutation area. No processing is performed. Otherwise, mutation points are identified for the spatial anomaly analysis set.

[0140] For each sand consolidation area, calculate its local spatial autocorrelation coefficient. The local spatial autocorrelation coefficient of a sand-fixing area is:

[0141] ;

[0142] in, It is The local spatial autocorrelation coefficient of the sand-fixing area is Is the spatial anomaly analysis set Middle The first sand fixation area and the The non-negative weighted coefficient of each sand-fixing area is in the range of (0,1);

[0143] Configure the correlation mutation threshold. If the local spatial autocorrelation coefficient of the sand-fixing area is less than the correlation mutation threshold, it indicates that there is a large difference in the characteristic value of the monitoring index between the sand-fixing area and the surrounding sand-fixing areas. This indicates that the sand-fixing area is the area where the spatial mutation point is located. In this case, the sand-fixing area is marked as a mutation area. Otherwise, no processing is performed.

[0144] The overall spatial autocorrelation score of the monitoring indicator eigenvalues ​​combined with the spatial anomaly analysis set , the local spatial autocorrelation coefficient of the sand-fixing area and the number of mutation areas in the spatial anomaly analysis set, and calculate the spatial anomaly score of the spatial anomaly analysis set:

[0145] ;

[0146] in, Is the spatial anomaly analysis set The spatial anomaly score of It is The local spatial autocorrelation coefficient of the sand-fixing area is Is the spatial anomaly analysis set The number of mutation regions, 、 and They are all non-negative weighted coefficients, ranging from (0,1);

[0147] Configure the spatial anomaly threshold and temporal anomaly threshold. If the spatial anomaly score of the spatial anomaly analysis set is greater than the spatial anomaly threshold, the moment where the spatial anomaly analysis set is located is marked as a temporal anomaly moment. Otherwise, no processing is performed. If the temporal anomaly score of the temporal anomaly analysis set is greater than the temporal anomaly threshold, the sand-fixing area where the temporal anomaly analysis set is located is marked as a spatial anomaly area.

[0148] In the implementation area of ​​the sand fixation project, the sand fixation areas with both temporal abnormal moment marks and spatial abnormal area marks are screened and marked as abnormal sand fixation areas.

[0149] The decision-making feedback module is used to adjust the acquisition frequency of sensors in the abnormal sand consolidation area and send acquisition instructions through the Internet of Things technology to obtain the characteristic values ​​of monitoring indicators in the abnormal sand consolidation area and monitor the changing trend of ecological benefits in the abnormal sand consolidation area.

[0150] See also Figure 8 Preferably, the specific steps of adjusting the acquisition frequency of the sensor in the abnormal sand-fixing area include:

[0151] When an abnormal sand consolidation area is detected, the characteristic values ​​of the monitoring indicators at the abnormal moments in the abnormal sand consolidation area are screened, and the sensors where the characteristic values ​​of the monitoring indicators are located are obtained, marked as abnormal sensors, and the acquisition frequency of the abnormal sensors is obtained;

[0152] Configure an abnormal sensor threshold, count the number of abnormal sensors in the abnormal sand-fixing area, and calculate the abnormal sensor ratio. If the abnormal sensor ratio in the sand-fixing area is greater than the abnormal sensor threshold, a sand-fixing area warning is issued, indicating that there is a significant deviation in the sand-fixing area. The warning notifies staff to conduct on-site inspections; otherwise, no action is taken.

