Miniature environment monitoring system and method based on big data

Abstract

The invention discloses a miniature environment monitoring system and method based on big data, and relates to the technical field of big data, the system comprises a database module, a relation analysis module, a change range module and a new data judgment module; the database module is used for collecting historical data of each monitoring pile sensor and the distance between the monitoring piles and forming a database; the relation analysis module is used for analyzing the influence relation between different sensors on the same monitoring pile and performing decorrelation processing; the change range module is used for calculating sensor parameter difference, establishing a regression model to fit a relationship between parameter change and distance, and determining a parameter change range; and the new data judgment module is used for judging the abnormal condition of the new data in combination with the relationship between the parameter change and the distance. The invention also provides a method for realizing the system. The invention can effectively improve the situation that the measurement result is influenced by the mutual influence of sensor measurement factors in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and in particular to a micro-environment monitoring system and method based on big data. Background Art

[0002] With the acceleration of industrialization and urbanization, environmental problems have become increasingly prominent, and accurate and real-time monitoring of the environment has become essential. Micro-environmental detection can monitor environmental parameters in a specific area at a micro level and provide more accurate information on environmental conditions. Since micro-sensors usually have the characteristics of rapid response, they can monitor environmental changes in real time and provide timely feedback. This is crucial for responding to sudden environmental events and taking timely measures to protect public health and the ecological environment.

[0003] In existing micro-environment monitoring technologies, when multiple micro-sensors are arranged in the same monitoring area, the sensor measurement factors may interfere with each other, affecting the accuracy of the measurement results; in addition, the existing technology lacks a unified data processing platform and scientific analysis methods, making it difficult to fully tap the value of multi-sensor data. Summary of the invention

[0004] The purpose of the present invention is to provide a micro-environment monitoring system and method based on big data to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solutions: a micro-environment monitoring system based on big data, the system comprising a database module, a relationship analysis module, a change range module, and a new data judgment module;

[0006] The database module is used to collect the historical data of each monitoring pile sensor and the distance between the monitoring piles and form a database; the relationship analysis module is used to analyze the influence relationship between different sensors on the same monitoring pile and perform decorrelation processing; the variation range module is used to calculate the sensor parameter difference, establish a regression model to fit the relationship between parameter change and distance, and determine the parameter variation range; the new data judgment module is used to judge the abnormality of the new data based on the relationship between parameter change and distance;

[0007] The output end of the database module is connected to the input end of the relationship analysis module and the change range module; the output end of the relationship analysis module is connected to the input end of the new data judgment module; the output end of the change range module is connected to the input end of the new data judgment module.

[0008] The database module includes a data collection unit and a database unit;

[0009] The data collection unit is used to collect historical data of different sensors and distance information between monitoring piles; the database unit is used to form a database based on the collected information and update it in real time;

[0010] The output end of the data collection unit is connected to the input end of the database unit; the output end of the database unit is connected to the input ends of the relationship analysis module and the variation range module.

[0011] The relationship analysis module includes a correlation analysis unit and a decorrelation processing unit;

[0012] The correlation analysis unit is used to calculate the Pearson correlation coefficient between sensors; the decorrelation processing unit is used to perform covariance analysis and feature decomposition on sensor data and generate principal components;

[0013] The output end of the correlation analysis unit is connected to the input end of the decorrelation processing unit; the output end of the decorrelation processing unit is connected to the input end of the new data judgment module.

[0014] The variation range module includes a difference calculation unit, a regression analysis unit and a range determination unit;

[0015] The difference calculation unit is used to calculate the difference between the sensor parameters corresponding to different monitoring piles; the regression analysis unit is used to fit the parameter difference and distance relationship using a polynomial regression model; the range determination unit is used to determine the sensor parameter variation range according to the difference and distance relationship;

[0016] The output end of the difference calculation unit is connected to the input end of the regression analysis unit; the output end of the regression analysis unit is connected to the input end of the range determination unit; the output end of the range determination unit is connected to the input end of the new data judgment module.

