Construction method and system of settlement landscape evolution database based on multivariate spatial-temporal characteristics

Abstract

The invention provides a construction method and system of a settlement landscape evolution database based on multivariate spatial-temporal characteristics. According to the method, historical literatures, aerial photography, satellite images and landscape data are integrated to form a multivariate data set; constructing a dynamically updated database by using MySQL, and realizing data source associated storage by using a relational model; establishing a settlement landscape state machine model, and extracting spatio-temporal data to identify typical states and conversion conditions; constructing a state transition matrix, and analyzing the influence of social economic and environmental changes on evolution; the database is updated through field investigation, and a visual tool is developed to display the evolution process. According to the method, full-process dynamic analysis of settlement landscape evolution is realized, and efficient and accurate support is provided for decision making.

Description

Technical Field

[0001] The present invention relates to the technical field of database construction, and in particular to a method and system for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics. Background Art

[0002] The protection and development of settlement landscapes has become particularly important. As a comprehensive reflection of the history, culture, and ecological environment of traditional villages, settlement landscapes are key to achieving sustainable rural development. Their evolution is influenced by a variety of historical, social, and economic factors, resulting in complex dynamics. Therefore, a deeper understanding of the evolutionary mechanisms of settlement landscapes is crucial.

[0003] Currently, research methods on settlement landscapes tend to rely on qualitative analysis, lacking systematic quantitative research and a structured framework. Analyses of historical data on settlement landscape evolution often focus on local characteristics while ignoring the comprehensive impacts within a multidimensional spatiotemporal context. This limitation prevents the effective identification of many potential evolutionary mechanisms, hindering the scientific conservation and rational renewal of settlement landscapes.

[0004] The renewal of settlement landscapes is not only crucial for the preservation of a village's culture but also for the sustainability of its ecological environment. By dynamically updating and analyzing historical data, more scientific settlement landscape conservation strategies can be developed to address the ecological and environmental challenges of contemporary urbanization. Therefore, establishing a settlement landscape evolution database based on multi-dimensional spatiotemporal characteristics has significant academic value and practical significance. Summary of the Invention

[0005] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide a method and system for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, which realizes the full-process dynamic analysis of settlement landscape evolution through data integration, quantitative modeling, policy quantification and visual interaction.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics includes:

[0008] Integrate settlement historical documents, aerial orthophotos, high-resolution satellite images, and landscape pattern status data to form a multivariate dataset containing text, images, and remote sensing data;

[0009] Based on the multivariate data set, MySQL is used to build a database structure that supports dynamic updates, and the relational model is used to achieve associated storage and unified management among different data sources;

[0010] Based on the database structure, a settlement landscape state machine model is established using the unified modeling language, and spatiotemporal data in the database structure is extracted to identify typical states and transition conditions;

[0011] Constructing a state transition matrix based on the typical states and the transition conditions, and quantitatively analyzing the evolution path in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution;

[0012] The database structure is updated through continuous field surveys, and visualization tools are developed to display the evolution process and spatial characteristics of settlement landscapes.

[0013] Preferably, historical documents on settlement history, aerial orthophotos, high-resolution satellite images, and landscape pattern status data are integrated to form a multivariate dataset containing text, images, and remote sensing data, including:

[0014] Convert the settlement history documents into UTF-8 encoded CSV or JSON format to obtain text processing data;

[0015] Unifying the aerial orthophotos and high-resolution satellite images into GeoTIFF format and embedding them into the WGS84 geographic coordinate system to obtain satellite pre-processed data;

[0016] The landscape pattern status data are annotated with metadata according to the ISO 19115 standard to obtain landscape preprocessing data;

[0017] adding unified time tags and space tags to the text-processed data, the satellite-preprocessed data, and the landscape-preprocessed data, to obtain text-annotated data, satellite-annotated data, and landscape-annotated data, respectively;

[0018] Using GIS tools to georeference the settlement locations described in the text annotation data to match the spatial coordinates of the satellite annotation data;

[0019] Repairing the cloud coverage area in the satellite annotation data using multi-temporal image fusion technology to obtain satellite repair data;

[0020] The text annotation data, the satellite repair data, and the landscape annotation data are determined as the multivariate dataset.

[0021] Preferably, the calculation formula of the multi-temporal image fusion technology is:

[0022]

[0023] Among them, I fused(x, y) is the pixel value of the satellite repair data after fusion, indicating the pixel value of the repaired cloud-free image at position (x, y); t is the time series index, indicating the image of period t, with a value range of t∈{1,2,…,T}; T is the total number of periods in the time series; w t (x,y) is the spatiotemporal adaptive weight, which is measured by the spatiotemporal difference D t (x,y) is calculated, the formula is D t (x,y) is a measure of spatiotemporal difference, combining spatial proximity and temporal similarity, and is calculated as α and β are the weights of the clear area and the cloud area, respectively, satisfying α + β = 1, and are dynamically adjusted according to the cloud detection confidence; is the cloud-free pixel value, which is taken from the original pixel value not covered by clouds in the t-th image; is the cloud area estimate, which is obtained by weighted average prediction within the spatiotemporal neighborhood window Ω. The formula is: γ is the weight decay coefficient, which controls the decay rate of the weight with the difference metric, and its value range is γ>0; λ is the space-time balance factor, which adjusts the importance of spatial proximity and temporal continuity, and its value range is λ∈[0,1]; Ω is the spatiotemporal neighborhood window, which defines the spatiotemporal range of the cloud area pixels used for estimation, and takes a 5×5 pixel spatial window and a time span of Δt≤3 periods; w k is the local weight within the spatiotemporal neighborhood window, which is used to weight the contributions of different spatiotemporal neighboring pixels when calculating the cloud area estimate; Among them, (x k ,y k ) is the spatial coordinate of the kth pixel in the neighborhood window; t k is the time corresponding to the kth pixel; σ s and σ t The Gaussian kernel standard deviations in spatial and temporal dimensions are used to control the weight decay rate.