[0153] Set the adjustment coefficient set based on the proportion of abnormal sensors in the abnormal sand consolidation area ,in , each adjustment coefficient is applied to the proportion of abnormal sensors in different abnormal sand fixation areas, that is:

[0154] ;

[0155] in, is the adjustment coefficient, is the proportion of abnormal sensors in the abnormal sand solidification area, 、 and are all abnormal sensor ratio thresholds, and , set according to experimental results;

[0156] Based on the adjustment coefficient, the acquisition frequency of the abnormal sensors in the abnormal sand consolidation area is adjusted according to the current acquisition frequency of the abnormal sensors in the abnormal sand consolidation area, that is:

[0157] ;

[0158] in, It is the first The collection frequency after the abnormal sensor is adjusted, It is the first The maximum acquisition frequency of abnormal sensors, It is the first The current collection frequency of abnormal sensors, is the ceiling function, is the minimum function;

[0159] Feedback the adjusted acquisition frequency of the abnormal sensor to the data acquisition module, and the data acquisition module sends an acquisition instruction based on the adjusted acquisition frequency of the abnormal sensor to obtain the monitoring data of the abnormal sensor to obtain the characteristic value of the monitoring indicator;

[0160] The characteristic values ​​of monitoring indicators in abnormal sand-fixing areas are summarized and displayed in the form of charts through a visual interface, and the real-time monitoring data of abnormal sensors and the record information of the adjustment of the abnormal sensor acquisition frequency are given.

[0161] Working principle and its effect:

[0162] The desert sand fixation monitoring and management system based on data analysis realizes efficient monitoring and management of desert sand fixation areas through the collaborative work of data acquisition module, data analysis module and decision feedback module.

[0163] The data collection module sends collection instructions to various sensors based on the characteristics of the monitored indicators, such as minute-by-minute meteorological indicators, hourly or daily soil indicators, half-hourly sand barrier displacement, and timed vegetation growth rates. The sensors transmit the monitoring data to the central management platform, which tags and checks the data for integrity, handles missing and outliers, and stores it in a cache. When new data is added to the cache, a dataset is constructed, missing and outliers are handled, sand consolidation zones are divided, characteristic values ​​of the monitored indicators are calculated, and a set of monitored indicator values ​​is constructed for storage.

[0164] The data analysis module analyzes the spatiotemporal characteristics of stored monitoring indicator value sets. Anomaly analysis sets are constructed based on anomaly analysis intervals. After being divided by time series, missing value processing and trend and mutation point identification are performed to calculate temporal anomaly scores. After being divided by spatial series, spatial anomaly scores are calculated through spatial interpolation and autocorrelation analysis. Based on these scores and thresholds, areas with abnormal sand consolidation are identified.

[0165] After detecting an abnormal area, the decision-making feedback module screens the characteristic values ​​of the indicators at the time of the abnormality and the corresponding sensors. Based on the proportion of abnormal sensors, it determines whether to issue an early warning. Simultaneously, it adjusts the acquisition frequency of the abnormal sensors according to a set of adjustment coefficients and provides feedback to the data acquisition module. Finally, the characteristic values ​​of the indicators monitored in the abnormal area are displayed in a chart, providing real-time data from the abnormal sensors and a record of the acquisition frequency adjustment.

[0166] The following effects have been achieved: On the one hand, the system can effectively guarantee data quality, ensure integrity and accuracy from the source of data collection, promptly detect and process missing values ​​and outliers, improve data reliability, and at the same time, provide timely warnings and troubleshooting for sensor failures and large-scale anomalies to maintain stable system operation. On the other hand, through deep spatiotemporal analysis, it can accurately identify abnormal areas and moments, comprehensively assess the degree and scope of anomalies, and provide a scientific basis for sand fixation management. Finally, the system's intelligent decision-making ability can dynamically adjust the sensor collection frequency according to abnormal conditions to more accurately grasp the changing trends of ecological benefits in abnormal areas. At the same time, it visualizes information to facilitate intuitive viewing and decision-making by staff, greatly improving the efficiency and scientific nature of desert sand fixation monitoring and management, facilitating the implementation and management of desert sand fixation projects, and promoting the smooth progress of desert control work.