[0017] The new data judgment module includes a new data acquisition unit and a change judgment unit;

[0018] The new data acquisition unit is used to acquire new parameters of the sensor; the change judgment unit is used to judge whether the parameter is abnormal according to the difference change of the sensor parameter and the change range;

[0019] The output end of the new data acquisition unit is connected to the input end of the change judgment unit.

[0020] A micro-environment monitoring method based on big data, the method comprising the following steps:

[0021] S1. Build a historical database, including the sensor parameters in the historical data and the distances between different monitoring piles;

[0022] S2, analyzing the influence relationship between different sensors of the same monitoring pile and performing decorrelation processing;

[0023] S3, determining the parameter change range according to the distance between the monitoring piles and the corresponding sensor parameter changes in the historical data;

[0024] S4. Decorrelate the newly collected data of the monitoring pile sensors to determine whether the difference between the sensors of different monitoring piles is within the variation range; and determine whether the corresponding sensor parameters of different monitoring piles are abnormal based on the determination results.

[0025] In step S1, the monitoring piles are provided with a plurality of sensors, and the types and quantities of sensors provided on different monitoring piles are consistent, and the monitoring piles are arranged at different positions to monitor the environment;

[0026] The monitoring pile is represented by [Z 1 ,Z 2 ,…,Z m ]; the sensor is represented by [C 1 ,C 2 ,…,C n ]; Monitoring pile Z i and Z j The distance between them is represented by D ij ;

[0027] Among them, m is a positive integer, indicating the number of monitoring piles, Z m represents the mth monitoring pile; n is a positive integer, indicating the number of sensors on each monitoring pile, C n represents the nth sensor; i,j∈{1,2,…,m}, represents the sequence of monitoring piles;

[0028] The format of the historical database is [sample ID, Z 1 (C 1 ),…Z 1 (C n ),…,Z m (C 1 ),Z m (C n ),D 12 ,…,D 1m ,D 23 ,…D 2m ,…,D (m-1)m ];

[0029] Among them, Z m (C n ) represents the parameter monitored by the nth sensor of the mth monitoring pile.

[0030] In step S2, the datasets of all sensors on each monitoring pile are extracted from the historical dataset [sample ID, Z 1 (C 1 ),…Z 1 (C n ),…,Z m (C 1 ),Z m (C n )];

[0031] For a sensor C on a specified monitoring pile a The corresponding historical data is represented by H a , the influence relationship is measured by calculating the Pearson correlation coefficient between it and other sensor data: r Ha,Hb =(Σ(H a -H a mean )(H b -H b mean )) / ((Σ(H a -H a mean) 2 ) 1 / 2 ×(Σ(H b -H b mean ) 2 ) 1 / 2 );

[0032] The Pearson correlation coefficient range is [-1,1], close to 1 or -1 indicates a high correlation, close to 0 indicates no correlation;

[0033] Where a is a positive integer, a∈[1,n]; H a and H b are the measurement values ​​of different sensors on the same monitoring pile, H a mean and H b mean are their means respectively;

[0034] Each monitoring pile builds a correlation matrix between sensors, and the matrix elements represent the correlation between different sensors. When the element value is close to -1 or 1, it means that the data of the corresponding two sensors are highly correlated, and there is a high possibility of influence.

[0035] Decorrelate the sensor data:

[0036] Construct the covariance matrix: Σ ab =(1 / (N-1))Σ q＝1 N (H a (q) -Ha mean )(H b (q) -H b mean );

[0037] Where N is a positive integer, indicating the number of samples; q is a positive integer, q≤N, indicating the sample sequence; H a (q) Indicates that the qth sample corresponds to a specified monitoring pile sensor C a data;

[0038] Perform eigendecomposition on the covariance matrix to obtain eigenvalues ​​and eigenvectors; the eigenvalue represents the variance explained by each principal component, and the eigenvector is used to define the new principal component axis: ΣV = λV; according to the eigenvalue, select the first k principal components with the largest explained variance; by projecting the original sensor data into the principal component space, obtain the decorrelated sensor data: H new =V T H;

[0039] Where V represents the eigenvector; λ represents the eigenvalue; k is a positive integer; H new Represents the data after decorrelation processing.

