A method for analyzing compliance of beach area use regulation based on multi-source data

Abstract

This invention relates to the field of multi-source data fusion and intelligent analysis technology, specifically a method for compliance analysis of beach area land use control based on multi-source data. The method includes: collecting data from multiple sources such as remote sensing images, GIS data, approval records, sensor monitoring data, and public opinion texts; introducing credibility calibration and spatiotemporal synchronization alignment to form a standardized beach area basic dataset; constructing a multi-dimensional correlation model of spatial morphology, behavioral disturbances, approval constraints, and temporal evolution characteristics to generate a unified land use feature vector; intelligently matching beach area land use characteristics with control rules; obtaining a comprehensive compliance score through multi-dimensional weighted fusion; identifying violation types and quantifying risk levels; predicting potential violation trends; triggering a three-level early warning based on anomaly detection; dynamically allocating management resources and generating hierarchical control strategies to achieve closed-loop optimization. This invention realizes intelligent dynamic compliance analysis of beach area land use control.

Description

Technical Field

[0001] This invention relates to the field of multi-source data fusion and intelligent analysis technology, specifically to a method for analyzing the compliance of beach area use control based on multi-source data. Background Technology

[0002] Beach areas are important transitional zones between river systems and the terrestrial environment, undertaking multiple functions such as flood control and retention, ecological protection, and resource utilization. Their land use control is of great significance to watershed safety and regional sustainable development. In recent years, with the rapid development of technologies such as remote sensing Earth observation, geographic information systems, Internet of Things sensor networks, and digital government approvals, the data sources available for beach area management have become increasingly abundant, covering heterogeneous information such as remote sensing images, spatial vector data, real-time sensor monitoring data, administrative approval records, and online public opinion texts. These data have their own characteristics in terms of spatiotemporal reference, collection granularity, structural form, and reliability, providing a multi-dimensional information foundation for comprehensively understanding the land use status of beach areas.

[0003] Chinese invention patent application CN120950615A discloses a dynamic monitoring method for land and space planning based on multi-source data fusion, belonging to the field of intelligent monitoring technology for land and space planning. The method includes: acquiring data to generate a multi-source heterogeneous dataset; performing spatiotemporal alignment processing using pre-set regional association rules of a planning knowledge base to generate a unified spatiotemporal dataset; constructing a dynamic knowledge graph centered on planning elements based on this dataset, with its node relationship weights updated in real time; guiding the multi-source data fusion direction through the graph relationship weights to generate a fusion feature vector; performing incremental learning monitoring processing on the feature vector and optimizing parameters, outputting the monitoring results of the planning implementation status; and updating the knowledge graph node relationship weights in a closed loop based on the monitoring results and simultaneously optimizing the incremental learning monitoring processing.

[0004] However, as water conservancy and natural resource management gradually moves towards refinement and intelligence, higher requirements are placed on the dynamic identification, compliance assessment, and early warning response of beach area use behavior. How to establish a unified spatiotemporal framework under the condition of multi-source heterogeneous data, realize the automatic extraction of use characteristics, computable matching with control rules, and form closed-loop management support has become a direction that current technological development needs to focus on. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the background technology by proposing a method for analyzing the compliance of beach area use control based on multi-source data.

[0006] The technical solution of this invention: a method for analyzing the compliance of beach area land use control based on multi-source data, comprising the following specific implementation steps: S1. Collect remote sensing images, GIS spatial data, administrative approval records, sensor monitoring data and public opinion text data related to the beach area from multiple sources. In the collection stage, introduce a data credibility calibration mechanism. Combine spatial coordinate unification, time axis normalization and semantic time back-inference processing to perform spatiotemporal synchronization alignment and consistency correction on multi-source data to form a standardized beach area basic dataset. S2. Based on the standardized beach area basic dataset, construct the correlation between spatial morphological features, behavioral disturbance features, approval constraint features and temporal evolution features to realize the dynamic expression of beach area use status and form a unified use feature vector that reflects use status, behavioral trends and development evolution process. S3. Intelligently match the characteristics of beach area use with the control rules to achieve dynamic quantitative analysis of beach area use control across the entire area. Obtain the comprehensive regional matching score through a multi-dimensional matching mechanism, identify the type of violation based on the matching results, quantify the risk level and predict potential violations by combining the trend of use evolution. S4. By combining multi-dimensional anomaly judgment with trend prediction, potential violations can be identified, three-level early warnings can be automatically triggered and a risk priority index can be generated. Management resources can be dynamically allocated, executable strategies can be generated, and continuous optimization can be achieved through closed-loop feedback.

[0007] Preferably, step S1 further includes: Data related to the beach area is collected synchronously from multiple sources. When the data enters the system, the credibility weight is calculated based on the data source quality, historical error and real-time stability, and the different types of data are initially classified and labeled. Spatial coordinate unification and temporal scale normalization are performed on data from different sources. Semantic time inference is performed on unstructured data such as public opinion texts in combination with sensor change signals to achieve consistent alignment of cross-modal data in the spatiotemporal dimension. A spatiotemporal consistency scoring model is constructed based on credibility weight and deviation calculation to quantify the conflict relationship between multi-source data. When the consistency is below the threshold, a time-first, credibility-first, and multi-temporal verification strategy is adopted to correct the conflicting data. After completing the conflict correction, all data are mapped to a unified spatiotemporal tensor structure to generate a standardized beach area basic dataset and unified feature output results.

[0008] Preferably, in step S1, when calculating the credibility weight, a comprehensive calculation is performed based on the data source quality score and the real-time stability factor. Semantic time inversion extracts event keywords related to beach area behavior from the text and matches them with abnormal sensor changes to invert the true time and location of the event. The spatiotemporal consistency scoring model employs an exponential decay mechanism to mitigate the impact of outlier data. During the conflict correction process, a sliding time window is used to verify conflicts between remote sensing and sensors, an approval validity delay factor is introduced for conflicts between approval and physical status, and conflicts between public opinion and physical data are filtered by the frequency of recurrence.

