Land use dynamic change monitoring method and system based on multimodal remote sensing data

Abstract

The present invention discloses a method and system for monitoring dynamic changes in land use based on multimodal remote sensing data, and relates to the technical field of multimodal remote sensing data analysis. The system of the present invention includes: a data acquisition and preprocessing module, a consistency baseline construction module, a real-time data analysis module, and an anomaly diagnosis and processing module; the data acquisition and preprocessing module realizes the spatiotemporal alignment of multi-source data; the consistency baseline construction module divides land use types based on geographical features and establishes a baseline database; the real-time data analysis module screens suspected anomaly areas through a difference index; the anomaly diagnosis and processing module distinguishes between data quality anomalies and emergency anomalies based on feature correlation relationships, the former triggering data repair and verification, and the latter marking high-priority areas and outputting location information. The system realizes high-precision monitoring of land use changes and intelligent anomaly classification through multimodal data fusion and correlation analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of multimodal remote sensing data analysis, and in particular to a method and system for monitoring dynamic changes in land use based on multimodal remote sensing data. Background Art

[0002] Monitoring dynamic land use change is a key research area within remote sensing applications, widely used in a variety of fields, including urban planning, resource management, and environmental protection. With the advancement of remote sensing technology, an increasing number of sensors and satellite platforms are being used to acquire multimodal remote sensing data, including optical, radar (such as synthetic aperture radar (SAR), and thermal infrared). These multimodal remote sensing data provide diverse perspectives and information sources, enabling effective monitoring of land use change. This data offers significant advantages, particularly in large-scale, timely, and continuous monitoring.

[0003] However, in practical applications, remote sensing data is often affected by various factors, such as cloudiness, sensor failures, and data transmission errors. This can lead to fluctuations in data quality and affect the accuracy of monitoring dynamic land use changes. Furthermore, land use change can also be subject to unusual events, such as forest fires and sudden changes in urbanization. These changes may differ from normal seasonal variations or fluctuations in ground properties and can have profound impacts on the environment. Therefore, accurately identifying and classifying data anomalies, and exploiting the physical relationships and consistency between multimodal remote sensing data to detect and correct these anomalies, have become key issues in remote sensing data processing. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for monitoring dynamic changes in land use based on multimodal remote sensing data to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] The method for monitoring dynamic changes in land use based on multimodal remote sensing data includes the following steps:

[0007] Step S100. Acquire multimodal remote sensing historical data within the monitoring period and geographic information data of the monitoring area, divide the monitoring area into different land use types based on the geographic information data; and based on the division results of the monitoring area, map the corresponding multimodal remote sensing historical data and establish a consistent baseline database for different land use types in a stable state;

[0008] Step S200: Acquire multimodal remote sensing real-time data for different land use types within the monitoring area, compare the multimodal remote sensing real-time data with the corresponding consistency baseline database, and obtain a corresponding difference index; based on the difference index, preliminarily screen out areas where the degree of deviation from the consistency baseline database is greater than a threshold, and mark them as suspected abnormal areas;

[0009] Step S300. Analyze the correlation relationship of the corresponding multimodal remote sensing real-time data for the suspected abnormal area, and identify the abnormal diagnosis results of the suspected abnormal area based on the correlation relationship; wherein the abnormal diagnosis results are divided into data quality abnormalities and emergency abnormalities;

[0010] Step S400. Mark the area identified as having abnormal data quality as an unreliable data area, repair the corresponding multimodal remote sensing real-time data, and compare the repaired multimodal remote sensing real-time data with the corresponding consistency baseline database, and output corresponding prompt information based on the comparison results; mark the area identified as abnormal emergency event as a high-priority change area, obtain the location information of the high-priority change area and output it to relevant personnel, who will then perform corresponding processing.

[0011] Furthermore, step S100 includes:

[0012] S101. Acquire multimodal historical remote sensing data within the monitoring period, including optical remote sensing, synthetic aperture radar, and infrared data; as well as geographic information data of the monitoring area, including land use status, vegetation cover, soil type, terrain elevation, etc.; preprocess the multimodal historical remote sensing data, including radiometric calibration, geometric correction, and atmospheric correction; and format and spatially register the geographic information data to ensure consistency in spatial and temporal dimensions between the geographic information data and the multimodal historical remote sensing data.