[0040] In step S3, for any two monitoring piles Z i and Z j The distance between them is represented by D ij ; Monitoring pile Z i and Z j Same type sensor C a The parameter difference between them is expressed as: ΔZ ij (C a )=|Z i (C a )-Z j (C a )|;

[0041] Fitting the relationship between sensor parameter changes and distance: Using the distance matrix D in historical data ij and the corresponding sensor parameter difference ΔZ ij (C a ), and use the polynomial regression model to fit the relationship function between parameter difference and distance: ΔZ ij (C a )=γ a D ij u +α a D ij +β a ; According to the least squares method, the parameter γ is fitted a , αa and β a ;

[0042] Among them, γ a , α a It is sensor C a The coefficient of the parameter variation with distance, β a is a constant term;

[0043] According to the result of the fitting function, the sensor C between different monitoring piles is determined. a The parameter variation range; calculate the difference between the fitted value and the actual observed value in the historical data to determine the upper and lower limits of the variation range: ΔZ ij (C a )∈[γ a D ij u +α a D ij +β a -Kσ a ,γ a D ij u +α a D ij +β a +Kσ a ];

[0044] Among them, σ a represents the standard deviation of the difference between the fitted value and the actual observed value; K represents the adjustment parameter.

[0045] In step S4, new data collection: collect new data from each sensor on each monitoring pile, and the data format is consistent with the historical data;

[0046] Decorrelation processing: decorrelation processing is performed on the new data using the principal component matrix obtained in step S2;

[0047] Calculate the difference: For any two sensors of the same type on the monitoring piles, calculate the difference between their newly collected data:

[0048] Determine the variation range: using the parameter variation range determined in step S3, by comparing the difference with the previously determined variation range;

[0049] Analysis and judgment results: Normal change: If there is no range, it is considered that the sensor parameter changes are normal; Abnormal change: If it exceeds the range, it means that there is an abnormality and feedback is given.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. The present invention clarifies the influence of different sensors on data results by analyzing the correlation between sensors, thereby improving the credibility of the data;

[0052] 2. The present invention constructs a historical database and uses big data analysis to determine a reasonable parameter variation range, so that the data of different monitoring piles can be kept consistent to a certain extent;

[0053] 3. The accurate, consistent and reliable monitoring data and scientific analysis results provided by the present invention can provide a precise basis for decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a schematic diagram of the structure of a micro-environment monitoring system based on big data of the present invention;

[0055] Figure 2 The figure is a schematic diagram of the steps of a micro-environment monitoring method based on big data of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not 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.

[0057] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a micro-environment monitoring system based on big data, the system includes a database module, a relationship analysis module, a change range module, and a new data judgment module;

[0058] The database module is used to collect the historical data of each monitoring pile sensor and the distance between the monitoring piles and form a database; the relationship analysis module is used to analyze the influence relationship between different sensors on the same monitoring pile and perform decorrelation processing; the variation range module is used to calculate the sensor parameter difference, establish a regression model to fit the relationship between parameter change and distance, and determine the parameter variation range; the new data judgment module is used to judge the abnormality of the new data based on the relationship between parameter change and distance;

[0059] The output end of the database module is connected to the input end of the relationship analysis module and the change range module; the output end of the relationship analysis module is connected to the input end of the new data judgment module; the output end of the change range module is connected to the input end of the new data judgment module.

[0060] The database module includes a data collection unit and a database unit;

[0061] The data collection unit is used to collect historical data of different sensors and distance information between monitoring piles; the database unit is used to form a database based on the collected information and update it in real time;

[0062] The output end of the data collection unit is connected to the input end of the database unit; the output end of the database unit is connected to the input ends of the relationship analysis module and the variation range module.