[0167] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. The desert sand fixation monitoring and management system based on data analysis is characterized by: Including data acquisition module, data analysis module and decision feedback module; The data acquisition module is used to collect monitoring data in the implementation area of ​​the sand fixation project, send collection instructions to sensors deployed in the implementation area through the Internet of Things, obtain monitoring data, perform integrity checks and preliminary outlier tests on the monitoring data, and then store them in the cache area; and when new monitoring data is added to the cache area, obtain new monitoring indicator feature values ​​through missing value processing, outlier processing and regional division, and construct a new monitoring indicator value set and store it in the storage area of ​​the central management platform; The data analysis module is used to perform spatiotemporal feature analysis on the monitoring indicator value set in the storage area, identify abnormal sand consolidation areas, mine the time-varying features of each monitoring indicator characteristic value in the monitoring indicator value set through time series analysis, identify the temporal variation trend, and obtain the spatial distribution features of the monitoring indicator characteristic value through spatial distribution analysis, identify the spatial abnormal trend, and locate the abnormal sand consolidation areas based on the temporal variation trend and the spatial abnormal trend; The specific steps of the spatiotemporal feature analysis include: Setting an abnormality analysis interval for each monitoring indicator characteristic value, and filtering the monitoring indicator characteristic values ​​from the monitoring indicator value set in the storage area after each abnormality analysis interval, to construct an abnormality analysis set of the monitoring indicator characteristic values; The anomaly analysis set is divided according to the time series to obtain the time series anomaly analysis set of the monitoring indicator characteristic values ​​of each sand consolidation area; For the time series anomaly analysis set after missing value processing, identify the time series change trend, including trend identification and mutation point identification, and calculate the time series anomaly score of the time series anomaly analysis set; The anomaly analysis set is divided according to the spatial sequence to obtain the spatial anomaly analysis set of the monitoring indicator characteristic values ​​at each moment; For the spatial anomaly analysis set after missing value processing, perform spatial change trend identification, including autocorrelation analysis and mutation point identification, and calculate the spatial anomaly score of the spatial anomaly analysis set; Configure the spatial anomaly threshold and temporal anomaly threshold. If the spatial anomaly score of the spatial anomaly analysis set is greater than the spatial anomaly threshold, the moment of the spatial anomaly analysis set is marked as a temporal anomaly moment. Otherwise, no processing is performed. If the time series anomaly score of the time series anomaly analysis set is greater than the time series anomaly threshold, the sand-fixing area where the time series anomaly analysis set is located is marked as a spatial anomaly area; In the implementation area of ​​the sand fixation project, screen the sand fixation areas with both time series abnormal moment marks and spatial abnormal area marks, and mark them as abnormal sand fixation areas; The decision feedback module is used to adjust the collection frequency of sensors in the abnormal sand-fixing sand area, send collection instructions through the Internet of Things technology to obtain the characteristic values ​​of monitoring indicators in the abnormal sand-fixing sand area, and monitor the changing trend of ecological benefits in the abnormal sand-fixing sand area.

2. The desert sand fixation monitoring and management system based on data analysis according to claim 1, characterized in that: The specific steps of identifying the time series change trend include: Set a smoothing time window, perform moving average on the time series anomaly analysis set, obtain the moving average of each monitoring indicator characteristic value in the time series anomaly analysis set, calculate the moving deviation between each monitoring indicator characteristic value and its moving average, and calculate the standard deviation of the moving deviation; Configure the deviation change threshold. If the standard deviation of the moving deviation of the time series anomaly analysis set is greater than the deviation change threshold, the time series anomaly analysis set is identified as a mutation point. Otherwise, no processing is performed. The step of identifying mutation points on the time series anomaly analysis set includes: Configure the deviation anomaly threshold. If the moving deviation of the monitoring indicator characteristic value is greater than the deviation anomaly threshold, the monitoring indicator characteristic value is marked as a potential mutation point. Otherwise, no action is taken. The relative change rate of each potential mutation point is calculated by the relative size of the monitoring indicator characteristic value at the potential mutation point relative to the monitoring indicator characteristic value at the adjacent time; Configure a change rate threshold. If the relative change rate of a potential mutation point is greater than the change rate threshold, the point is marked as a mutation point, and the timestamp and time series anomaly analysis set of the mutation point are marked. Otherwise, no processing is performed. The time series anomaly score of the time series anomaly analysis set is calculated by combining the standard deviation of the moving deviation of the characteristic values ​​of the monitoring indicators in the time series anomaly analysis set and the number of mutation points.