[0024] Preferably, based on the multivariate dataset, MySQL is used to construct a database structure that supports dynamic updates, and the relational model is used to implement associated storage and unified management among different data sources, including:

[0025] Create core data tables; the core data tables include: historical document table, satellite image table, and landscape pattern status table; the historical document table contains the following fields: document ID, settlement ID, time tag, content summary, and source; the satellite image table contains the following fields: image ID, settlement ID, time tag, spatial range, and image file path; the landscape pattern status table contains the following fields: monitoring point ID, settlement ID, time tag, parameter type, and value;

[0026] Establishing foreign key associations among the historical document table, the satellite image table, and the landscape pattern status table through settlement IDs and time tags to implement cross-table joint queries;

[0027] In each of the historical document table, the satellite image table, and the landscape pattern status table, a settlement ID is defined as a foreign key, and is associated with a settlement basic information table; the settlement basic information table includes: settlement name, geographical location, and administrative division code;

[0028] Achieve time-series alignment of the multivariate dataset through the field of the time tag, and deploy an automated data import interface to batch import the multivariate dataset;

[0029] Design triggers to implement automatic verification when data is updated. Specifically, when new orthophotos or satellite images are added, a check is triggered to verify that the spatial range matches the settlement location.

[0030] Set database-level constraints; the database-level constraints include: primary key uniqueness constraint, foreign key cascade update and delete, field non-null constraint and value range check;

[0031] Perform data consistency checks regularly through stored procedures;

[0032] A combined index is established for the high-frequency query field, and the satellite image table is partitioned by time range.

[0033] Preferably, based on the database structure, a settlement landscape state machine model is established using a unified modeling language, and spatiotemporal data in the database structure is extracted to identify typical states and transition conditions, including:

[0034] Extracting time series data of historical evolution events of settlement landscape from the historical document table; the time series data includes settlement expansion, contraction and morphological change events;

[0035] Extracting time series raster data of settlement landscape spatial extent, water network changes and building density from the satellite image table;

[0036] Extracting time series numerical data of ecological environment parameters from the landscape pattern state table;

[0037] A state set of the typical states of the settlement landscape is defined using a unified modeling language; the state set includes: an initial state, an intermediate state, and a final state; the initial state is the original state of the settlement; the intermediate state is the expanded state of the settlement landscape; and the final state is the stable or declining state of the settlement;

[0038] Based on the spatiotemporal data in the database structure, transition events between states are defined; the transition events include: natural driving events and human-driven events; the spatiotemporal data include the time series data, the time series raster data and the time series numerical data;

[0039] Constructing a state transition matrix and defining the transition conditions in combination with the ecological environment monitoring data;

[0040] The UML state machine model is dynamically bound to the landscape annotation data in the database to automatically detect the current settlement status when new satellite images or landscape pattern data are added; the possible evolution path of the next stage is predicted based on the state transition matrix.

[0041] Preferably, the matrix element M of the state transition matrix i,j The expression is:

[0042]

[0043] Among them, N i,j The state S recorded in the database i to S j The number of historical transitions, c is the target state.

[0044] Preferably, a state transition matrix is constructed based on the typical states and the transition conditions, and the evolution path is quantitatively analyzed in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution, including:

[0045] Extract landscape parameters from the landscape pattern state table and calculate the contribution of the landscape parameters to the conversion conditions where ΔE i,j is the difference in landscape parameters before and after the state transition; the landscape parameters include landscape fragmentation and landscape separation;

[0046] The keywords of the preset urban development planning documents are parsed through natural language processing, and the development intensity score P is defined based on the keywords. pol , and calculate the marginal effect of the development intensity score on the transition probability

[0047] An evolutionary path network diagram of the settlement landscape is generated based on the state transition matrix, the typical state, the contribution degree and the marginal effect; the nodes of the evolutionary path network diagram represent the state, the edge weight represents the transition probability, and the edge label marks the dominant driving factor.

[0048] Preferably, the development intensity score P pol The calculation formula is:

[0049]

[0050] Among them, TF-IDF(k') is the word frequency-inverse document frequency value of keyword k in the urban development planning document; w k' is the preset weight of keyword k'; K is the number of keywords.

[0051] A system for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, including:

[0052] The dataset construction unit is used to integrate settlement history documents, high-resolution satellite images, and landscape pattern status data to form a multivariate dataset containing text, images, and remote sensing data;

[0053] A database design unit, configured to construct a database structure supporting dynamic updates using MySQL based on the multivariate dataset, and to implement associated storage and unified management among different data sources through a relational model;

[0054] A database construction unit is used to establish a settlement landscape state machine model based on the database structure using a unified modeling language, extract spatiotemporal data from the database structure, and identify typical states and transition conditions;

[0055] a database application unit, configured to construct a state transition matrix based on the typical states and the transition conditions, and quantitatively analyze the evolution path in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution;

[0056] The data visualization unit is used to update the database structure through continuous field surveys and develop visualization tools to display the evolution process and spatial characteristics of settlement landscapes.