[0009] Preferably, step S2 further includes: Based on remote sensing images and GIS boundary data, the spatial structure of the beach area is dynamically reconstructed into regional units. By analyzing the texture gradient, edge continuity and surface reflection changes in the remote sensing images, the boundaries of human activities are identified. An adaptive segmentation mechanism with disturbed edges is adopted to make the regional units fit the actual use boundaries. Spatial morphological features are extracted and a regional spatial topology map is constructed. By integrating information from remote sensing time-series changes, abnormal sensor fluctuations, and public opinion events, a regional disturbance intensity model is constructed to dynamically identify the actual use behavior of the beach area. The use activities are classified and identified by combining the historical status of the area with the current disturbance pattern. The administrative approval documents are structured and parsed to extract the approved purpose, approval period, development intensity and restrictions, and an approval constraint vector is constructed. The real-time behavioral characteristics and the scope of approval are dynamically analyzed, and hidden violations are identified by calculating the approval behavior deviation coefficient. By integrating spatial morphological characteristics, behavioral disturbance characteristics, approval constraint characteristics, and temporal evolution characteristics, a beach area use evolution vector is constructed, generating a unified use characteristic library.

[0010] Preferably, in step S2, the spatial morphological features extracted during the dynamic region unit reconstruction process include region area, boundary complexity, texture dispersion, and neighborhood connectivity. The regional disturbance intensity model is used to jointly analyze the changes in remote sensing images, abnormal fluctuations in sensor values, and the activity level of public opinion events; the classification and identification of the use activities include temporary dumping activities, engineering construction activities, illegal reclamation activities, temporary road formation activities, ecological restoration activities, and intermittent occupation activities; The approval behavior deviation coefficient is used to quantify the difference between actual behavior and approval constraints. Abnormal deviation is triggered when continuous construction disturbance or large-area mechanical compaction marks are detected. The vector of beach area use evolution is integrated with temporal evolution characteristics, which are represented by a perturbation change trend function to mark the area as a continuously expanding area, an intermittently occupied area, or an ecological restoration area.

[0011] Preferably, step S3 further includes: Key features of the beach area control rules are extracted from unstructured text and graphic information and transformed into a unified vector form, including permitted use categories, constraint space, time period, development intensity and environmental constraints, and a weight matrix is ​​constructed. The spatial morphology, behavioral disturbances, approval constraints, and temporal evolution characteristics are matched with the rule vector in multiple dimensions. The multidimensional matching includes usage category matching, spatial range matching, behavioral intensity matching, and temporal constraint matching. The regional comprehensive compliance score is obtained through weighted fusion. Violation types are classified according to deviation thresholds of each dimension, and regional risk levels are quantified by combining time evolution trends. Current violations and potential evolution trends are identified through weighting functions and trend functions. Based on the type of violation and the level of risk, tiered control recommendations are generated.

[0012] Preferably, in step S3, the use category matching is performed by analyzing the intersection between the use categories identified by the behavioral perturbation and the rule-allowed use categories to obtain the degree of conformity; Spatial extent matching uses GIS spatial overlay analysis to calculate the intersection area ratio between beach area units and rule-constrained areas; Behavioral intensity matching calculates the compliance degree by comparing the disturbance intensity with the maximum development intensity allowed by the rules; Time-constrained matching compares the time evolution characteristics with the rule period to determine whether the behavior is within the allowed period; The comprehensive compliance matching score is obtained by weighting and integrating the matching results of each dimension according to their respective weights. The higher the score, the stronger the regional compliance. Violations include mismatch between intended use and category, exceeding the permitted space, exceeding the permitted intensity of the activity, and time-related violations; The risk level is dynamically calculated by combining the matching score with the trend of usage evolution.

[0013] Preferably, step S4 further includes: Abnormal behaviors are identified by comprehensively matching scores, behavioral trends, and regional sensitivity. Beach area units are divided into four categories: violation of usage category, violation of space occupation range, violation of behavior intensity limit, and violation of time period. At the same time, multi-label abnormal attributes are recorded. Dynamic hierarchical early warnings are generated based on abnormal behavior, risk level, and behavioral trends. Abnormalities are divided into three levels: high, medium, and low. Early warning information is automatically generated and can be pushed through multiple channels. Priority indices are calculated based on comprehensive factors such as violation type, risk level, trend evolution, and regional sensitivity. Resources for beach area management are dynamically allocated, with high-risk areas given priority for patrols, drone monitoring, and sensor deployment. Resource scheduling is optimized based on feedback data to achieve closed-loop adjustment. Based on abnormal behavior, risk level, trend prediction and resource priority, generate executable control strategies, including immediate intervention strategies, medium and long-term management strategies and preventive measures, and provide feedback on the implementation effect to achieve closed-loop management.

[0014] Preferably, in step S4, among the three levels of early warning, the first level early warning is high risk. It is triggered when the comprehensive compliance matching score is lower than the first matching threshold and the risk level is greater than the first risk threshold. Immediate on-site intervention and enhanced dynamic monitoring are required. Level 2 warning is a medium risk level. It is triggered when the comprehensive compliance matching score is between the first matching threshold and the second matching threshold, or when the risk level is between the first risk threshold and the second risk threshold. Short-term inspection and key monitoring are required. Level 3 warning is low risk. It is triggered when the comprehensive compliance matching score is greater than the second matching threshold and the risk level is less than the second risk threshold, and is included in daily monitoring. The priority index is calculated by taking into account the type of violation, risk level, speed of trend evolution, and regional sensitivity.