3. The desert sand fixation monitoring and management system based on data analysis according to claim 1, characterized in that: The specific steps of identifying the spatial change trend include: Perform global spatial autocorrelation analysis on the spatial anomaly analysis set. Combined with the spatial location of the sand fixation area, obtain the difference between the monitoring indicator characteristic value and its mean value for each sand fixation area in the spatial anomaly analysis set, and calculate the overall spatial autocorrelation score of the monitoring indicator characteristic value of the spatial anomaly analysis set within the sand fixation area. Configure the autocorrelation score threshold. If the overall spatial autocorrelation score of the monitoring indicator characteristic values ​​of the spatial anomaly analysis set in the sand fixation area is greater than the autocorrelation score threshold, no processing is performed. Otherwise, the spatial anomaly analysis set is identified as a sudden change point. For each sand-fixing area, the local spatial autocorrelation coefficient is calculated and a correlation mutation threshold is configured. If the local spatial autocorrelation coefficient of the sand-fixing area is less than the correlation mutation threshold, the sand-fixing area is marked as a mutation area. Otherwise, no processing is performed. The overall spatial autocorrelation score of the monitoring indicator eigenvalues ​​combined with the spatial anomaly analysis set , the local spatial autocorrelation coefficient of the sand-fixing area and the number of mutation areas in the spatial anomaly analysis set are used to calculate the spatial anomaly score of the spatial anomaly analysis set.

4. The desert sand fixation monitoring and management system based on data analysis according to claim 1, characterized in that: The monitoring data collection steps include: Set the sensor data collection frequency, and based on the data collection frequency, send collection instructions to sensors in the sand fixation project implementation area through the Internet of Things; When the sensor receives the collection instruction, it measures the monitoring data, and the monitoring data is returned to the central management platform through the Internet of Things in the form of digital or analog signals; After the central management platform performs sensor tagging and time stamping on the received monitoring data, it checks the integrity of the monitoring data and determines whether there are missing values. If there are missing values ​​in the received monitoring data, the data acquisition module resends the acquisition instruction based on the sensor tag to obtain replacement monitoring data to fill the missing values. Determine whether the substitute monitoring data is a missing value. If the substitute monitoring data is a missing value, issue a sensor warning; otherwise, do not perform any operation; Perform a preliminary outlier check on the monitoring data after the integrity check. Set the sensor data range according to the sensor type. Filter the monitoring data outside the sensor data range based on the sensor tag and sensor data range of the monitoring data and mark it as abnormal monitoring data. The monitoring data that have undergone integrity checks and preliminary outlier tests are classified according to sensor tags and stored in the cache area of ​​the central management platform.

5. The desert sand fixation monitoring and management system based on data analysis according to claim 1, characterized in that: The monitoring indicators include soil monitoring indicators, meteorological monitoring indicators, vegetation monitoring indicators and sand fixation equipment monitoring indicators; The step of obtaining the characteristic value of the monitoring indicator includes: When new monitoring data is added to the cache area, the new monitoring data of the corresponding sensor is selected according to the monitoring indicator type, a new monitoring data set is constructed, and the location coordinates of each sensor in the new monitoring data set are obtained; Perform missing value processing on the newly added monitoring data set, filter out missing values ​​in the newly added monitoring data set, count the number of missing values, and calculate the missing value frequency of the newly added monitoring data set; Configure a missing frequency threshold. If the missing value frequency of the newly added monitoring dataset exceeds the missing frequency threshold, the sensor fault troubleshooting process is initiated and the newly added monitoring dataset is set to an empty set. Otherwise, the missing values ​​of the newly added monitoring dataset are filled using spatial interpolation. Perform outlier processing on the newly added monitoring data sets, filter the data points marked as abnormal monitoring data in each newly added monitoring data set, count the number of data points marked as abnormal monitoring data, and calculate the outlier frequency of the newly added monitoring data sets; Configure an abnormal frequency threshold. If the frequency of abnormal values ​​in the newly added monitoring dataset is greater than the abnormal frequency threshold, cluster analysis is performed on the abnormal monitoring data to locate the abnormal area. Otherwise, the abnormal monitoring data in the newly added monitoring dataset is filled in through spatial interpolation.