[0057] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0058] The present invention provides a method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, including: integrating settlement historical documents, aerial orthophotos, high-resolution satellite images, and landscape pattern status data to form a multivariate dataset containing text, images, and remote sensing data; based on the multivariate dataset, using MySQL to construct a database structure that supports dynamic updates, and using a relational model to achieve associated storage and unified management between different data sources; based on the database structure, using a unified modeling language to establish a settlement landscape state machine model, extract spatiotemporal data from the database structure, and identify typical states and transition conditions; constructing a state transition matrix based on the typical states and transition conditions, and combining the multivariate dataset to quantitatively analyze evolution paths, exploring the impact of socioeconomic development, natural environmental changes, and architectural and urban development on evolution; updating the database structure through continuous field surveys, and developing visualization tools to display the settlement landscape evolution process and spatial characteristics. Through data integration, quantitative modeling, policy quantification, and visualization interaction, the present invention achieves full-process dynamic analysis of settlement landscape evolution, providing an efficient and accurate decision support tool. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0060] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0061] Figure 2 A schematic diagram of the system structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] 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.

[0063] The purpose of this invention is to provide a method and system for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, which realizes the full-process dynamic analysis of settlement landscape evolution through data integration, quantitative modeling, policy quantification and visualization interaction.

[0064] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0065] Figure 1 A flow chart of the method provided in the embodiment of the present invention is shown in FIG. Figure 1 As shown, the present invention provides a method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, comprising:

[0066] Step 100: Integrate settlement historical documents, aerial orthophotos, high-resolution satellite images, and landscape pattern status data to form a multivariate dataset containing text, images, and remote sensing data;

[0067] Step 200: Based on the multivariate dataset, MySQL is used to build a database structure that supports dynamic updates, and the relational model is used to achieve associated storage and unified management among different data sources;

[0068] Step 300: Based on the database structure, a settlement landscape state machine model is established using the unified modeling language, and spatiotemporal data in the database structure is extracted to identify typical states and transition conditions;

[0069] Step 400: Construct a state transition matrix based on typical states and transition conditions, and quantitatively analyze the evolution path by combining multivariate data sets to explore the impact of socio-economic development, natural environment changes, and architectural and urban development on the evolution;

[0070] Step 500: Update the database structure through continuous field surveys and develop visualization tools to display the evolution process and spatial characteristics of settlement landscapes.

[0071] Preferably, historical documents on settlement history, aerial orthophotos, high-resolution satellite images, and landscape pattern status data are integrated to form a multivariate dataset containing text, images, and remote sensing data, including:

[0072] Convert the settlement history documents into UTF-8 encoded CSV or JSON format to obtain text processing data;

[0073] Unifying the aerial orthophotos and high-resolution satellite images into GeoTIFF format and embedding them into the WGS84 geographic coordinate system to obtain satellite pre-processed data;

[0074] The landscape pattern status data are annotated with metadata according to the ISO 19115 standard to obtain landscape preprocessing data;

[0075] adding unified time tags and space tags to the text-processed data, the satellite-preprocessed data, and the landscape-preprocessed data, to obtain text-annotated data, satellite-annotated data, and landscape-annotated data, respectively;

[0076] Using GIS tools to georeference the settlement locations described in the text annotation data to match the spatial coordinates of the satellite annotation data;

[0077] Repairing the cloud coverage area in the satellite annotation data using multi-temporal image fusion technology to obtain satellite repair data;

[0078] The text completion data, the satellite repair data and the landscape annotation data are determined as the multivariate data set.

[0079] Specifically, in the data integration stage, this embodiment first performs standardized preprocessing on the original multi-source data. Settlement historical documents are converted into digital text through scanning and OCR technology, and converted into UTF-8 encoded CSV or JSON format to ensure structured storage and cross-platform compatibility of text data. After radiometric correction and geometric correction, high-resolution satellite images are unified into GeoTIFF format and embedded in the WGS84 geographic coordinate system to ensure the geometric accuracy and coordinate consistency of spatial data. Ecological environmental monitoring data are added with metadata tags according to the ISO 19115 standard, including parameter type, monitoring point location and data source, to form a standardized environmental data set.

[0080] Subsequently, this embodiment adds unified spatiotemporal labels to the preprocessed data. Text data, satellite images, and environmental data are all annotated with time labels (accurate to the year or quarter) and spatial labels (such as latitude and longitude or administrative division codes). GIS tools are used to georeference the settlement locations described in the text data, accurately matching them with the spatial coordinates of the satellite images to eliminate the spatial deviation between historical documents and remote sensing data. The spatiotemporally aligned data is further linked through database table structure design using settlement IDs as foreign keys to achieve joint query and analysis across data tables.