[0015] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects: This invention designs a method for compliance analysis of beach area land use control based on multi-source data. Through deep fusion and spatiotemporal alignment of multi-source heterogeneous data, it significantly improves the comprehensiveness and reliability of beach area land use control analysis. The introduction of data credibility calibration and spatiotemporal consistency scoring mechanisms effectively reduces the interference of errors or conflicts from a single data source on the analysis results, making the basic data layer more stable and reliable. Based on this, a correlation model is constructed for spatial morphology, behavioral disturbances, approval constraints, and temporal evolution characteristics. This model overcomes the limitations of traditional supervision, which only identifies changes in land features but cannot determine the nature of land use. It can dynamically identify complex land use activities such as construction, reclamation, temporary dumping, and ecological restoration, and achieve automatic deviation analysis between actual behavior and approved permits, promptly detecting hidden violations such as exceeding the scope, time limit, or disguised development. By intelligently matching and multi-dimensional weighted fusion with regulatory rules, it can generate a comprehensive regional compliance score, accurately classify violation types and quantify risk levels, and predict potential violations by combining land use evolution trends, thus shifting from passive discovery to proactive early warning. Through a three-level early warning mechanism and risk priority index, it dynamically allocates management resources such as patrol teams, drone monitoring, and sensor deployment, significantly improving regulatory efficiency and response speed. The generated hierarchical control strategies, including immediate intervention, medium- and long-term management, and preventive measures, support closed-loop feedback optimization, transforming the land use control of the beach area from a static, manual, and lagging model to a dynamic, intelligent, and predictive model, significantly reducing the risk of ecological damage and illegal occupation, and providing quantifiable and executable technical support for natural resource management and water conservancy supervision. Attached Figure Description

[0016] Figure 1 This is a flowchart of a method for analyzing the compliance of beach area use control based on multi-source data, as proposed in this invention. Detailed Implementation

[0017] Example 1, as Figure 1 As shown, the present invention proposes a method for compliance analysis of beach area land use control based on multi-source data, which includes the following specific implementation steps: S1. By collecting multi-source heterogeneous data such as remote sensing images, GIS spatial data, administrative approval records, sensor monitoring data and public opinion text data related to the beach area, and introducing a data credibility calibration mechanism during the collection stage, combined with spatial coordinate unification, time axis normalization and semantic time back-calculation processing, the multi-source data is spatiotemporally synchronized and consistently corrected to form a standardized beach area basic dataset. S2. Based on the standardized beach area basic dataset output in step S1, multi-dimensional feature modeling of beach area use status is performed. By constructing the correlation between spatial morphological features, behavioral disturbance features, approval constraint features and temporal evolution features, the dynamic expression of beach area use status is realized: beach area use units are reconstructed based on remote sensing texture changes and human disturbance edges, and then remote sensing, sensor and public opinion data are integrated to identify regional behavioral activities. Subsequently, the approval and permit content is mapped into computable constraint features and coupled with real-time behavior for analysis, and finally a unified use feature vector that can reflect the use status, behavioral trends and development evolution process is formed. S3. Intelligent matching of beach area use characteristics with control rules enables comprehensive, dynamic, and quantitative analysis of beach area use control. Textual and graphical control rules are transformed into vectors, and a rule weight matrix is ​​constructed, allowing for quantifiable processing of rules in a computer. Through a multi-dimensional matching mechanism, use category, spatial scope, behavioral intensity, and time constraints are weighted and fused with rule characteristics to obtain a comprehensive regional matching score. Then, based on the matching results, violation types are identified, and risk levels are quantified by combining use evolution trends, enabling prediction of potential violations. Intelligent reports are generated to provide management departments with real-time and actionable regulatory decision-making support. S4. By combining multi-dimensional anomaly judgment with trend prediction, potential violations are identified, a three-level early warning is automatically triggered, and a risk priority index is generated. Management resources such as patrol teams, drone monitoring, and sensor deployment are dynamically allocated, and finally, executable strategies are generated, including immediate intervention measures, medium and long-term management plans, and preventive monitoring schemes. These strategies are continuously optimized through closed-loop feedback.

[0018] In an optional embodiment, step S1 is used to construct the basic data layer for beach area use control compliance analysis. By performing distributed collection, credibility assessment, spatiotemporal unified mapping, and consistency correction on multi-source heterogeneous data, data from different sources, with different structures, and at different time scales can be computably fused within the same analytical framework, thereby providing stable input for subsequent use feature modeling. The specific implementation process is as follows: S11. Collect relevant data about the beach area simultaneously from multiple channels, including remote sensing imagery, GIS data, approval records, sensor data, and public opinion information. When the data enters the system, calculate the credibility weight based on the data source quality, historical error, and real-time stability. Perform initial classification and labeling of different types of data to control data quality differences from the source and provide a credibility constraint basis for subsequent fusion. Specifically: Periodically acquire multi-temporal remote sensing image data of the beach area from satellite remote sensing platform. Image data is downloaded in batches at fixed time intervals, and metadata such as imaging angle, cloud coverage, and sensor type are recorded simultaneously; at the same time, vector data of the beach area boundary and the current land use layer are extracted from the geographic information system. And ensure the consistency of boundary closure by verifying spatial topological relationships; On the government side, historical and real-time administrative approval data are obtained through interface services. This includes construction permits, temporary land use approvals, and records of changes in land use, and it also provides structured parsing of text-based approval content. On the IoT side, data streams from sensors such as water level, flow velocity, surface humidity, and soil disturbance are continuously received. The data is cached at a granular level of seconds or minutes; simultaneously, public opinion text data related to activities in the beach area are collected from publicly available information platforms. This includes news reports, complaint information, and social media content; To ensure the reliability of subsequent processing, an initial credibility weight is calculated when the data enters the system: ; In the formula, This represents the credibility weight of the k-th type of data source; This represents the quality score of the k-th data source, which is calculated based on historical error rate, data integrity rate, and manual sampling results. The real-time stability factor represents the k-th data source category, calculated based on real-time conditions such as current sampling frequency, network latency, and packet loss rate; n represents the total number of data source categories. S12. Spatial coordinate unification and temporal scale normalization are performed on data from different sources. Remote sensing and GIS data are mapped to a unified spatial benchmark using coordinate transformation functions. Simultaneously, all temporal information is normalized to a standard time axis. Furthermore, semantic temporal back-calculation is performed on unstructured data such as public opinion texts in conjunction with sensor change signals, achieving consistent alignment of cross-modal data across spatiotemporal dimensions. Specifically: At the spatial level, all data is uniformly transformed to a preset standard coordinate system (such as a local engineering coordinate system or the WGS84 unified projection system). The transformation process is achieved through a spatial mapping function. ; In the formula, Represents spatial coordinates in a unified coordinate system; Represents the original spatial coordinates; This represents a spatial coordinate transformation function, established based on the transformation relationship between the target coordinate system and the original coordinate system; At the time level, data with different time granularities are normalized, and all timestamps are mapped to a unified timeline: ; In the formula, t represents the original timestamp; Indicates the start time of the analysis time window, which is set according to the regulatory cycle; Indicates the end time of the analysis time window, which is set according to the current analysis cycle; Represents the normalized time coordinates; For unstructured data such as public opinion texts, a semantic time-based deduction mechanism is further introduced. Specifically, this involves extracting event keywords related to beach area activities from the text (such as "construction," "reclamation," and "drainage renovation") and matching them with abnormal sensor changes to deduce the true time and location of the event. ; In the formula, Indicates the unified time location corresponding to a public opinion event; Indicates a semantic time alignment function; S13. A spatiotemporal consistency scoring model is constructed based on credibility weight and deviation calculation to quantify the conflict relationship between multi-source data. An exponential decay mechanism is used to weaken the impact of outlier data. When consistency falls below a threshold, a time-first, credibility-first, and multi-temporal verification strategy is employed to correct conflicting data, thereby improving the stability of data fusion. Specifically: Calculate the degree of deviation between each data source and the current baseline state, and introduce an exponential decay function to construct a consistency score: ; In the formula, C represents the spatiotemporal consistency score of multi-source data; This represents the degree of deviation between the k-th type of data and the baseline state, calculated based on spatial location deviation, semantic difference, or time offset. This represents the conflict attenuation coefficient, which is set based on system experience parameters or historical training results. When the consistency score is below the threshold At this point, the conflict resolution process begins: First, prioritize data sources with the latest time series and high reliability; second, verify conflicts between remote sensing and sensors using a sliding time window, i.e., confirm whether changes persist through continuous multi-temporal data; for conflicts between approval and physical status, introduce an approval validity delay factor to determine whether there are any cases of non-implementation or delayed execution; for conflicts between public opinion and physical data, filter by recurrence frequency to eliminate the influence of individual false alarms or unstructured noise. S14. After completing the conflict correction, map all data to a unified spatiotemporal tensor structure, and perform feature extraction and weighted fusion on different types of data through a multi-source feature mapping function to generate a standardized beach area basic dataset and unified feature output results, specifically: After completing consistency correction, all data are mapped to a unified spatiotemporal representation space to construct a tensor structure for multi-source basic data in the beach area: ; The final preprocessed result is then generated using a multi-source fusion function: ; In the formula, This represents the unified set of multi-source data tensors; This represents remote sensing image data after spatiotemporal unification. This represents GIS spatial data after coordinate unification. This represents the approval data after the completion time has been standardized. Represents sensor data after a unified time scale; This represents the public opinion text data after time mapping; F represents the final spatiotemporal synchronization preprocessing result. The feature mapping function represents the k-th class of data, and corresponding feature extraction models are constructed based on different data types. This represents the k-th type of standardized data, generated from preprocessed data sources. The final output includes: a basic dataset F of the beach area under a unified spatiotemporal benchmark; a highly reliable multi-source fusion data structure with conflict correction; and a standardized feature input interface that can be directly used for beach area use modeling.