[0013] S102. Divide the monitoring area into several sub-areas according to a preset minimum area unit; for each sub-area, obtain geographic information data corresponding to the sub-area and extract corresponding geographic data features from the geographic information data, wherein the geographic data features include terrain features, soil features, vegetation features, and hydrological features; normalize the extracted geographic data features to form a corresponding geographic data feature vector V, where V = [v1, v2, ..., vn], where v1 represents the geographic data feature of the first dimension, v2 represents the geographic data feature of the second dimension, and so on, vn represents the geographic data feature of the nth dimension; aggregate the geographic data feature vectors V of all sub-areas and calculate the average value to obtain the corresponding geographic data feature average vector V0; and identify the land use type of each sub-area based on the geographic data feature vector V;

[0014] S103. Based on the classified land use types, the corresponding multimodal historical remote sensing data is associated with each sub-region. For all sub-regions of the same land use type, the corresponding multimodal historical remote sensing data and the stable state marking period are obtained, where the stable state marking period is analyzed by relevant personnel. Based on the stable state marking period, the multimodal historical remote sensing data of all sub-regions of the same land use type are intercepted and time series analysis is performed on the intercepted multimodal remote sensing historical data to obtain the multimodal remote sensing historical data characteristics of each sub-region during the stable state marking period. The multimodal remote sensing historical data characteristics are reduced in dimension using principal component analysis to obtain the main eigenvector Z. The characteristic mean and standard deviation of the main eigenvector Z for each land use type are calculated to construct a consistent baseline database. The database includes the key characteristics of each land use type in the stable state, such as vegetation index, soil moisture, surface temperature, etc., as well as its stable pattern in different time periods. The stable state data of each sub-region will be integrated to form a high-dimensional feature space to ensure that the database can comprehensively reflect the stable state characteristics of each land use type.

[0015] Furthermore, in S102, the land use type of each sub-region is identified based on the geographic data feature vector V. The specific analysis process is as follows:

[0016] For each sub-region, the corresponding geographic data feature vector V and the geographic data feature average vector V0 are represented in the same radar chart, and the number of axes corresponding to the radar chart is n; on each axis of the radar chart, the elements whose geographic data feature vector V is greater than the geographic data feature average vector V0 are marked as key features, and the key features are summarized to form the key feature set G of the corresponding sub-region; the key feature set Gi corresponding to each land use type in the database is obtained, where i represents the number of land use types; the key feature set G of each sub-region is compared with the key feature set Gi corresponding to each land use type in turn, and the corresponding matching coefficient Pi is calculated, and Pi=N(G∩Gi) / N(Gi), where N(G∩Gi) represents the number of key features in the intersection of the key feature set G and the key feature set Gi, and N(Gi) represents the number of key features corresponding to the i-th land use type; the matching coefficients Pi corresponding to all land use types are summarized, and the land use type with the largest matching coefficient is selected as the land use type of the corresponding sub-region.

[0017] Furthermore, step S200 includes:

[0018] S201. Obtain multimodal remote sensing real-time data of sub-regions with different land use types within the monitoring area, analyze the multimodal remote sensing real-time data according to the analysis method of multimodal historical data, and obtain a real-time data feature vector S with the same dimension as the consistency baseline database; according to the land use type of the corresponding sub-region, obtain the corresponding consistency baseline database, calculate the feature mean and standard deviation of the real-time data feature vector S of the current sub-region and the main feature vector Z corresponding to the consistency baseline database, and obtain the corresponding difference index R, and ,

[0019] Among them S k represents the kth eigenvalue of the real-time data eigenvector, μ_Z k represents the mean of the kth eigenvalue of the main eigenvector Z in the consistency baseline database, σ_Z k represents the standard deviation of the kth eigenvalue of the main eigenvector Z in the consistency baseline database;

[0020] S202. Compare the difference index R of sub-regions of different land use types in the monitoring area with a preset threshold R0. If there is a sub-region with a difference index R greater than the preset threshold R0, mark the corresponding sub-region as a suspected abnormal region.

[0021] Furthermore, step S300 includes:

[0022] S301. For the suspected abnormal area, obtain the real-time data feature vector S of the corresponding multimodal remote sensing real-time data, analyze the correlation between different elements in the real-time data feature vector S of the multimodal remote sensing real-time data, and thus form the correlation matrix A m×m , and the association matrix A m×m Each element aij in represents the correlation between the i-th feature and the j-th feature, and the Pearson correlation coefficient is used to calculate the linear correlation between the feature elements; obtain the consistency baseline database of land use types corresponding to the suspected abnormal area, and analyze the correlation between different elements in the main feature vector Z in the consistency baseline database according to the analysis method of the real-time data feature vector S, so as to obtain the baseline correlation matrix B m×m ;

[0023] S302. Calculate the correlation matrix A of the suspected abnormal area m×m And the corresponding baseline correlation matrix B m×mThe difference between them is used to extract elements whose values ​​are not 0 and are greater than or equal to the preset threshold, and the corresponding real-time data feature quantity N is obtained according to the elements; if 0<N<N1 is satisfied, the abnormal diagnosis result of data quality abnormality is output; if N≥N1, the abnormal diagnosis result of emergency abnormality is output; where N1 indicates that the correlation does not meet the characteristic element quantity threshold of the consistency baseline database, and the value is less than m, which is obtained by relevant personnel based on specific situation analysis.

[0024] Among them, if the correlation between only a few features and other features does not meet the requirements of the consistency baseline database, and the differences between these features are within the set tolerance range, it is classified as a data quality anomaly; if the correlation between multiple features does not meet the requirements of the consistency baseline database, and the differences between these features are large, it is classified as an emergency event anomaly.