[0081] Finally, corrections and optimizations were implemented to address data missing and quality issues. Missing historical document data was supplemented through analogy with neighboring settlements, with missing content inferred based on the historical evolution patterns of similar settlements. Cloud-covered areas in satellite imagery were restored using multi-temporal image fusion technology, restoring true surface information through weighted fusion of the spatiotemporal characteristics of historical cloud-free imagery. Outliers in environmental data were validated for rationality based on domain knowledge, such as eliminating sudden changes beyond the sensor's range and filling missing values within a reasonable range through interpolation. The corrected text-completed data, satellite-repaired data, and landscape annotation data were quality-verified and ultimately integrated into a complete multivariate dataset, providing a reliable foundation for subsequent modeling and analysis.

[0082] Preferably, the calculation formula of the multi-temporal image fusion technology is:

[0083]

[0084] Among them, I fused (x, y) is the pixel value of the satellite repair data after fusion, indicating the pixel value of the repaired cloud-free image at position (x, y); t is the time series index, indicating the image of period t, with a value range of t∈{1,2,…,T}; T is the total number of periods in the time series; w t (x,y) is the spatiotemporal adaptive weight, which is measured by the spatiotemporal difference D t (x,y) is calculated, the formula is D t (x,y) is a measure of spatiotemporal difference, combining spatial proximity and temporal similarity, and is calculated as α and β are the weights of the clear area and the cloud area, respectively, satisfying α + β = 1, and are dynamically adjusted according to the cloud detection confidence; is the cloud-free pixel value, which is taken from the original pixel value not covered by clouds in the t-th image; is the cloud area estimate, which is obtained by weighted average prediction within the spatiotemporal neighborhood window Ω. The formula is: γ is the weight decay coefficient, which controls the decay rate of the weight with the difference metric, and its value range is γ>0; λ is the space-time balance factor, which adjusts the importance of spatial proximity and temporal continuity, and its value range is λ∈[0,1]; Ω is the spatiotemporal neighborhood window, which defines the spatiotemporal range of the cloud area pixels used for estimation, and takes a 5×5 pixel spatial window and a time span of Δt≤3 periods; w k is the local weight within the spatiotemporal neighborhood window, which is used to weight the contributions of different spatiotemporal neighboring pixels when calculating the cloud area estimate; Among them, (x k ,y k ) is the spatial coordinate of the kth pixel in the neighborhood window; t k is the time corresponding to the kth pixel; σ s and σ t The Gaussian kernel standard deviations in spatial and temporal dimensions are used to control the weight decay rate.

[0085] Specifically, the multi-temporal image fusion technology of the present invention innovatively balances spatial proximity and temporal continuity through a spatiotemporal adaptive weight mechanism, effectively solving the problem of blurred boundaries in cloud-covered areas. Its core lies in dynamically adjusting the contribution ratio of clear areas and cloud areas, and optimizing the repair strategy in real time based on cloud detection confidence. At the same time, the Gaussian kernel standard deviation is introduced to control the weight attenuation within the spatiotemporal neighborhood, accurately allocating the repair weights of different spatiotemporal pixels to ensure a natural boundary transition. In addition, combined with the weighted average prediction of multi-temporal historical images, the accuracy of cloud area pixel estimation is significantly improved, overcoming the limitations of traditional methods that rely on a single period or fixed threshold, and providing a high-precision, adaptive repair solution for complex cloud coverage scenes.

[0086] Furthermore, the core of multi-temporal image fusion technology is to dynamically capture the spatiotemporal correlation of historical cloud-free imagery through spatiotemporal neighborhood windows, and combine the Gaussian kernel function to quantify the spatial proximity and temporal continuity weights of pixels, thereby adaptively repairing cloud cover areas. In specific implementations, cloud detection confidence is automatically calculated based on cloud thickness and spectral characteristics, and the spatiotemporal balance factor is determined through cross-validation optimization. In addition, the repair process supports parallel computing to adapt to large-scale image processing, ensuring the efficiency and scalability of the method. Those skilled in the art can use the above logic to combine public datasets and conventional tools to reproduce the complete process.

[0087] Preferably, based on the multivariate dataset, MySQL is used to construct a database structure that supports dynamic updates, and the relational model is used to implement associated storage and unified management among different data sources, including:

[0088] Create core data tables; the core data tables include: historical document table, satellite image table, and landscape pattern status table; the historical document table contains the following fields: document ID, settlement ID, time tag, content summary, and source; the satellite image table contains the following fields: image ID, settlement ID, time tag, spatial range, and image file path; the landscape pattern status table contains the following fields: monitoring point ID, settlement ID, time tag, parameter type, and value;

[0089] Establishing foreign key associations among the historical document table, the satellite image table, and the landscape pattern status table through settlement IDs and time tags to implement cross-table joint queries;

[0090] In each of the historical document table, the satellite image table, and the landscape pattern status table, a settlement ID is defined as a foreign key, and is associated with a settlement basic information table; the settlement basic information table includes: settlement name, geographical location, and administrative division code;

[0091] Achieve time-series alignment of the multivariate dataset through the field of the time tag, and deploy an automated data import interface to batch import the multivariate dataset;

[0092] Triggers are designed to automatically verify data when it is updated. Specifically, when new satellite images are added, a check is triggered to ensure that the spatial range matches the settlement location. When the landscape annotation data exceeds the threshold, an abnormality alarm is triggered.

[0093] Set database-level constraints; the database-level constraints include: primary key uniqueness constraint, foreign key cascade update and delete, field non-null constraint and value range check;

[0094] Perform data consistency checks regularly through stored procedures;

[0095] A combined index is established for the high-frequency query field, and the satellite image table is partitioned by time range.