[0019] In an optional embodiment, step S2 establishes a cross-source behavior association mechanism to uniformly map the relationships between static landforms, dynamic disturbance behaviors, and approval constraints to the same feature space, thereby solving the problem in traditional beach area supervision that "only changes in land features are identified, but the nature of land use cannot be determined." The specific implementation process is as follows: S21. Based on the remote sensing imagery and GIS boundary data output in step S1, the spatial structure of the beach area is dynamically reconstructed into regional units. By analyzing the texture gradient, edge continuity, and surface reflection changes in the remote sensing imagery, human activity boundaries are identified. A perturbation edge adaptive segmentation mechanism is adopted to replace the traditional fixed grid division method, enabling regional units to conform to the actual use boundaries. Subsequently, spatial morphological features such as area, boundary complexity, texture dispersion, and neighborhood connectivity are extracted, and a regional spatial topology map is constructed to identify gradual use changes such as coastal expansion and local encroachment. Specifically: Reading remote sensing images after uniform coordinate correction It combines GIS boundary data to identify the natural edges of the beach area, the distribution of water systems, and the historical shoreline change trajectory; it also performs joint analysis on texture gradients, edge continuity, and reflectance changes in remote sensing images to automatically identify boundaries of human activities. An adaptive segmentation mechanism with disturbed edges is adopted, which uses traces of human activity as the basis for the formation of area boundaries. For example, when local soil piles, construction access roads, or enclosure structures appear in the beach area, the area is automatically separated from the original natural unit, making the area division more consistent with the actual use status. For each region unit Extract spatial morphological feature vectors: ; ; In the formula, This represents the i-th regional unit after the beach area is divided. Represents the spatial morphological feature vector of the i-th region; Indicates the area characteristics of a region, used to reflect the scale of regional development; Indicate boundary complexity; It represents the dispersion of surface texture, reflecting the surface heterogeneity and disturbance intensity within the region; The neighborhood connectivity index represents the spatial connectivity between a region and its neighboring regions, and is used to identify expansion trends. This represents the boundary length of the i-th region unit; This represents the area of ​​the i-th region unit; After the regional division is completed, a regional spatial topology map is further constructed to record the adjacency relationships, expansion directions and connection paths between regions, which will be used for subsequent analysis of gradual use change behaviors such as "expansion along the river" and "strip encroachment". S22. By integrating remote sensing temporal changes, sensor anomalies, and public opinion event information, the actual use of the beach area is dynamically identified; by constructing a regional disturbance intensity model, surface texture changes, continuous sensor anomalies, and semantic events are jointly analyzed, and combined with the historical status of the area, construction, reclamation, temporary dumping, and ecological restoration activities are identified, enabling early identification of short-cycle, weak-trace violations. Specifically: Regional behavioral disturbance information is extracted from the sensor data stream, remote sensing time-series change data, and public opinion event data output in step S1, and a regional disturbance intensity model is established: ; In the formula, This represents the overall disturbance intensity of the i-th region, used to characterize the degree of practical application activities; It represents the magnitude of change in remote sensing images, reflecting the changes in the regional surface at different points in time. That is, the image data after spatiotemporal alignment in step S1 is calculated jointly by texture, spectrum, shadow and micro-topography differences. The abnormal fluctuation value of the sensor represents the abnormal changes in environmental parameters such as water level, flow rate, and humidity. It is obtained after the sensor data acquisition results in step S1 are processed by time-series filtering and anomaly detection. It represents the activity level of public opinion events and is used to quantify the frequency and intensity of events involving regional activities in text information, namely the public opinion text data in step S1, which is calculated through keyword extraction and semantic clustering. , and The weights represent the fusion weights of the disturbance intensity, corresponding to the contributions of remote sensing, sensor, and public opinion data to the overall disturbance, respectively. They can be allocated based on the accuracy or reliability of historical data, or they can be dynamically adjusted. Among them, the change amplitude of remote sensing is not simply compared with the difference in grayscale, but also combines: changes in surface texture; changes in spectral reflectance; shadow movement characteristics; and changes in micro-topography elevation differences. For example, when construction machinery is in operation, even if the main surface has not changed significantly, the shadow structure and texture arrangement will change in advance, thus enabling early identification. For abnormal sensor data, a continuous sliding time window analysis mechanism is adopted. Only when the abnormality persists for multiple time windows is it considered a real disturbance. For example, short-term fluctuations in water level are not directly identified as reclamation activities, but when a combination of continuous surface vibration and decreased humidity occurs, it can be identified as construction activity. At the same time, the system performs semantic clustering on event keywords in public opinion data; For example, keywords such as "road construction," "filling," "fence," and "entry of construction machinery" will be mapped to "human development activities"; keywords such as "ecological restoration" and "vegetation restoration" will be mapped to "ecological governance activities." Subsequently, based on the region's historical status and current disturbance patterns, the regional use activities were categorized and identified, including: temporary dumping activities; engineering construction activities; illegal reclamation activities; temporary road construction activities; ecological restoration activities; and intermittent occupation activities. S23. The administrative approval text is structured and parsed to extract information such as approved purpose, approval period, development intensity, and restrictions. An approval constraint vector is constructed, and further, a dynamic deviation analysis is performed between real-time behavioral characteristics and the scope of approval. By calculating the approval behavior deviation coefficient, hidden violations such as construction beyond the approved scope, occupation beyond the approved period, and disguised development are identified. This achieves coupled analysis of approval constraints and actual usage behavior. Specifically: The approval document was structured and parsed to extract: approved usage category; approved area scope; approval period; permitted development intensity; temporary occupation conditions; and ecological protection restrictions. Construct the approval constraint vector: ; In the formula, This represents the feature vector of the approval constraints in the i-th region; Indicates the approved usage category, such as "ecological restoration" or "temporary construction"; It indicates the effective time limit for approval, representing the validity period of the use permit in terms of time; Indicates development intensity limits, characterizing the permitted intensity of construction or occupancy; This indicates special constraints, such as ecological protection restrictions or temporary occupation restrictions; Furthermore, an approval behavior deviation analysis mechanism is introduced to dynamically compare real-time behavioral characteristics with approval constraints: ; In the formula, The actual behavioral characteristic value of the i-th region is used for comparison with the approval constraints and is derived from the disturbance intensity and behavior type calculated in step S22. This represents the theoretical characteristic value of the approval and permit in the i-th region, used for comparison with actual behavior, and is derived from the approval vector. The calculations yield information including license type, strength, and duration. This represents the deviation coefficient of the approval behavior, quantifying the difference between actual behavior and approval constraints; For example, if the approval only permits temporary stockpiling, but continuous construction disturbance is detected, the deviation coefficient will continue to increase; if the approval permits low-intensity ecological restoration, but large-scale mechanical compaction marks appear in the area, it will also trigger abnormal deviations. S24. Spatial morphological characteristics, behavioral disturbance characteristics, approval constraint characteristics, and temporal evolution characteristics are uniformly integrated to construct a beach area use evolution vector. By analyzing regional disturbance change trends, the use statuses such as continuous expansion, periodic occupation, and ecological restoration are dynamically expressed, and a unified use characteristic library is generated, specifically: After completing the modeling of spatial morphology, behavioral disturbances, and approval constraints, the various features are further integrated to construct a unified feature vector for a single purpose: ; The temporal evolution characteristics are represented by a perturbation change trend function: ; In the formula, This represents the unified land use feature vector (beach area land use evolution vector), which serves as the input for subsequent land use control analysis; This indicates the characteristics of behavioral perturbations, i.e., the overall perturbation intensity. and identified usage categories ; It represents the characteristics of time evolution and is used to describe the rate and trend of changes in regional land use; This represents the intensity of regional disturbance in the k-th time period; Indicates the number of time windows for statistics, used to smooth out trend changes; When regional disturbances increase continuously over a long period, it is marked as a "continuously expanding region"; when disturbances fluctuate periodically, it is marked as a "partially occupied region"; when disturbances decrease rapidly and are accompanied by vegetation recovery, it is identified as an "ecological restoration region". Finally, a unified usage characteristic library is output, including: regional spatial morphology; actual usage behavior type; approval constraints; usage evolution trend; and regional dynamic risk characteristics.