[0025] Furthermore, step S400 includes:

[0026] S401. Mark the area identified as having abnormal data quality as an unreliable data area. Combine the multimodal remote sensing historical data of the same area with the multimodal remote sensing real-time data for the unreliable data area to perform data repair. Calculate the real-time difference index Rs of the repaired multimodal remote sensing real-time data according to the difference index R calculation formula. If the real-time difference index Rs is less than or equal to the threshold R0, output a prompt message "Data repair successful". If the real-time difference index Rs is greater than the threshold R0, output a prompt message "Abnormalities still exist after data repair, please check further" and relevant personnel will take appropriate measures.

[0027] S402. Mark the area identified as an emergency abnormality as a high-priority change area, obtain the location information of the high-priority change area and output it to relevant personnel, who will then handle it accordingly.

[0028] The land use dynamic change monitoring system based on multimodal remote sensing data includes: data acquisition and preprocessing module, consistency baseline construction module, real-time data analysis module, and abnormal diagnosis and processing module;

[0029] The data acquisition and preprocessing module acquires multimodal remote sensing historical data within the monitoring period and geographic information data of the monitoring area; preprocesses the multimodal historical remote sensing data, unifies the format of the geographic information data and performs spatial registration, so that the geographic information data and the multimodal historical remote sensing data are consistent in spatial and temporal dimensions;

[0030] The consistency baseline construction module divides the monitoring area into multiple sub-areas and identifies the land use type of each sub-area based on its geographical data characteristics; analyzes the steady-state data of each land use type and establishes a consistency baseline database;

[0031] The real-time data analysis module obtains multimodal remote sensing real-time data of different land use types in the monitoring area and compares it with the consistency baseline database; it calculates the difference index between the real-time data characteristics and the baseline database characteristics, thereby preliminarily screening and marking suspected abnormal areas;

[0032] The anomaly diagnosis and processing module analyzes suspected abnormal areas to identify data quality anomalies or sudden event anomalies; it identifies the type of anomaly by analyzing the characteristic correlation relationship of multimodal remote sensing real-time data; if it is diagnosed as a data quality anomaly, the unreliable data is repaired and the repaired data is re-compared; if it is diagnosed as a sudden event anomaly, it is marked as a high-priority change area, and the relevant location information is promptly provided to relevant personnel for processing.

[0033] Furthermore, the data acquisition and preprocessing module includes a data acquisition unit and a data preprocessing unit;

[0034] The data acquisition unit acquires the multimodal remote sensing historical data within the monitoring period and the geographic information data of the monitoring area; the data preprocessing unit preprocesses the multimodal historical remote sensing data, unifies the format and spatially aligns the geographic information data, so that the geographic information data and the multimodal historical remote sensing data are consistent in spatial and temporal dimensions.

[0035] Furthermore, the consistency baseline construction module includes a land use type analysis unit and a consistency baseline construction unit;

[0036] The land use type analysis unit divides the monitoring area into multiple sub-areas and identifies the land use type of each sub-area based on its geographical data characteristics; the consistency baseline construction unit analyzes the steady-state data of each land use type and establishes a consistency baseline database.

[0037] Furthermore, the real-time data analysis module includes a consistency baseline database comparison unit and a suspected abnormal area identification unit;

[0038] The consistency baseline database comparison unit obtains multimodal remote sensing real-time data of different land use types in the monitoring area and compares it with the consistency baseline database; the suspected abnormal area identification unit calculates the difference index between the real-time data characteristics and the baseline database characteristics, thereby preliminarily screening and marking suspected abnormal areas;

[0039] The abnormality diagnosis and processing module includes an abnormality diagnosis unit and a prompt information output unit;

[0040] The abnormality diagnosis unit analyzes suspected abnormal areas to identify data quality abnormalities or emergency abnormalities; it identifies the type of abnormality by analyzing the characteristic correlation relationship of multimodal remote sensing real-time data; if the prompt information output unit diagnoses it as data quality abnormality, it will repair the unreliable data and re-compare the repaired data; if it diagnoses it as an emergency abnormality, it will be marked as a high-priority change area and the relevant location information will be provided to relevant personnel in a timely manner for processing.