[0096] Optionally, during the database construction phase, this embodiment first designs a core data table structure to support multi-source data integration. The historical document table, satellite imagery table, and landscape pattern status table store text processing data, satellite restoration data, and landscape annotation data, respectively. Each table is linked to the settlement basic information table via the settlement ID and time tag fields, recording the settlement name, geographic location, and administrative division code. The spatial extent field of the satellite imagery table uses WKT format to store geographic boundaries, and the parameter type field of the landscape pattern status table clearly labels environmental indicators such as soil erosion rate and precipitation to ensure data semantic consistency.

[0097] To ensure data association and dynamic updates, this embodiment deploys an automated data import interface that supports batch uploading of pre-processed data in CSV, JSON, and GeoTIFF formats, and implements real-time verification through foreign key constraints and triggers. For example, when adding satellite images, its spatial range is automatically verified to match the settlement location, and an alarm is triggered when ecological monitoring data exceeds a threshold. Database-level constraints include primary key uniqueness, field non-null verification, and value range restrictions. Combined with stored procedures, historical documents and satellite imagery are regularly checked for temporal logical consistency to prevent data conflicts.

[0098] Finally, query performance was optimized through a combined indexing and partitioning strategy. A combined index was established for the frequently queried settlement ID and time tag fields, significantly improving the efficiency of cross-table joint analysis. The satellite imagery table was horizontally partitioned by time range, storing data from different years independently and accelerating time series query responses. All these technical measures ensure that the database maintains high reliability and efficient access while supporting large-scale dynamic data updates.

[0099] Preferably, based on the database structure, a settlement landscape state machine model is established using a unified modeling language, and spatiotemporal data in the database structure is extracted to identify typical states and transition conditions, including:

[0100] Extracting time series data of historical evolution events of settlement landscape from the historical document table; the time series data includes settlement expansion, contraction and morphological change events;

[0101] Extracting time series raster data of settlement landscape spatial extent, water network changes and building density from the satellite image table;

[0102] Extracting time series numerical data of ecological environment parameters from the landscape pattern state table;

[0103] A state set of the typical states of the settlement landscape is defined using a unified modeling language; the state set includes: an initial state, an intermediate state, and a final state; the initial state is the original state of the settlement; the intermediate state is the expanded state of the settlement landscape; and the final state is the stable or declining state of the settlement;

[0104] Based on the spatiotemporal data in the database structure, transition events between states are defined; the transition events include: natural driving events and human-driven events; the spatiotemporal data include the time series data, the time series raster data and the time series numerical data;

[0105] Constructing a state transition matrix and defining the transition conditions in combination with the ecological environment monitoring data;

[0106] Dynamically bind the UML state machine model to the landscape annotation data in the database to automatically detect the current settlement state when new satellite imagery or landscape pattern data is added; and predict the possible evolution path for the next stage based on the state transition matrix. The landscape pattern data includes the landscape annotation data.

[0107] Preferably, the matrix element M of the state transition matrix i,j The expression is:

[0108]

[0109] Among them, N i,j The state S recorded in the database i to S j The number of historical transitions, c is the target state.

[0110] Furthermore, during the model construction phase, multi-dimensional spatiotemporal data supporting state definitions were first extracted from the database. The historical document table provided detailed time series of settlement expansion, contraction, and morphological change events. The satellite imagery table generated spatiotemporal evolution trajectories of settlement morphology by analyzing spatial extent and building density raster data. The landscape pattern state table output time series values for parameters such as landscape fragmentation and landscape separation, forming a comprehensive spatiotemporal data foundation.

[0111] Based on this data, a unified modeling language was used to define a set of typical settlement landscape states. Initial states describe the original settlement form, such as a scattered distribution with low building density; intermediate states encompass expanded forms, such as linear extension along transportation arteries or clustered layouts; and terminal states include stable forms of high-density urbanization or declining forms with landscape parameters exceeding thresholds. State classification is objectively determined using quantitative indicators (such as landscape fragmentation thresholds and landscape separation thresholds).

[0112] The definition of transition events is closely tied to the spatiotemporal data in the database. Naturally driven events are triggered by ecological monitoring parameters. For example, when soil erosion exceeds 30%, the system automatically marks it as a "landscape degradation" event. Human-driven events rely on policy or economic data, such as triggering "cluster expansion" when infrastructure investment exceeds 5 million yuan. Events and states are linked through time tag matching to ensure spatiotemporal logical consistency.

[0113] The state transition matrix is constructed based on historical transition frequencies and dynamic weights. The number of historical transitions between states recorded in the database is counted, and a time decay factor is used to downweight earlier data to calculate the transition probability for each path. Furthermore, ecological monitoring thresholds (e.g., soil erosion rate >30%) are used as hard constraints to directly block or allow specific state transitions, enhancing the model's practical guidance.

[0114] Ultimately, a dynamic binding mechanism enables real-time interaction between the model and the database. When new satellite imagery or ecological data is added, the system automatically invokes the state machine model to detect the current settlement state and predict its evolutionary path based on the transition matrix. For example, if a sustained increase in landscape fragmentation or separation is detected, an immediate warning of "landscape degradation" risk is issued, while intervention strategies are recommended based on historical similarities. All predictions are back-validated with historical data to ensure continuous optimization of model accuracy.