[0020] In an optional embodiment, step S3 uses the beach area use evolution vector output in step S2. Using this as input, the spatial morphology, behavioral disturbances, approval constraints, and evolutionary trends of the beach area are intelligently matched with the beach area land use control rules to generate compliance analysis results for each area, including violation types, risk levels, and possible evolutionary trend predictions. The specific implementation process is as follows: S31. Extract key features from unstructured text and graphic information of the beach area control rules, and transform them into a unified vector form, including permitted use categories, constraint space, time period, development intensity, and environmental constraints. Simultaneously, construct a weight matrix to reflect the importance and sensitivity of the rules, specifically: Automatically reads regulatory documents, drawings, and approval records; performs semantic analysis on text descriptions, including keyword extraction, constraint parsing, and time limit identification; and performs spatial vectorization processing on illustrated boundaries, riverbanks, and protected area markings. For example, "No soil dumping in water buffer zone" will be interpreted as: Use category: soil dumping is prohibited; Space constraint: buffer zone range; Mandatory constraint level: high. Convert the parsing results into vector form: ; In the formula, The feature vectorization representation of the regulatory rules; This indicates the permitted use categories under the rules, such as "temporary soil stockpiling" or "ecological restoration"; A spatial region vector representing rule constraints, such as buffer zone range or water protection zone; Indicates the effective time or period of the rule, such as the permitted construction dates or protection period; This indicates the upper limit of permitted development intensity or behavior, such as the upper limit of soil piling height or the upper limit of construction area; This indicates additional environmental protection conditions, such as prohibiting mechanical compaction and restricting the occupation of water areas; Each rule is assigned a weight based on its mandatory level, regional sensitivity, and historical violation frequency to ensure that it is more sensitive to deviations in highly sensitive regions or high-priority rules in subsequent matching. S32. Multi-dimensional matching of spatial morphology, behavioral disturbances, approval constraints, and temporal evolution characteristics with rule vectors is performed, including matching by usage category, spatial scope, behavioral intensity, and temporal constraints. A weighted fusion is then used to obtain a comprehensive regional compliance score, specifically: The application evolution vector generated in step S2 With regular feature vectors Intelligent matching is performed to achieve a one-to-one correspondence analysis between area usage and regulatory rules: Match 1, i.e., usage category matching: the usage category identified from the behavioral perturbation. Category of permitted uses according to rules Performing an intersection analysis yields the degree of conformity in usage categories: ; A value closer to 1 indicates that the usage type fully complies with the rules, while a value closer to 0 indicates a serious mismatch. Matching 2, namely spatial range matching: using GIS spatial overlay analysis methods to match beach area units. With the constrained area of ​​the rules The intersection area ratio is calculated as follows: It can identify acts of encroachment, such as partial construction encroachment on the protective buffer zone; Matching 3, namely behavior intensity matching: matching the perturbation intensity With the rules allowing maximum development intensity Compare and calculate the degree of agreement: Through continuous numerical analysis, construction or stacking exceeding limits can be identified; Matching 4, namely time-constrained matching: matching time-evolutionary features With the rule cycle Compare and determine whether the behavior is within the allowed period: ; Calculate the overall matching score: The matching results from each dimension are weighted and combined. ; A higher score indicates stronger regional compliance, while a lower score indicates a greater potential risk of violation. In the formula, This indicates the matching degree of the usage category, with a value range of [0,1], where 1 represents a perfect match and 0 represents a complete mismatch. Indicates the degree of spatial range matching; This function represents the actual area calculation of a spatial region. Given a spatial polygon or GIS vector region, it returns the physical area covered by that region. Indicates the degree of behavioral intensity matching; This indicates the time constraint matching degree, where 1 represents compliance with the time constraint rule and 0 represents non-compliance. This indicates the overall compliance matching score of the beach area unit; , , and Indicates the weighting coefficient; S33. Classify violation types according to deviation thresholds for each dimension, such as inappropriate use, spatial boundary violations, excessive behavioral intensity, and temporal violations. Quantify regional risk levels by combining temporal evolution trends. Through weighted formulas and trend functions, identify current violations and predict potential evolution trends to achieve dynamic and continuous risk analysis. Specifically: Based on the matching score thresholds for each dimension, violation types are categorized as follows: Type A: Inconsistent application category ( ); Type B: Space usage exceeds the limit ( ); Type C: Behavioral intensity exceeds limits ( ); Type D: Time violation ( ); Considering the dynamic changes in matching scores and usage evolution trends, the risk level is calculated as follows: ; In the formula, , and This represents the set matching score threshold used to determine the type of violation. Match thresholds to usage categories. To match the threshold for the spatial range, The threshold for matching behavior intensity; Indicates the risk level of the beach area unit; Indicates the weight of the corresponding dimension; This represents a time-evolution weighted function for disturbances; a long-term upward trend increases the risk level, while cyclical fluctuations or downward trends decrease the risk level. By combining the evolution of disturbances over a continuous time window, the potential development trend of violations in the next 1-2 weeks or month can be predicted, and possible expansionary violations can be identified in advance. S34. Generate tiered control recommendations based on violation type and risk level: immediate on-site handling measures (e.g., fencing, patrols), adjustment of monitoring priority for adjacent areas, and optimization of medium- and long-term regulatory strategies (e.g., regional restrictions, adjustments to development approval). The report can be exported as an electronic file or pushed in real time through the management system.