[0041] Compared with the existing technology, the beneficial effects of the present invention are as follows: the present invention accurately divides the monitoring area into different land use types through the combination of geographic information data and remote sensing data, and establishes a consistent baseline database based on the stable state marking period; this division method makes monitoring more accurate, and can conduct detailed analysis on different types of land use, ensuring that the monitoring results are more targeted. By establishing a consistent baseline database and comparing the differences between the characteristics of multimodal remote sensing real-time data and historical data, the present invention can efficiently screen out abnormal areas and accurately diagnose abnormal types; this analysis method based on the difference index and the correlation matrix can timely discover abnormal events such as forest fires and sudden changes in the urbanization process during the monitoring process, and has high real-time and accuracy. For areas with abnormal data quality, the present invention provides a data repair mechanism, which repairs unreliable data by comparing multimodal historical data and evaluating it through the difference index; the repaired data can be verified twice to ensure the reliability and accuracy of the data; compared with the existing technology, the present invention can effectively avoid manual intervention when processing data anomalies, thereby improving the degree of automation and data processing efficiency. For abnormal emergencies, the present invention can prioritize the identification of relevant areas and promptly feed back location information to relevant personnel for processing, ensuring a rapid response to emergencies. This feature is of great significance in practical applications in areas such as environmental protection and disaster monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0043] Figure 1 It is a module schematic diagram of the land use dynamic change monitoring system based on multimodal remote sensing data of the present invention. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0045] See also Figure 1 , the present invention provides a technical solution:

[0046] The land use dynamic change monitoring system based on multimodal remote sensing data includes: data acquisition and preprocessing module, consistency baseline construction module, real-time data analysis module, and abnormal diagnosis and processing module;

[0047] The data acquisition and preprocessing module acquires multimodal remote sensing historical data within the monitoring period and geographic information data of the monitoring area; preprocesses the multimodal historical remote sensing data, unifies the format of the geographic information data and performs spatial registration, so that the geographic information data and the multimodal historical remote sensing data are consistent in spatial and temporal dimensions;

[0048] The consistency baseline construction module divides the monitoring area into multiple sub-areas and identifies the land use type of each sub-area based on its geographical data characteristics; analyzes the steady-state data of each land use type and establishes a consistency baseline database;

[0049] The real-time data analysis module obtains multimodal remote sensing real-time data of different land use types in the monitoring area and compares it with the consistency baseline database; it calculates the difference index between the real-time data characteristics and the baseline database characteristics, thereby preliminarily screening and marking suspected abnormal areas;

[0050] The anomaly diagnosis and processing module analyzes suspected abnormal areas to identify data quality anomalies or sudden event anomalies; it identifies the type of anomaly by analyzing the characteristic correlation relationship of multimodal remote sensing real-time data; if it is diagnosed as a data quality anomaly, the unreliable data is repaired and the repaired data is re-compared; if it is diagnosed as a sudden event anomaly, it is marked as a high-priority change area, and the relevant location information is promptly provided to relevant personnel for processing.

[0051] The data acquisition and preprocessing module includes a data acquisition unit and a data preprocessing unit;

[0052] The data acquisition unit acquires the multimodal remote sensing historical data within the monitoring period and the geographic information data of the monitoring area; the data preprocessing unit preprocesses the multimodal historical remote sensing data, unifies the format and spatially aligns the geographic information data, so that the geographic information data and the multimodal historical remote sensing data are consistent in spatial and temporal dimensions.

[0053] The consistent baseline construction module includes a land use type analysis unit and a consistent baseline construction unit;

[0054] The land use type analysis unit divides the monitoring area into multiple sub-areas and identifies the land use type of each sub-area based on its geographical data characteristics; the consistency baseline construction unit analyzes the steady-state data of each land use type and establishes a consistency baseline database.

[0055] The real-time data analysis module includes a consistency baseline database comparison unit and a suspected abnormal area identification unit;

[0056] The consistency baseline database comparison unit obtains multimodal remote sensing real-time data of different land use types in the monitoring area and compares it with the consistency baseline database; the suspected abnormal area identification unit calculates the difference index between the real-time data characteristics and the baseline database characteristics, thereby preliminarily screening and marking suspected abnormal areas;

[0057] The abnormality diagnosis and processing module includes an abnormality diagnosis unit and a prompt information output unit;

[0058] The abnormality diagnosis unit analyzes suspected abnormal areas to identify data quality abnormalities or emergency abnormalities; it identifies the type of abnormality by analyzing the characteristic correlation relationship of multimodal remote sensing real-time data; if the prompt information output unit diagnoses it as data quality abnormality, it will repair the unreliable data and re-compare the repaired data; if it diagnoses it as an emergency abnormality, it will be marked as a high-priority change area and the relevant location information will be provided to relevant personnel in a timely manner for processing.

[0059] The method for monitoring dynamic changes in land use based on multimodal remote sensing data includes the following steps:

[0060] Step S100. Acquire multimodal remote sensing historical data within the monitoring period and geographic information data of the monitoring area, divide the monitoring area into different land use types based on the geographic information data; and based on the division results of the monitoring area, map the corresponding multimodal remote sensing historical data and establish a consistent baseline database for different land use types in a stable state;

[0061] Step S200: Acquire multimodal remote sensing real-time data for different land use types within the monitoring area, compare the multimodal remote sensing real-time data with the corresponding consistency baseline database, and obtain a corresponding difference index; based on the difference index, preliminarily screen out areas where the degree of deviation from the consistency baseline database is greater than a threshold, and mark them as suspected abnormal areas;

[0062] Step S300. Analyze the correlation relationship of the corresponding multimodal remote sensing real-time data for the suspected abnormal area, and identify the abnormal diagnosis results of the suspected abnormal area based on the correlation relationship; wherein the abnormal diagnosis results are divided into data quality abnormalities and emergency abnormalities;

[0063] Step S400. Mark the area identified as having abnormal data quality as an unreliable data area, repair the corresponding multimodal remote sensing real-time data, and compare the repaired multimodal remote sensing real-time data with the corresponding consistency baseline database, and output corresponding prompt information based on the comparison results; mark the area identified as abnormal emergency event as a high-priority change area, obtain the location information of the high-priority change area and output it to relevant personnel, who will then perform corresponding processing.