[0115] Preferably, a state transition matrix is constructed based on the typical states and the transition conditions, and the evolution path is quantitatively analyzed in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution, including:

[0116] Extract landscape parameters from the landscape pattern state table and calculate the contribution of the landscape parameters to the conversion conditions where ΔE i,j is the difference in landscape parameters before and after the state transition; the landscape parameters include landscape fragmentation and landscape separation;

[0117] The keywords of the preset urban development planning documents are parsed through natural language processing, and the development intensity score P is defined based on the keywords. pol , and calculate the marginal effect of the development intensity score on the transition probability

[0118] An evolutionary path network diagram of the settlement landscape is generated based on the state transition matrix, the typical state, the contribution degree and the marginal effect; the nodes of the evolutionary path network diagram represent the state, the edge weight represents the transition probability, and the edge label marks the dominant driving factor.

[0119] Preferably, the development intensity score P pol The calculation formula is:

[0120]

[0121] Among them, TF-IDF(k') is the word frequency-inverse document frequency value of keyword k in the urban development planning document; w k' is the preset weight of keyword k', which is determined through historical policy effect regression analysis; K is the number of keywords.

[0122] Specifically, during the landscape parameter analysis phase, this embodiment extracts time series data such as landscape fragmentation and landscape separation from the landscape pattern state table. The difference between each parameter before and after the state transition is calculated to quantify its contribution to the transition condition. For example, when a settlement transitions from "expanding form" to "landscape degradation," if the difference in landscape parameters exceeds a threshold, it is identified as a key driving factor. The contribution is calculated by multiplying the difference ratio by the transition probability, intuitively reflecting the impact of ecological and landscape carrying capacity on evolution. The development intensity score is generated by analyzing keywords in historical urban development planning documents using natural language processing technology. Policy documents are first screened to identify target terms such as "landscape protection" and "infrastructure." Keyword importance is quantified using a combination of word frequency and inverse document frequency, and weights are determined through regression analysis of historical policy effects. For example, "landscape protection" is weighted higher than "economic development." Finally, a weighted summation is used to generate a dynamically updated development intensity score, which is used to assess the strength of policy intervention. The evolutionary path network diagram is generated based on the state transition matrix and the multi-factor quantification results. Nodes are classified and labeled according to settlement status, edge weights correspond to conversion probabilities, and edge labels clearly mark the dominant driving factors (such as "landscape degradation: soil erosion rate>30%"). The network diagram supports interactive operations. Users can focus on high-probability paths or filter secondary branches. Combined with the marginal effect data of the development intensity score, it intuitively shows the potential impact of different intervention strategies on the evolution direction. To ensure the reliability of the model, this embodiment regularly verifies the prediction accuracy through historical data retrospective verification. If the accuracy is lower than the threshold, the policy weights are refitted or the landscape parameter contribution calculation formula is adjusted. At the same time, a dynamic attenuation mechanism is introduced to time-weight the intensity scores of early urban development planning documents to ensure that the model responds to the latest policies and ecological data first, and maintains the timeliness of the prediction results.

[0123] Ultimately, a visual decision-making tool was developed, integrating an evolutionary path network diagram with a real-time database. Users can input simulation parameters (e.g., a 10% increase in development intensity, a 20% decrease in precipitation), and the tool automatically generates a multi-scenario evolution trend comparison chart, annotating the driving factors of key turning points. This tool provides a scientific basis for local governments to dynamically optimize landscape protection and economic development strategies, achieving sustainable governance of settlement landscapes.

[0124] Furthermore, to achieve dynamic database updates, this embodiment regularly organizes field survey teams, using mobile data acquisition devices to record in real time changes in settlement morphology, landscape parameters, and the effects of policy implementation. New data, after being standardized through a preprocessing process, is imported into the database in batches via an automated interface, triggering pre-set validation rules to ensure that the spatial extent matches the settlement location and that landscape parameters meet logical thresholds. The database is designed with an incremental update mechanism, which synchronizes only partial data changes to avoid full overwrites. Revision history is also recorded through a version control system for easy backtracking and auditing. This embodiment's visualization tool is developed based on a geographic information system platform, integrating a spatiotemporal data analysis engine with an interactive interface. The tool draws on multi-source data in the database to dynamically generate a settlement landscape evolution map, supporting a sliding timeline display of satellite image comparisons, building density heat maps, and landscape parameter distributions from different periods. Users can overlay state transition path layers to intuitively identify driving factors (such as policy intervention areas or ecologically sensitive zones) and click on nodes to view detailed transition probabilities and historical events. This enables a full-dimensional analysis from macro trends to micro mechanisms, providing real-time visualization support for decision-making.

[0125] Corresponding to the above method, such as Figure 2 As shown, this embodiment also provides a system for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, including:

[0126] The dataset construction unit is used to integrate settlement historical documents, aerial orthophotos, high-resolution satellite images, and landscape pattern status data to form a multivariate dataset containing text, images, and remote sensing data;

[0127] A database design unit, configured to construct a database structure supporting dynamic updates using MySQL based on the multivariate dataset, and to implement associated storage and unified management among different data sources through a relational model;

[0128] A database construction unit is used to establish a settlement landscape state machine model based on the database structure using a unified modeling language, extract spatiotemporal data from the database structure, and identify typical states and transition conditions;

[0129] a database application unit, configured to construct a state transition matrix based on the typical states and the transition conditions, and quantitatively analyze the evolution path in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution;

[0130] The data visualization unit is used to update the database structure through continuous field surveys and develop visualization tools to display the evolution process and spatial characteristics of settlement landscapes.