[0021] In an optional embodiment, step S4 takes the compliance analysis results (violation type, risk level, and evolution trend) of the beach area unit output in step S3 as input, and realizes real-time monitoring and intelligent management of abnormal beach area use through anomaly identification, early warning triggering, risk priority ranking, and intelligent control strategy generation. The specific implementation process is as follows: S41. By comprehensively matching scores, behavioral trends, and regional sensitivity, abnormal behaviors are identified, and beach area units are divided into four categories: usage category violation, space occupation exceeding the scope, behavior intensity exceeding the limit, and time period violation. Simultaneously, multi-label abnormal attributes are recorded, including violation intensity, duration, and trend changes, specifically: Combined with comprehensive compliance matching score Beach area unit risk level and its time evolution characteristics Set multi-dimensional trigger conditions: When the overall compliance matching score When the value is below the set first matching threshold, it indicates that the current use of the beach area deviates significantly from the rule range; when the risk level of the beach area unit... When the risk level exceeds the set first risk threshold, it indicates that the potential risk in the area has reached the level of immediate concern; when the time evolution characteristics... When there are continuous upward trends over multiple periods or rapid abnormal fluctuations, it is marked as a potential expansionary violation. The beach area units are divided into four categories based on their violation characteristics: Application category violation: The actual use does not match the permitted use, such as piling soil or carrying out construction in an ecological buffer zone; Exceeding the permitted space: The actual activity area exceeds the boundary of the controlled area, such as some of the materials being piled up in the prohibited construction zone; Exceeding the limit of intensity of behavior: The quantity of behavior exceeds the control limit, such as the height of soil piling or the intensity of construction exceeding the specification; Time period violation: The activity occurred during a prohibited period, such as construction during the protection season; For each beach area unit, record anomaly labels, including violation type, violation intensity, duration, and trend change rate; S42. Generate dynamic tiered early warnings based on abnormal behavior, risk level, and behavioral trends. Abnormalities are categorized into high, medium, and low levels, and early warning information is automatically generated, including area identification, violation type, risk level, trend analysis, and handling suggestions. Multi-channel push notifications are supported, providing real-time operational guidance to management departments. Specifically: Transform abnormal behavior into actionable tiered alerts to achieve dynamic management and intelligent notifications: Level 1 Warning (High Risk): When the overall compliance matching score... The risk level of the beach area unit is below the set first matching threshold. When the risk exceeds the set first risk threshold, or when the time evolution characteristics continue to rise within multiple time windows, it is marked as high risk. At this time, the violation is serious and may cause immediate environmental impact or regulatory risks, requiring immediate on-site intervention (such as patrol teams, fencing, and drone deployment) and enhanced dynamic monitoring. Level 2 Warning (Medium Risk): When the comprehensive compliance matching score is between the first and second matching thresholds, or the risk level of the beach area unit is... If the risk level is between the first and second risk thresholds, it is marked as medium risk. At this point, the violation is obvious but has not caused serious impact immediately. The trend may be rising, requiring short-term inspections, key monitoring, trend tracking, and early intervention if necessary. Level 3 Warning (Low Risk): When the comprehensive compliance matching score is greater than the second matching threshold, and the risk level of the beach area unit is... When the risk level is below the second risk threshold, it is marked as low risk. At this point, the degree of violation is low, the risk is controllable or will not expand in the short term, and it needs to be included in daily monitoring and regular inspections are sufficient, without the need for immediate intervention. S43. A priority index is calculated based on the comprehensive violation type, risk level, trend evolution, and regional sensitivity to dynamically allocate beach area management resources. High-risk areas are prioritized for patrols, drone monitoring, and sensor deployment. Resource scheduling is optimized based on feedback data to achieve closed-loop adjustments, ensuring maximum resource utilization efficiency and continuously improving control effectiveness. Specifically: Taking into account the type of violation, risk level, speed of trend evolution, and regional sensitivity, a priority index is calculated: ; In the formula, This represents the risk priority index, a control priority indicator for beach area unit i, used to guide the allocation of management resources; , , and This indicates the risk priority weight, which corresponds to the proportion of risk level, behavioral trend, usage compliance, and regional sensitivity in the priority calculation; The regional sensitivity index of beach area unit i is automatically matched according to priority based on available management resources from beach area control policies, ecological protection levels, laws and regulations, or management experience. For example, core ecological protection areas, wetland protection areas, and coastal protection areas are assigned higher sensitivity values, while general development areas or low-sensitivity areas are assigned lower values. In this embodiment, the value range is 0~1: 0 represents a low-sensitivity area where the impact of violations or risk events is limited, and 1 represents a high-sensitivity area where any violation may cause serious consequences. Available management resources are automatically matched based on priority, including patrol teams, drones, video surveillance, and sensor deployment; resource usage is dynamically optimized, such as delaying patrols in low-risk areas and allocating multiple monitoring methods to high-risk areas; subsequent monitoring data is fed back to adjust priority indices and resource allocation strategies, achieving closed-loop optimization and ensuring optimal utilization of management resources; S44. Generate executable control strategies based on abnormal behavior, risk level, trend prediction, and resource priority, including immediate intervention, medium- and long-term management, and preventive measures, and provide feedback on implementation effectiveness. Optimize anomaly identification, early warning triggering, and resource scheduling to achieve closed-loop management of beach area use control, specifically: Based on anomaly type, risk level, trend prediction, and resource priority, a multi-level control strategy is generated: Real-time intervention strategies: on-site inspections, temporary fencing, and drone monitoring and control; Medium- and long-term management strategies: adjustment of development permits, regional restrictions, ecological restoration plans, and publicity of regulations; Preventive measures: Real-time monitoring of high-risk areas, optimization of alarm parameters, and increased data collection frequency; After the strategy is implemented, data from the beach area will continue to be collected, and the actual execution effect will be fed back to steps S1-S3 to optimize anomaly identification, early warning triggering, and resource scheduling, thereby achieving complete closed-loop management.