[0063] The relationship analysis module includes a correlation analysis unit and a decorrelation processing unit;

[0064] The correlation analysis unit is used to calculate the Pearson correlation coefficient between sensors; the decorrelation processing unit is used to perform covariance analysis and feature decomposition on sensor data and generate principal components;

[0065] The output end of the correlation analysis unit is connected to the input end of the decorrelation processing unit; the output end of the decorrelation processing unit is connected to the input end of the new data judgment module.

[0066] The variation range module includes a difference calculation unit, a regression analysis unit and a range determination unit;

[0067] The difference calculation unit is used to calculate the difference between the sensor parameters corresponding to different monitoring piles; the regression analysis unit is used to fit the parameter difference and distance relationship using a polynomial regression model; the range determination unit is used to determine the sensor parameter variation range according to the difference and distance relationship;

[0068] The output end of the difference calculation unit is connected to the input end of the regression analysis unit; the output end of the regression analysis unit is connected to the input end of the range determination unit; the output end of the range determination unit is connected to the input end of the new data judgment module.

[0069] The new data judgment module includes a new data acquisition unit and a change judgment unit;

[0070] The new data acquisition unit is used to acquire new parameters of the sensor; the change judgment unit is used to judge whether the parameter is abnormal according to the difference change of the sensor parameter and the change range;

[0071] The output end of the new data acquisition unit is connected to the input end of the change judgment unit.

[0072] A micro-environment monitoring method based on big data, the method comprising the following steps:

[0073] S1. Build a historical database, including the sensor parameters in the historical data and the distances between different monitoring piles;

[0074] S2, analyzing the influence relationship between different sensors of the same monitoring pile and performing decorrelation processing;

[0075] S3, determining the parameter change range according to the distance between the monitoring piles and the corresponding sensor parameter changes in the historical data;

[0076] S4. Decorrelate the newly collected data of the monitoring pile sensors to determine whether the difference between the sensors of different monitoring piles is within the variation range; and determine whether the corresponding sensor parameters of different monitoring piles are abnormal based on the determination results.

[0077] In step S1, the monitoring piles are provided with a plurality of sensors, and the types and quantities of sensors provided on different monitoring piles are consistent, and the monitoring piles are arranged at different positions to monitor the environment;

[0078] The monitoring pile is represented by [Z 1 ,Z 2 ,…,Z m ]; the sensor is represented by [C 1 ,C 2 ,…,C n ]; Monitoring pile Z i and Z j The distance between them is represented by D ij ;

[0079] Among them, m is a positive integer, indicating the number of monitoring piles, Z m represents the mth monitoring pile; n is a positive integer, indicating the number of sensors on each monitoring pile, C n represents the nth sensor; i,j∈{1,2,…,m}, represents the sequence of monitoring piles;

[0080] The format of the historical database is [sample ID, Z 1 (C 1 ),…Z 1 (C n ),…,Z m (C 1 ),Z m (C n ),D 12 ,…,D 1m ,D 23 ,…D 2m ,…,D (m-1)m ];

[0081] Among them, Z m (C n ) represents the parameter monitored by the nth sensor of the mth monitoring pile.

[0082] In step S2, the datasets of all sensors on each monitoring pile are extracted from the historical dataset [sample ID, Z 1 (C 1 ),…Z 1 (C n ),…,Z m (C 1 ),Z m (C n )];

[0083] For a sensor C on a specified monitoring pile a The corresponding historical data is represented by H a , the influence relationship is measured by calculating the Pearson correlation coefficient between it and other sensor data: r Ha,Hb =(Σ(H a -H a mean )(H b -H b mean )) / ((Σ(H a -H a mean) 2 ) 1 / 2 ×(Σ(H b -H b mean ) 2 ) 1 / 2 );

[0084] The Pearson correlation coefficient range is [-1,1], close to 1 or -1 indicates a high correlation, close to 0 indicates no correlation;

[0085] Where a is a positive integer, a∈[1,n]; H a and H b are the measurement values ​​of different sensors on the same monitoring pile, H a mean and H b mean are their means respectively;

[0086] Each monitoring pile builds a correlation matrix between sensors, and the matrix elements represent the correlation between different sensors. When the element value is close to -1 or 1, it means that the data of the corresponding two sensors are highly correlated, and there is a high possibility of influence.