[0064] Step S100 includes:

[0065] S101. Acquire multimodal historical remote sensing data within the monitoring period, including optical remote sensing, synthetic aperture radar, and infrared data; as well as geographic information data of the monitoring area, including land use status, vegetation cover, soil type, terrain elevation, etc.; preprocess the multimodal historical remote sensing data, including radiometric calibration, geometric correction, and atmospheric correction; and format and spatially register the geographic information data to ensure consistency in spatial and temporal dimensions between the geographic information data and the multimodal historical remote sensing data.

[0066] S102. Divide the monitoring area into several sub-areas according to a preset minimum area unit; for each sub-area, obtain geographic information data corresponding to the sub-area and extract corresponding geographic data features from the geographic information data, wherein the geographic data features include terrain features, soil features, vegetation features, and hydrological features; normalize the extracted geographic data features to form a corresponding geographic data feature vector V, where V = [v1, v2, ..., vn], where v1 represents the geographic data feature of the first dimension, v2 represents the geographic data feature of the second dimension, and so on, vn represents the geographic data feature of the nth dimension; aggregate the geographic data feature vectors V of all sub-areas and calculate the average value to obtain the corresponding geographic data feature average vector V0; and identify the land use type of each sub-area based on the geographic data feature vector V;

[0067] S103. Based on the classified land use types, the corresponding multimodal historical remote sensing data is associated with each sub-region. For all sub-regions of the same land use type, the corresponding multimodal historical remote sensing data and the stable state marking period are obtained, where the stable state marking period is analyzed by relevant personnel. Based on the stable state marking period, the multimodal historical remote sensing data of all sub-regions of the same land use type are intercepted and time series analysis is performed on the intercepted multimodal remote sensing historical data to obtain the multimodal remote sensing historical data characteristics of each sub-region during the stable state marking period. The multimodal remote sensing historical data characteristics are reduced in dimension using principal component analysis to obtain the main eigenvector Z. The characteristic mean and standard deviation of the main eigenvector Z for each land use type are calculated to construct a consistent baseline database. The database includes the key characteristics of each land use type in the stable state, such as vegetation index, soil moisture, surface temperature, etc., as well as its stable pattern in different time periods. The stable state data of each sub-region will be integrated to form a high-dimensional feature space to ensure that the database can comprehensively reflect the stable state characteristics of each land use type.

[0068] In S102, the land use type of each sub-region is identified based on the geographic data feature vector V. The specific analysis process is as follows:

[0069] For each sub-region, the corresponding geographic data feature vector V and the geographic data feature average vector V0 are represented in the same radar chart, and the number of axes corresponding to the radar chart is n; on each axis of the radar chart, the elements whose geographic data feature vector V is greater than the geographic data feature average vector V0 are marked as key features, and the key features are summarized to form the key feature set G of the corresponding sub-region; the key feature set Gi corresponding to each land use type in the database is obtained, where i represents the number of land use types; the key feature set G of each sub-region is compared with the key feature set Gi corresponding to each land use type in turn, and the corresponding matching coefficient Pi is calculated, and Pi=N(G∩Gi) / N(Gi), where N(G∩Gi) represents the number of key features in the intersection of the key feature set G and the key feature set Gi, and N(Gi) represents the number of key features corresponding to the i-th land use type; the matching coefficients Pi corresponding to all land use types are summarized, and the land use type with the largest matching coefficient is selected as the land use type of the corresponding sub-region.

[0070] In this embodiment, it is assumed that the key feature set G of a certain sub-region is: G = {slope, vegetation coverage}, and the key feature set corresponding to each land use type is obtained from the database. It is assumed that there are: G1, G2 and G3, and they are expressed as: G1 = {slope, soil moisture, vegetation coverage}; G2 = {slope, soil moisture, moisture content}; G3 = {vegetation coverage, moisture content};

[0071] Calculate the corresponding matching coefficient Pi, which is:

[0072] P1=N(G∩G1) / N(G1)=2 / 3; P2=N(G∩G2) / N(G2)=1 / 3; P3=N(G∩G3) / N(G3)=1 / 2;

[0073] Since P1>P3>P2, the land use type corresponding to this sub-region is the land use type corresponding to G1.