[0022] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for analyzing compliance of beach use regulation based on multi-source data, characterized in that, The specific implementation steps include the following: S1. Collect remote sensing images, GIS spatial data, administrative approval records, sensor monitoring data and public opinion text data related to the beach area from multiple sources. In the collection stage, introduce a data credibility calibration mechanism. Combine spatial coordinate unification, time axis normalization and semantic time back-inference processing to perform spatiotemporal synchronization alignment and consistency correction on multi-source data to form a standardized beach area basic dataset. S2. Based on the standardized beach area basic dataset, construct the correlation between spatial morphological features, behavioral disturbance features, approval constraint features and temporal evolution features to realize the dynamic expression of beach area use status and form a unified use feature vector that reflects use status, behavioral trends and development evolution process. S3. Intelligently match the characteristics of beach area use with the control rules to achieve dynamic quantitative analysis of beach area use control across the entire area. Obtain the comprehensive regional matching score through a multi-dimensional matching mechanism, identify the type of violation based on the matching results, quantify the risk level and predict potential violations by combining the trend of use evolution. S4. By combining multi-dimensional anomaly judgment with trend prediction, potential violations can be identified, three-level early warnings can be automatically triggered and a risk priority index can be generated. Management resources can be dynamically allocated, executable strategies can be generated, and continuous optimization can be achieved through closed-loop feedback.

2. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 1, characterized in that, Step S1 further includes: Data related to the beach area is collected synchronously from multiple sources. When the data enters the system, the credibility weight is calculated based on the data source quality, historical error and real-time stability, and the different types of data are initially classified and labeled. Spatial coordinate unification and temporal scale normalization are performed on data from different sources. Semantic time inference is performed on unstructured data such as public opinion texts in combination with sensor change signals to achieve consistent alignment of cross-modal data in the spatiotemporal dimension. A spatiotemporal consistency scoring model is constructed based on credibility weight and deviation calculation to quantify the conflict relationship between multi-source data. When the consistency is below the threshold, a time-first, credibility-first, and multi-temporal verification strategy is adopted to correct the conflicting data. After conflict correction, all data are mapped to a unified spatiotemporal tensor structure to generate a standardized beach area basic dataset and unified feature output results.

3. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 2, characterized in that, In step S1, when calculating the credibility weight, a comprehensive calculation is performed based on the data source quality score and the real-time stability factor. Semantic time inversion extracts event keywords related to beach area behavior from the text and matches them with abnormal sensor changes to invert the true time and location of the event. The spatiotemporal consistency scoring model employs an exponential decay mechanism to mitigate the impact of outlier data. During the conflict correction process, a sliding time window is used to verify conflicts between remote sensing and sensors, an approval validity delay factor is introduced for conflicts between approval and physical status, and conflicts between public opinion and physical data are filtered by the frequency of recurrence.

4. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 3, characterized in that, Step S2 further includes: Based on remote sensing images and GIS boundary data, the spatial structure of the beach area is dynamically reconstructed into regional units. By analyzing the texture gradient, edge continuity and surface reflection changes in the remote sensing images, the boundaries of human activities are identified. An adaptive segmentation mechanism with disturbed edges is adopted to make the regional units fit the actual use boundaries. Spatial morphological features are extracted and a regional spatial topology map is constructed. By integrating information from remote sensing time-series changes, abnormal sensor fluctuations, and public opinion events, a regional disturbance intensity model is constructed to dynamically identify the actual use behavior of the beach area. The use activities are classified and identified by combining the historical status of the area with the current disturbance pattern. The administrative approval documents are structured and parsed to extract the approved purpose, approval period, development intensity and restrictions, and an approval constraint vector is constructed. The real-time behavioral characteristics and the scope of approval are dynamically analyzed, and hidden violations are identified by calculating the approval behavior deviation coefficient. By integrating spatial morphological characteristics, behavioral disturbance characteristics, approval constraint characteristics, and temporal evolution characteristics, a beach area use evolution vector is constructed, generating a unified use characteristic library.

5. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 4, characterized in that, In step S2, during the dynamic region unit reconstruction process, the extracted spatial morphological features include region area, boundary complexity, texture dispersion, and neighborhood connectivity. The regional disturbance intensity model is used to jointly analyze the changes in remote sensing images, abnormal fluctuations in sensor values, and the activity level of public opinion events; the classification and identification of the use activities include temporary dumping activities, engineering construction activities, illegal reclamation activities, temporary road formation activities, ecological restoration activities, and intermittent occupation activities; The approval behavior deviation coefficient is used to quantify the difference between actual behavior and approval constraints. Abnormal deviation is triggered when continuous construction disturbance or large-area mechanical compaction marks are detected. The vector of beach area use evolution is integrated with temporal evolution characteristics, which are represented by a perturbation change trend function to mark the area as a continuously expanding area, an intermittently occupied area, or an ecological restoration area.

6. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 5, characterized in that, Step S3 further includes: Key features of the beach area control rules are extracted from unstructured text and graphic information and transformed into a unified vector form, including permitted use categories, constraint space, time period, development intensity and environmental constraints, and a weight matrix is ​​constructed. The spatial morphology, behavioral disturbances, approval constraints, and temporal evolution characteristics are matched with the rule vector in multiple dimensions. The multidimensional matching includes usage category matching, spatial range matching, behavioral intensity matching, and temporal constraint matching. The regional comprehensive compliance score is obtained through weighted fusion. Violation types are classified according to deviation thresholds of each dimension, and regional risk levels are quantified by combining time evolution trends. Current violations and potential evolution trends are identified through weighting functions and trend functions. Based on the type of violation and the level of risk, tiered control recommendations are generated.

7. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 6, characterized in that, In step S3, the use category matching is performed by analyzing the intersection between the use categories identified by the behavioral perturbation and the rule-permitted use categories to obtain the degree of conformity; Spatial extent matching uses GIS spatial overlay analysis to calculate the intersection area ratio between beach area units and rule-constrained areas; Behavioral intensity matching calculates the compliance degree by comparing the disturbance intensity with the maximum development intensity allowed by the rules; Time-constrained matching compares the time evolution characteristics with the rule period to determine whether the behavior is within the allowed period; The comprehensive compliance matching score is obtained by weighting and integrating the matching results of each dimension according to their respective weights. The higher the score, the stronger the regional compliance. Violations include mismatch between intended use and category, exceeding the permitted space, exceeding the permitted intensity of the activity, and time-related violations; The risk level is dynamically calculated by combining the matching score with the trend of usage evolution.

8. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 7, characterized in that, Step S4 further includes: Abnormal behaviors are identified by comprehensively matching scores, behavioral trends, and regional sensitivity. Beach area units are divided into four categories: violation of usage category, violation of space occupation range, violation of behavior intensity limit, and violation of time period. At the same time, multi-label abnormal attributes are recorded. Dynamic hierarchical early warnings are generated based on abnormal behavior, risk level, and behavioral trends. Abnormalities are divided into three levels: high, medium, and low. Early warning information is automatically generated and can be pushed through multiple channels. Priority indices are calculated based on comprehensive factors such as violation type, risk level, trend evolution, and regional sensitivity. Resources for beach area management are dynamically allocated, with high-risk areas given priority for patrols, drone monitoring, and sensor deployment. Resource scheduling is optimized based on feedback data to achieve closed-loop adjustment. Based on abnormal behavior, risk level, trend prediction and resource priority, generate executable control strategies, including immediate intervention strategies, medium and long-term management strategies and preventive measures, and provide feedback on the implementation effect to achieve closed-loop management.

9. The method for analyzing the compliance of beach area land use control based on multi-source data according to claim 8, characterized in that, In step S4, among the three levels of early warning, the first level early warning is high risk. It is triggered when the comprehensive compliance matching score is lower than the first matching threshold and the risk level is greater than the first risk threshold. Immediate on-site intervention and enhanced dynamic monitoring are required. Level 2 warning is a medium risk level. It is triggered when the comprehensive compliance matching score is between the first matching threshold and the second matching threshold, or when the risk level is between the first risk threshold and the second risk threshold. Short-term inspection and key monitoring are required. Level 3 warning is low risk. It is triggered when the comprehensive compliance matching score is greater than the second matching threshold and the risk level is less than the second risk threshold, and is included in daily monitoring. The priority index is calculated by taking into account the type of violation, risk level, speed of trend evolution, and regional sensitivity.

Images