[0087] Decorrelate the sensor data:

[0088] Construct the covariance matrix: Σ ab =(1 / (N-1))Σ q＝1 N (H a (q) -Ha mean )(H b (q) -H b mean );

[0089] Where N is a positive integer, indicating the number of samples; q is a positive integer, q≤N, indicating the sample sequence; H a (q) Indicates that the qth sample corresponds to a specified monitoring pile sensor C a data;

[0090] Perform eigendecomposition on the covariance matrix to obtain eigenvalues ​​and eigenvectors; the eigenvalue represents the variance explained by each principal component, and the eigenvector is used to define the new principal component axis: ΣV = λV; according to the eigenvalue, select the first k principal components with the largest explained variance; by projecting the original sensor data into the principal component space, obtain the decorrelated sensor data: H new =V T H;

[0091] Where V represents the eigenvector; λ represents the eigenvalue; k is a positive integer; H new Represents the data after decorrelation processing.

[0092] In step S3, for any two monitoring piles Z i and Z j The distance between them is represented by D ij ; Monitoring pile Z i and Z j Same type sensor C a The parameter difference between them is expressed as: ΔZ ij (C a )=|Z i (C a )-Z j (C a )|;

[0093] Fitting the relationship between sensor parameter changes and distance: Using the distance matrix D in historical data ij and the corresponding sensor parameter difference ΔZ ij (C a ), and use the polynomial regression model to fit the relationship function between parameter difference and distance: ΔZ ij (C a )=γ a D ij u +α a D ij +β a ; According to the least squares method, the parameter γ is fitted a , αa and β a ;

[0094] Among them, γ a , α a It is sensor C a The coefficient of the parameter variation with distance, β a is a constant term;

[0095] According to the result of the fitting function, the sensor C between different monitoring piles is determined. a The parameter variation range; calculate the difference between the fitted value and the actual observed value in the historical data to determine the upper and lower limits of the variation range: ΔZ ij (C a )∈[γ a D ij u +α a D ij +β a -Kσ a ,γ a D ij u +α a D ij +β a +Kσ a ];

[0096] Among them, σ a represents the standard deviation of the difference between the fitted value and the actual observed value; K represents the adjustment parameter.

[0097] In step S4, new data collection: collect new data from each sensor on each monitoring pile, and the data format is consistent with the historical data;

[0098] Decorrelation processing: decorrelation processing is performed on the new data using the principal component matrix obtained in step S2;

[0099] Calculate the difference: For any two sensors of the same type on the monitoring piles, calculate the difference between their newly collected data:

[0100] Determine the variation range: using the parameter variation range determined in step S3, by comparing the difference with the previously determined variation range;

[0101] Analysis and judgment results: Normal change: If there is no range, it is considered that the sensor parameter changes are normal; Abnormal change: If it exceeds the range, it means that there is an abnormality and feedback is given.

[0102] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A micro-environment monitoring system based on big data, characterized by: The system includes a database module, a relationship analysis module, a change range module, and a new data judgment module; The database module is used to collect the historical data of each monitoring pile sensor and the distance between the monitoring piles and form a database; the relationship analysis module is used to analyze the influence relationship between different sensors on the same monitoring pile and perform decorrelation processing; the variation range module is used to calculate the sensor parameter difference, establish a regression model to fit the relationship between parameter change and distance, and determine the parameter variation range; the new data judgment module is used to judge the abnormality of the new data based on the relationship between parameter change and distance; The output end of the database module is connected to the input ends of the relationship analysis module and the variation range module; The output end of the relationship analysis module is connected to the input end of the new data judgment module; the output end of the change range module is connected to the input end of the new data judgment module.