[0074] Step S200 includes:

[0075] S201. Obtain multimodal remote sensing real-time data of sub-regions with different land use types within the monitoring area, analyze the multimodal remote sensing real-time data according to the analysis method of multimodal historical data, and obtain a real-time data feature vector S with the same dimension as the consistency baseline database; according to the land use type of the corresponding sub-region, obtain the corresponding consistency baseline database, calculate the feature mean and standard deviation of the real-time data feature vector S of the current sub-region and the main feature vector Z corresponding to the consistency baseline database, and obtain the corresponding difference index R, and ,

[0076] Among them S k represents the kth eigenvalue of the real-time data eigenvector, μ_Z k represents the mean of the kth eigenvalue of the main eigenvector Z in the consistency baseline database, σ_Z k represents the standard deviation of the kth eigenvalue of the main eigenvector Z in the consistency baseline database;

[0077] S202. Compare the difference index R of sub-regions of different land use types in the monitoring area with a preset threshold R0. If there is a sub-region with a difference index R greater than the preset threshold R0, mark the corresponding sub-region as a suspected abnormal region.

[0078] Step S300 includes:

[0079] S301. For the suspected abnormal area, obtain the real-time data feature vector S of the corresponding multimodal remote sensing real-time data, analyze the correlation between different elements in the real-time data feature vector S of the multimodal remote sensing real-time data, and thus form the correlation matrix A m×m , and the association matrix A m×mEach element aij in represents the correlation between the i-th feature and the j-th feature, and the Pearson correlation coefficient is used to calculate the linear correlation between the feature elements; obtain the consistency baseline database of land use types corresponding to the suspected abnormal area, and analyze the correlation between different elements in the main feature vector Z in the consistency baseline database according to the analysis method of the real-time data feature vector S, so as to obtain the baseline correlation matrix B m×m ;

[0080] S302. Calculate the correlation matrix A of the suspected abnormal area m×m And the corresponding baseline correlation matrix B m×m The difference between them is used to extract elements whose values ​​are not 0 and are greater than or equal to the preset threshold, and the corresponding real-time data feature quantity N is obtained according to the elements; if 0<N<N1 is satisfied, the abnormal diagnosis result of data quality abnormality is output; if N≥N1, the abnormal diagnosis result of emergency abnormality is output; where N1 indicates that the correlation does not meet the characteristic element quantity threshold of the consistency baseline database, and the value is less than m, which is obtained by relevant personnel based on specific situation analysis.

[0081] Among them, if the correlation between only a few features and other features does not meet the requirements of the consistency baseline database, and the differences between these features are within the set tolerance range, it is classified as a data quality anomaly; if the correlation between multiple features does not meet the requirements of the consistency baseline database, and the differences between these features are large, it is classified as an emergency event anomaly.

[0082] Step S400 includes:

[0083] S401. Mark the area identified as having abnormal data quality as an unreliable data area. Combine the multimodal remote sensing historical data of the same area with the multimodal remote sensing real-time data for the unreliable data area to perform data repair. Calculate the real-time difference index Rs of the repaired multimodal remote sensing real-time data according to the difference index R calculation formula. If the real-time difference index Rs is less than or equal to the threshold R0, output a prompt message "Data repair successful". If the real-time difference index Rs is greater than the threshold R0, output a prompt message "Abnormalities still exist after data repair, please check further" and relevant personnel will take appropriate measures.

[0084] S402. Mark the area identified as an emergency abnormality as a high-priority change area, obtain the location information of the high-priority change area and output it to relevant personnel, who will then handle it accordingly.