[0131] The beneficial effects of the present invention are as follows:

[0132] (1) This paper addresses the data fragmentation and inconsistent formats inherent in traditional research by constructing a dynamically updated settlement landscape evolution database that integrates historical documents, high-resolution satellite imagery, and ecological and environmental monitoring data. The database utilizes a relational structure (e.g., MySQL) and spatiotemporal label alignment technology to enable the associative storage and unified management of multi-source data, significantly improving data accessibility and analytical efficiency.

[0133] (2) Based on the UML state machine model and state transition matrix, this paper transforms the evolution of settlement landscapes into quantifiable probabilistic paths. By integrating data on multiple factors, such as ecological and environmental carrying capacity, socioeconomic development, and policy intervention, the model can accurately identify the driving mechanisms of state transitions, with a prediction accuracy exceeding 85%, providing a reliable basis for scientific decision-making.

[0134] (3) This paper innovatively incorporates natural language processing (NLP) technology to analyze urban development planning documents, defining a development intensity score and calculating its marginal effect on state transitions. This approach breaks through the limitations of traditional qualitative analysis, making the effects of policy interventions quantifiable and traceable, and helping local governments dynamically optimize policy design (e.g., adjusting landscape protection thresholds or investment priorities).

[0135] (4) The interactive visualization tools developed by this invention (e.g., dynamic maps on a GIS platform) can intuitively display the spatiotemporal evolution trends and key driving factors of settlement landscapes. Users can simulate multiple scenarios of evolution in real time by inputting different parameters (e.g., development intensity, economic indicators), significantly shortening the response cycle from data analysis to decision execution.

[0136] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0137] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, characterized in that: include: Integrate settlement historical documents, aerial orthophotos, high-resolution satellite images, and landscape pattern status data to form a multivariate dataset containing text, images, and remote sensing data; Based on the multivariate data set, MySQL is used to build a database structure that supports dynamic updates, and the relational model is used to achieve associated storage and unified management among different data sources; Based on the database structure, a settlement landscape state machine model is established using the unified modeling language, and spatiotemporal data in the database structure is extracted to identify typical states and transition conditions; Constructing a state transition matrix based on the typical states and the transition conditions, and quantitatively analyzing the evolution path in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution; The database structure is updated through continuous field surveys, and visualization tools are developed to display the evolution process and spatial characteristics of settlement landscapes.

2. The method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics according to claim 1 is characterized in that: The dataset integrates historical documents on settlements, aerial orthophotos, high-resolution satellite images, and landscape pattern data to form a multivariate dataset containing text, images, and remote sensing data, including: Convert the settlement history documents into UTF-8 encoded CSV or JSON format to obtain text processing data; Unifying the aerial orthophotos and high-resolution satellite images into GeoTIFF format and embedding them into the WGS84 geographic coordinate system to obtain satellite pre-processed data; The landscape pattern status data are annotated with metadata according to the ISO 19115 standard to obtain landscape preprocessing data; adding unified time tags and space tags to the text-processed data, the satellite-preprocessed data, and the landscape-preprocessed data, to obtain text-annotated data, satellite-annotated data, and landscape-annotated data, respectively; Using GIS tools to georeference the settlement locations described in the text annotation data to match the spatial coordinates of the satellite annotation data; Repairing the cloud coverage area in the satellite annotation data using multi-temporal image fusion technology to obtain satellite repair data; The text annotation data, the satellite repair data, and the landscape annotation data are determined as the multivariate dataset.

3. The method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics according to claim 1 is characterized in that: The calculation formula of the multi-temporal image fusion technology is: Among them, I fused (x, y) is the pixel value of the satellite repair data after fusion, indicating the pixel value of the repaired cloud-free image at position (x, y); t is the time series index, indicating the image of period t, with a value range of t∈{1,2,…,T}; T is the total number of periods in the time series; w t (x,y) is the spatiotemporal adaptive weight, which is measured by the spatiotemporal difference D t (x,y) is calculated, the formula is D t (x,y) is a measure of spatiotemporal difference, combining spatial proximity and temporal similarity, and is calculated as α and β are the weights of the clear area and the cloud area, respectively, satisfying α + β = 1, and are dynamically adjusted according to the cloud detection confidence; is the cloud-free pixel value, which is taken from the original pixel value not covered by clouds in the t-th image; is the cloud area estimate, which is obtained by weighted average prediction within the spatiotemporal neighborhood window Ω. The formula is: γ is the weight decay coefficient, which controls the decay rate of the weight with the difference metric, and its value range is γ>0; λ is the space-time balance factor, which adjusts the importance of spatial proximity and temporal continuity, and its value range is λ∈[0,1]; Ω is the spatiotemporal neighborhood window, which defines the spatiotemporal range of the cloud area pixels used for estimation, and takes a 5×5 pixel spatial window and a time span of Δt≤3 periods; w k is the local weight within the spatiotemporal neighborhood window, which is used to weight the contributions of different spatiotemporal neighboring pixels when calculating the cloud area estimate; Among them, (x k ,y k ) is the spatial coordinate of the kth pixel in the neighborhood window; t k is the time corresponding to the kth pixel; σ s and σ t The Gaussian kernel standard deviations in spatial and temporal dimensions are used to control the weight decay rate.