2. The micro-environment monitoring system based on big data according to claim 1, characterized in that: The database module includes a data collection unit and a database unit; The data collection unit is used to collect historical data of different sensors and distance information between monitoring piles; the database unit is used to form a database based on the collected information and update it in real time; The output end of the data collection unit is connected to the input end of the database unit; the output end of the database unit is connected to the input ends of the relationship analysis module and the variation range module.

3. The micro-environment monitoring system based on big data according to claim 2 is characterized in that: The relationship analysis module includes a correlation analysis unit and a decorrelation processing unit; The correlation analysis unit is used to calculate the Pearson correlation coefficient between sensors; the decorrelation processing unit is used to perform covariance analysis and feature decomposition on sensor data and generate principal components; The output end of the correlation analysis unit is connected to the input end of the decorrelation processing unit; the output end of the decorrelation processing unit is connected to the input end of the new data judgment module.

4. The micro-environment monitoring system based on big data according to claim 3 is characterized in that: The variation range module includes a difference calculation unit, a regression analysis unit and a range determination unit; The difference calculation unit is used to calculate the difference between the sensor parameters corresponding to different monitoring piles; the regression analysis unit is used to fit the parameter difference and distance relationship using a polynomial regression model; the range determination unit is used to determine the sensor parameter variation range according to the difference and distance relationship; The output end of the difference calculation unit is connected to the input end of the regression analysis unit; the output end of the regression analysis unit is connected to the input end of the range determination unit; the output end of the range determination unit is connected to the input end of the new data judgment module.

5. The micro-environment monitoring system based on big data according to claim 4 is characterized in that: The new data judgment module includes a new data acquisition unit and a change judgment unit; The new data acquisition unit is used to acquire new parameters of the sensor; the change judgment unit is used to judge whether the parameter is abnormal according to the difference change of the sensor parameter and the change range; The output end of the new data acquisition unit is connected to the input end of the change judgment unit.

6. A micro-environment monitoring method based on big data, characterized in that: The method comprises the following steps: S1. Build a historical database, including the sensor parameters in the historical data and the distances between different monitoring piles; S2, analyzing the influence relationship between different sensors of the same monitoring pile and performing decorrelation processing; S3, determining the parameter change range according to the distance between the monitoring piles and the corresponding sensor parameter changes in the historical data; S4. Decorrelate the newly collected data of the monitoring pile sensors to determine whether the difference between the sensors of different monitoring piles is within the variation range; and determine whether the corresponding sensor parameters of different monitoring piles are abnormal based on the determination results.

7. The micro-environment monitoring method based on big data according to claim 6 is characterized in that: In step S1, the monitoring piles are provided with a plurality of sensors, and the types and quantities of sensors provided on different monitoring piles are consistent, and the monitoring piles are arranged at different positions to monitor the environment; The monitoring pile is represented by [Z1, Z2, …, Z m ]; the sensor is represented by [C1,C2,…,C n ]; Monitoring pile Z i and Z j The distance between them is represented by D ij ; Among them, m is a positive integer, indicating the number of monitoring piles, Z m represents the mth monitoring pile; n is a positive integer, indicating the number of sensors on each monitoring pile, C n represents the nth sensor; i,j∈{1,2,…,m}, represents the sequence of monitoring piles; The format of the historical database is [sample ID, Z1(C1),…Z1(C n ),…,Z m (C1),Z m (C n ),D 12 ,…,D 1m ,D 23 ,…D 2m ,…,D (m-1)m ]; Among them, Z m (C n ) represents the parameter monitored by the nth sensor of the mth monitoring pile.