[0085] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0086] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for monitoring dynamic changes in land use based on multimodal remote sensing data, characterized by: The method comprises the following steps: Step S100. Acquire multimodal remote sensing historical data within the monitoring period and geographic information data of the monitoring area, divide the monitoring area into different land use types based on the geographic information data; and based on the division results of the monitoring area, map the corresponding multimodal remote sensing historical data and establish a consistent baseline database for different land use types in a stable state; The step S100 includes: S101. Acquire multimodal historical remote sensing data within the monitoring period and geographic information data of the monitoring area, preprocess the multimodal historical remote sensing data, including radiometric calibration, geometric correction, and atmospheric correction; and format and spatially register the geographic information data to ensure consistency between the geographic information data and the multimodal historical remote sensing data in spatial and temporal dimensions. S102. Divide the monitoring area into several sub-areas according to a preset minimum area unit; for each sub-area, obtain geographic information data corresponding to the sub-area and extract corresponding geographic data features from the geographic information data; normalize the extracted geographic data features to form a corresponding geographic data feature vector V, where V = [v1, v2, ..., vn], where v1 represents the geographic data feature of the first dimension, v2 represents the geographic data feature of the second dimension, and so on, and vn represents the geographic data feature of the nth dimension; aggregate the geographic data feature vectors V of all sub-areas and calculate the average value to obtain the corresponding geographic data feature average vector V0; and identify the land use type of each sub-area based on the geographic data feature vector V; S103. Based on the classified land use types, the corresponding multimodal historical remote sensing data is associated with each sub-region. For all sub-regions of the same land use type, the corresponding multimodal historical remote sensing data and the stable state marking period are obtained, where the stable state marking period is analyzed by relevant personnel. Based on the stable state marking period, the multimodal historical remote sensing data of all sub-regions of the same land use type are intercepted, and time series analysis is performed on the intercepted multimodal remote sensing historical data to obtain the multimodal historical remote sensing data characteristics of each sub-region during the stable state marking period. The principal component analysis method is used to reduce the dimensionality of the multimodal historical remote sensing data characteristics to obtain the main eigenvector Z. The characteristic mean and standard deviation of the main eigenvector Z for each land use type are calculated to construct a consistency baseline database. Step S200: Acquire multimodal remote sensing real-time data for different land use types within the monitoring area, compare the multimodal remote sensing real-time data with the corresponding consistency baseline database, and obtain a corresponding difference index; based on the difference index, preliminarily screen out areas where the degree of deviation from the consistency baseline database is greater than a threshold, and mark them as suspected abnormal areas; The step S200 includes: S201. Obtain multimodal remote sensing real-time data of sub-regions with different land use types within the monitoring area, analyze the multimodal remote sensing real-time data according to the analysis method of multimodal historical data, and obtain a real-time data feature vector S with the same dimension as the consistency baseline database; according to the land use type of the corresponding sub-region, obtain the corresponding consistency baseline database, calculate the feature mean and standard deviation of the real-time data feature vector S of the current sub-region and the main feature vector Z corresponding to the consistency baseline database, and obtain the corresponding difference index R, and ， Among them S k represents the kth eigenvalue of the real-time data eigenvector, μ_Z k represents the mean of the kth eigenvalue of the main eigenvector Z in the consistency baseline database, σ_Z k represents the standard deviation of the kth eigenvalue of the main eigenvector Z in the consistency baseline database; S202. Compare the difference index R of sub-regions of different land use types within the monitoring area with a preset threshold R0. If there is a sub-region with a difference index R greater than the preset threshold R0, mark the corresponding sub-region as a suspected abnormal area; Step S300. Analyze the correlation relationship of the corresponding multimodal remote sensing real-time data for the suspected abnormal area, and identify the abnormal diagnosis results of the suspected abnormal area based on the correlation relationship; wherein the abnormal diagnosis results are divided into data quality abnormalities and emergency abnormalities; Step S400. Mark the area identified as having abnormal data quality as an unreliable data area, repair the corresponding multimodal remote sensing real-time data, and compare the repaired multimodal remote sensing real-time data with the corresponding consistency baseline database, and output corresponding prompt information based on the comparison results; mark the area identified as abnormal emergency event as a high-priority change area, obtain the location information of the high-priority change area and output it to relevant personnel, who will then perform corresponding processing.

2. The method for monitoring dynamic changes in land use based on multimodal remote sensing data according to claim 1, characterized in that: In S102, the land use type of each sub-region is identified based on the geographic data feature vector V. The specific analysis process is as follows: For each sub-region, the corresponding geographic data feature vector V and the geographic data feature average vector V0 are represented in the same radar chart, and the number of axes corresponding to the radar chart is n; on each axis of the radar chart, the elements whose geographic data feature vector V is greater than the geographic data feature average vector V0 are marked as key features, and the key features are summarized to form the key feature set G of the corresponding sub-region; the key feature set Gi corresponding to each land use type in the database is obtained, where i represents the number of land use types; the key feature set G of each sub-region is compared with the key feature set Gi corresponding to each land use type in turn, and the corresponding matching coefficient Pi is calculated, and Pi=N(G∩Gi) / N(Gi), where N(G∩Gi) represents the number of key features in the intersection of the key feature set G and the key feature set Gi, and N(Gi) represents the number of key features corresponding to the i-th land use type; the matching coefficients Pi corresponding to all land use types are summarized, and the land use type with the largest matching coefficient is selected as the land use type of the corresponding sub-region.

3. The method for monitoring dynamic changes in land use based on multimodal remote sensing data according to claim 1, characterized in that: The step S300 includes: S301. For the suspected anomaly area, obtain the real-time data feature vector S of the corresponding multimodal remote sensing real-time data, analyze the correlation between different elements in the real-time data feature vector S of the multimodal remote sensing real-time data, thereby forming a correlation matrix Am×m, where each element aij in the correlation matrix Am×m represents the correlation between the i-th feature and the j-th feature, and the Pearson correlation coefficient is used to calculate the linear correlation between the feature elements; obtain a consistency baseline database of land use types corresponding to the suspected anomaly area, and analyze the correlation between different elements in the main feature vector Z in the consistency baseline database according to the analysis method of the real-time data feature vector S, thereby obtaining a baseline correlation matrix Bm×m; S302. Calculate the difference between the correlation matrix Am×m of the suspected abnormal area and the corresponding baseline correlation matrix Bm×m, extract the elements whose values ​​are not 0 and are greater than or equal to the preset threshold, and obtain the corresponding real-time data feature number N based on the elements; if 0<N<N1 is satisfied, then output the abnormal diagnosis result of data quality abnormality; if N≥N1, then output the abnormal diagnosis result of emergency abnormality; where N1 indicates that the correlation does not meet the characteristic element number threshold of the consistency baseline database, and the value is less than m.