6. The desert sand fixation monitoring and management system based on data analysis according to claim 5, characterized in that: The step of obtaining the characteristic value of the monitoring indicator further includes: The sand fixation area of ​​the sand fixation project implementation area is divided into sand fixation areas. According to the sensor location coordinates within the sand fixation area, the newly added monitoring data set after missing value processing and outlier processing is divided to obtain a new monitoring indicator set for each sand fixation area. The newly added monitoring indicator set is divided according to the sensor type. A new monitoring indicator subset for each type of sensor is obtained. The mean of the monitoring data in each new monitoring indicator subset is calculated as the monitoring indicator characteristic value of the new monitoring indicator subset in the sand fixation area. Obtain the monitoring indicator characteristic values ​​of each monitoring indicator in all sand fixation areas under each sensor type, build a new monitoring indicator value set for the sand fixation area, mark it with a timestamp, and store it in the storage area of ​​the central management platform.

7. The desert sand fixation monitoring and management system based on data analysis according to claim 5, characterized in that: The step of filling missing values ​​in the newly added monitoring data set includes: Divide the newly added monitoring data set according to the sensor type to obtain the newly added monitoring data subset; Get the location coordinates of the sensor where the missing value in the newly added monitoring data subset is located , configure the compensation radius, and divide the compensation area of ​​missing values ​​according to the compensation radius , filter the monitoring data of non-missing values ​​in the compensation area, and calculate the compensation monitoring data of missing values ​​based on inverse distance weighted interpolation.

8. The desert sand fixation monitoring and management system based on data analysis according to claim 5, characterized in that: The step of performing cluster analysis on abnormal monitoring data includes: Obtain the location coordinates of the sensor where the abnormal monitoring data is located in the newly added monitoring data set, set the number of clusters, and divide the sensor where the abnormal monitoring data is located into multiple clusters based on the cluster number using a clustering algorithm; Count the number of abnormal monitoring data in each cluster, and calculate the abnormal concentration of the sensor where the abnormal monitoring data is located in each cluster based on the area of ​​the cluster; Configure the concentration threshold. If the abnormal concentration of a cluster is greater than the concentration threshold, the area where the cluster is located is divided into an abnormal area, an abnormal warning is issued for the cluster in the abnormal area, and the abnormal monitoring data in the cluster in the abnormal area is set to null; otherwise, the abnormal monitoring data of the cluster is filled through spatial interpolation.

9. The desert sand fixation monitoring and management system based on data analysis according to claim 1, characterized in that: The step of adjusting the acquisition frequency of the sensor in the abnormal sand solidification area includes: When an abnormal sand consolidation area is detected, the characteristic values ​​of the monitoring indicators at the abnormal moments in the abnormal sand consolidation area are screened, and the sensors where the characteristic values ​​of the monitoring indicators are located are obtained, marked as abnormal sensors, and the acquisition frequency of the abnormal sensors is obtained; Configure an abnormal sensor threshold, count the number of abnormal sensors in the abnormal sand-fixing area, and calculate the abnormal sensor ratio. If the abnormal sensor ratio in the sand-fixing area is greater than the abnormal sensor threshold, issue a sand-fixing area warning; otherwise, no action is taken. A set of adjustment coefficients is set based on the proportion of abnormal sensors in the abnormal sand-fixing area, and each adjustment coefficient is applied to the proportion of abnormal sensors in different abnormal sand-fixing areas; Based on the adjustment coefficient, the collection frequency of the abnormal sensors in the abnormal sand consolidation area is adjusted according to the current collection frequency of the abnormal sensors in the abnormal sand consolidation area.

Images