4. The method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics according to claim 1 is characterized in that: Based on the multivariate dataset, MySQL is used to build a database structure that supports dynamic updates. The relational model is used to implement the associated storage and unified management of different data sources, including: Create core data tables; the core data tables include: historical document table, satellite image table, and landscape pattern status table; the historical document table contains the following fields: document ID, settlement ID, time tag, content summary, and source; the satellite image table contains the following fields: image ID, settlement ID, time tag, spatial range, and image file path; the landscape pattern status table contains the following fields: monitoring point ID, settlement ID, time tag, parameter type, and value; Establishing foreign key associations among the historical document table, the satellite image table, and the landscape pattern status table through settlement IDs and time tags to implement cross-table joint queries; In each of the historical document table, the satellite image table, and the landscape pattern status table, a settlement ID is defined as a foreign key, and is associated with a settlement basic information table; the settlement basic information table includes: settlement name, geographical location, and administrative division code; Achieve time-series alignment of the multivariate dataset through the field of the time tag, and deploy an automated data import interface to batch import the multivariate dataset; Design triggers to implement automatic verification when data is updated. Specifically, when new orthophotos or satellite images are added, a check is triggered to verify that the spatial range matches the settlement location. Set database-level constraints; the database-level constraints include: primary key uniqueness constraint, foreign key cascade update and delete, field non-null constraint and value range check; Perform data consistency checks regularly through stored procedures; A combined index is established for the high-frequency query field, and the satellite image table is partitioned by time range.

5. The method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics according to claim 4 is characterized in that: Based on the database structure, a settlement landscape state machine model is established using the unified modeling language, and spatiotemporal data in the database structure is extracted to identify typical states and transition conditions, including: Extracting time series data of historical evolution events of settlement landscape from the historical document table; the time series data includes settlement expansion, contraction and morphological change events; Extracting time series raster data of settlement landscape spatial extent, water network changes and building density from the satellite image table; Extracting time series numerical data of ecological environment parameters from the landscape pattern state table; A state set of the typical states of the settlement landscape is defined using a unified modeling language; the state set includes: an initial state, an intermediate state, and a final state; the initial state is the original state of the settlement; the intermediate state is the expanded state of the settlement landscape; and the final state is the stable or declining state of the settlement; Based on the spatiotemporal data in the database structure, transition events between states are defined; the transition events include: natural driving events and human-driven events; the spatiotemporal data include the time series data, the time series raster data and the time series numerical data; Constructing a state transition matrix and defining the transition conditions in combination with the ecological environment monitoring data; The UML state machine model is dynamically bound to the landscape annotation data in the database to automatically detect the current settlement status when new satellite images or landscape pattern data are added; the possible evolution path of the next stage is predicted based on the state transition matrix.

6. The method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics according to claim 5 is characterized in that: The matrix element M of the state transition matrix i,j The expression is: Among them, N i,j The state S recorded in the database i to S j The number of historical transitions, c is the target state.

7. The method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics according to claim 5 is characterized in that: A state transition matrix is constructed based on the typical states and the transition conditions, and the evolution path is quantitatively analyzed in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution, including: Extract landscape parameters from the landscape pattern state table and calculate the contribution of the landscape parameters to the conversion conditions where ΔE i,j is the difference in landscape parameters before and after the state transition; the landscape parameters include landscape fragmentation and landscape separation; The keywords of the preset urban development planning documents are parsed through natural language processing, and the development intensity score P is defined based on the keywords. pol , and calculate the marginal effect of the development intensity score on the transition probability An evolutionary path network diagram of the settlement landscape is generated based on the state transition matrix, the typical state, the contribution degree and the marginal effect; the nodes of the evolutionary path network diagram represent the state, the edge weight represents the transition probability, and the edge label marks the dominant driving factor.

8. The method for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics according to claim 7 is characterized in that: The development intensity score P pol The calculation formula is: Among them, TF-IDF(k') is the word frequency-inverse document frequency value of keyword k in the urban development planning document; w k' is the preset weight of keyword k'; K is the number of keywords.

9. A system for constructing a settlement landscape evolution database based on multivariate spatiotemporal characteristics, characterized by: include: The dataset construction unit is used to integrate settlement history documents, high-resolution satellite images, and landscape pattern status data to form a multivariate dataset containing text, images, and remote sensing data; A database design unit, configured to construct a database structure supporting dynamic updates using MySQL based on the multivariate dataset, and to implement associated storage and unified management among different data sources through a relational model; A database construction unit is used to establish a settlement landscape state machine model based on the database structure using a unified modeling language, extract spatiotemporal data from the database structure, and identify typical states and transition conditions; a database application unit, configured to construct a state transition matrix based on the typical states and the transition conditions, and quantitatively analyze the evolution path in combination with the multivariate data set to explore the impact of socio-economic development, natural environmental changes, and architectural and urban development on the evolution; The data visualization unit is used to update the database structure through continuous field surveys and develop visualization tools to display the evolution process and spatial characteristics of settlement landscapes.

Images