4. The method for monitoring dynamic changes in land use based on multimodal remote sensing data according to claim 3, characterized in that: The step S400 includes: S401. Mark the area identified as having abnormal data quality as an unreliable data area. Perform data repair on the multimodal remote sensing real-time data in the unreliable data area by combining it with the multimodal remote sensing historical data for the same area. Calculate the real-time difference index Rs using the formula for the difference index R for the repaired multimodal remote sensing real-time data. If the real-time difference index Rs is less than or equal to the threshold R0, output a prompt message stating "Data repair successful." If the real-time difference index Rs is greater than the threshold R0, output a prompt message stating "Abnormalities still exist after data repair, please check further," and relevant personnel will take appropriate action. S402. Mark the area identified as an emergency abnormality as a high-priority change area, obtain the location information of the high-priority change area and output it to relevant personnel, who will then handle it accordingly.

5. A system for monitoring dynamic changes in land use based on multimodal remote sensing data, applied to the method for monitoring dynamic changes in land use based on multimodal remote sensing data according to any one of claims 1 to 4, characterized in that: The system includes: a data acquisition and preprocessing module, a consistency baseline construction module, a real-time data analysis module, and an abnormality diagnosis and processing module; The data acquisition and preprocessing module acquires multimodal remote sensing historical data within the monitoring period and geographic information data of the monitoring area; preprocesses the multimodal historical remote sensing data, and performs format unification and spatial registration on the geographic information data, so that the geographic information data and the multimodal historical remote sensing data are consistent in spatial and temporal dimensions; The consistency baseline construction module divides the monitoring area into multiple sub-areas and identifies the land use type of each sub-area based on its geographical data characteristics; analyzes the steady-state data of each land use type and establishes a consistency baseline database; The real-time data analysis module obtains multimodal remote sensing real-time data of different land use types in the monitoring area and compares it with the consistency baseline database; calculates the difference index between the real-time data characteristics and the baseline database characteristics, thereby preliminarily screening and marking suspected abnormal areas; The abnormality diagnosis and processing module analyzes suspected abnormal areas to identify data quality abnormalities or emergency abnormalities; identifies the type of abnormality by analyzing the characteristic correlation relationship of multimodal remote sensing real-time data; if the diagnosis is data quality abnormality, the unreliable data is repaired and the repaired data is re-compared; if the diagnosis is an emergency abnormality, it is marked as a high-priority change area, and the relevant location information is promptly provided to relevant personnel for processing.

6. The land use dynamic change monitoring system based on multimodal remote sensing data according to claim 5, characterized in that: The data acquisition and preprocessing module includes a data acquisition unit and a data preprocessing unit; The data acquisition unit acquires multimodal remote sensing historical data within the monitoring period and geographic information data of the monitoring area; the data preprocessing unit preprocesses the multimodal historical remote sensing data, unifies the format of the geographic information data and performs spatial registration, so that the geographic information data and the multimodal historical remote sensing data are consistent in spatial and temporal dimensions.

7. The land use dynamic change monitoring system based on multimodal remote sensing data according to claim 5, characterized in that: The consistency baseline construction model includes a land use type analysis unit and a consistency baseline construction unit; The land use type analysis unit divides the monitoring area into multiple sub-areas and identifies the land use type of each sub-area based on its geographical data characteristics; the consistency baseline construction unit analyzes the stable state data of each land use type and establishes a consistency baseline database.

8. The land use dynamic change monitoring system based on multimodal remote sensing data according to claim 5, characterized in that: The real-time data analysis module includes a consistency baseline database comparison unit and a suspected abnormal area identification unit; The consistency baseline database comparison unit obtains multimodal remote sensing real-time data of different land use types in the monitoring area and compares it with the consistency baseline database; the suspected abnormal area identification unit calculates the difference index between the real-time data characteristics and the baseline database characteristics, thereby preliminarily screening and marking the suspected abnormal areas; The abnormality diagnosis and processing module includes an abnormality diagnosis unit and a prompt information output unit; The abnormality diagnosis unit analyzes the suspected abnormal area to identify data quality abnormalities or emergency abnormalities; identifies the abnormality type by analyzing the feature correlation relationship of multimodal remote sensing real-time data; the prompt information output unit repairs the unreliable data if the data quality is diagnosed as abnormal, and re-compares the repaired data; For abnormalities diagnosed as emergencies, they are marked as high-priority change areas, and the relevant location information is promptly provided to relevant personnel for processing.

Images