8. The micro-environment monitoring method based on big data according to claim 7 is characterized in that: In step S2, the data set of all sensors on each monitoring pile is extracted from the historical data set [sample ID, Z1(C1),…Z1(C n ),…,Z m (C1),Z m (C n )]; For a sensor C on a specified monitoring pile a All the corresponding historical data are represented by H a , the influence relationship is measured by calculating the Pearson correlation coefficient between it and other sensor data: r Ha,Hb =(Σ(H a -H a mean )(H b -H b mean )) / ((Σ(H a -H a mean) 2 ) 1 / 2 ×(Σ(H b -H b mean ) 2 ) 1 / 2 ); The Pearson correlation coefficient range is [-1,1], close to 1 or -1 indicates a high correlation, close to 0 indicates no correlation; Where a is a positive integer, a∈[1,n]; H a and H b are the measurement values ​​of different sensors on the same monitoring pile, H a mean and H b mean are their means respectively; Each monitoring pile builds a correlation matrix between sensors, and the matrix elements represent the correlation between different sensors. When the element value is close to -1 or 1, it means that the data of the corresponding two sensors are highly correlated, and there is a high possibility of influence. Decorrelate the sensor data: Construct the covariance matrix: Σ ab =(1 / (N-1))Σ q＝1 N (H a (q) -H a mean )(H b (q) -H b mean ); Where N is a positive integer, indicating the number of samples; q is a positive integer, q≤N, indicating the sample sequence; H a (q) Indicates that the qth sample corresponds to a specified monitoring pile sensor C a data; Perform eigendecomposition on the covariance matrix to obtain eigenvalues ​​and eigenvectors; the eigenvalue represents the variance explained by each principal component, and the eigenvector is used to define the new principal component axis: ΣV = λV; according to the eigenvalue, select the first k principal components with the largest explained variance; by projecting the original sensor data into the principal component space, obtain the decorrelated sensor data: H new =V T H; Where V represents the eigenvector; λ represents the eigenvalue; k is a positive integer; H new Represents the data after decorrelation processing.

9. The micro-environment monitoring method based on big data according to claim 8, characterized in that: In step S3, for any two monitoring piles Z i and Z j The distance between them is represented by D ij ; Monitoring pile Z i and Z j Same type sensor C a The parameter difference between them is expressed as: ΔZ ij (C a )=|Z i (C a )-Z j (C a )|; Fitting the relationship between sensor parameter changes and distance: Using the distance matrix D in historical data ij and the corresponding sensor parameter difference ΔZ ij (C a ), and use the polynomial regression model to fit the relationship function between parameter difference and distance: ΔZ ij (C a )=γ a D ij u +α a D ij +β a ; According to the least squares method, the parameter γ is fitted a , α a and β a ; Among them, γ a , α a It is sensor C a The coefficient of the parameter variation with distance, β a is a constant term; According to the result of the fitting function, the sensor C between different monitoring piles is determined. a The parameter variation range; calculate the difference between the fitted value and the actual observed value in the historical data to determine the upper and lower limits of the variation range: ΔZ ij (C a )∈[γ a D ij u +α a D ij +β a -Kσ a ,γ a D ij u +α a D ij +β a +Kσ a ]; Among them, σ a represents the standard deviation of the difference between the fitted value and the actual observed value; K represents the adjustment parameter.

10. The micro-environment monitoring method based on big data according to claim 9, characterized in that: In step S4, new data collection: collect new data from each sensor on each monitoring pile, and the data format is consistent with the historical data; Decorrelation processing: decorrelation processing is performed on the new data using the principal component matrix obtained in step S2; Calculate the difference: For any two sensors of the same type on the monitoring piles, calculate the difference between their newly collected data: Determine the variation range: using the parameter variation range determined in step S3, by comparing the difference with the previously determined variation range; Analysis and judgment results: Normal change: If there is no range, it is considered that the sensor parameter changes are normal; Abnormal change: If it exceeds the range, it means that there is an abnormality and feedback is given.

Images