Planned plot change monitoring method and system based on time sequence remote sensing image

Abstract

The invention discloses a planning plot change monitoring method and system based on a time sequence remote sensing image, and belongs to the technical field of image data processing and analysis, and the method comprises the steps: obtaining a multi-temporal remote sensing image and planning plot data; preprocessing the multi-temporal remote sensing image to generate preprocessed time sequence remote sensing data; performing structured processing on the planned plot data to generate a structured planning knowledge base; based on the preprocessed time sequence remote sensing data and the structured planning knowledge base, performing change sensitivity pre-evaluation on each planning land parcel to generate a change sensitivity parameter of each planning land parcel; dynamically distributing analysis strategy parameters for each planning land parcel according to the change sensitivity parameters; and performing layered change detection on the planned plot based on the analysis strategy parameters to generate a plot change monitoring result. According to the invention, the time sequence remote sensing image and the planning land parcel data are adopted, and the change sensitivity pre-evaluation and dynamic analysis strategy are combined, so that the automatic and precise change monitoring of the planning land parcel can be realized.

Description

Technical Field

[0001] This invention relates to the field of image data processing and analysis technology, and in particular to a method and system for monitoring changes in planned land parcels based on time-series remote sensing imagery. Background Technology

[0002] Temporal remote sensing image change monitoring technology utilizes multiple remote sensing images covering the same area acquired at different times. Through image processing and analysis, it identifies areas where land cover type and status have changed. In the fields of urban planning and natural resource management, this technology is applied to land use change monitoring, urban expansion analysis, and the detection of illegal construction. It is an important means of obtaining dynamic information on national land space and supporting the supervision of planning implementation.

[0003] In related technologies, Chinese invention patent CN112053359B discloses a method, device, electronic device, and storage medium for detecting changes in remote sensing images. The method includes: preprocessing periodic time-series remote sensing images by radiometric correction, orthorectification, resampling, registration, cloud and shadow removal, and band or index extraction to obtain pixel-by-pixel and band or index-by-band periodic time-series data; using an empirically defined LSTM network structure, learning pixel-by-pixel from the time-series data prior to the monitoring period to obtain a pixel-level time-series evolution LSTM model; using each model to predict the band or index data for each pixel during the monitoring period, and calculating the difference between the model predictions and the actual monitoring values ​​to obtain a non-periodic change marker map for each band or index during the monitoring period; and using a classifier trained on samples of changes of interest to classify all non-periodic change marker maps to obtain the classification detection result of the change of interest.

[0004] However, the aforementioned existing technical solutions have the following technical shortcomings. First, they lack targeted analysis guided by planning semantics. Existing technologies focus on pixel-level time-series modeling and change detection, but do not incorporate planned land parcel data as analysis units, making it impossible to associate them with semantic information such as specific planning uses and control indicators. This results in the system only being able to identify physical changes, unable to determine whether changes conform to planning intent or constitute illegal construction, failing to meet the deeper needs of "compliance" monitoring in urban planning management. Second, the uniform allocation of resources makes it difficult to balance efficiency and accuracy. Existing technologies use a uniform LSTM modeling and classification process for all pixels, failing to differentiate the change risk levels of different areas. For large-scale monitoring scenarios, this approach leads to wasted computational resources in low-risk areas, while high-risk areas may experience missed reports due to insufficient analysis depth, failing to achieve a balance between monitoring efficiency and accuracy. Furthermore, the output of this method is a pixel-level change classification map, not overlaid with spatial data such as planning control lines and basic geographic elements, and lacks visualization and statistical analysis functions tailored to planning management needs. The monitoring results are difficult to directly use for spatial control and decision support, requiring additional manual interpretation and integration. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a method and system for monitoring changes in planned land parcels based on time-series remote sensing imagery. By employing time-series remote sensing imagery and planned land parcel data, combined with change sensitivity pre-assessment and dynamic analysis strategies, it is possible to achieve automated and precise change monitoring of planned land parcels.

[0006] The above objectives can be achieved through the following approach: A method and system for monitoring changes in planned land parcels based on time-series remote sensing imagery includes: acquiring multi-temporal remote sensing imagery and planned land parcel data; preprocessing the multi-temporal remote sensing imagery to generate preprocessed time-series remote sensing data; performing structured processing on the planned land parcel data to generate a structured planning knowledge base; based on the preprocessed time-series remote sensing data and the structured planning knowledge base, performing a pre-assessment of change sensitivity for each planned land parcel to generate change sensitivity parameters for each planned land parcel; dynamically assigning analysis strategy parameters to each planned land parcel according to the change sensitivity parameters; and performing hierarchical change detection on the planned land parcels based on the analysis strategy parameters to generate land parcel change monitoring results.

[0007] Optionally, the step of performing a change sensitivity pre-assessment for each planned land parcel and generating change sensitivity parameters for each planned land parcel includes: extracting historical fluctuation characteristics of each planned land parcel from the preprocessed time-series remote sensing data to generate historical stability parameters; extracting policy texts associated with the planned land parcels from the structured planning knowledge base and parsing the regulatory intensity of the policy texts to generate policy sensitivity parameters; obtaining verified change information of adjacent planned land parcels and calculating the spatial correlation between the current planned land parcel and the adjacent planned land parcels to generate neighborhood correlation parameters; and integrating the historical stability parameters, policy sensitivity parameters, and neighborhood correlation parameters to calculate change sensitivity parameters.

[0008] Optionally, the step of dynamically allocating analysis strategy parameters for each planned land parcel includes: classifying each planned land parcel into different sensitivity levels based on the value of the change sensitivity parameter; obtaining the analysis time window parameter and analysis intensity level parameter corresponding to each sensitivity level; and generating analysis strategy parameters by combining the corresponding analysis time window parameter and analysis intensity level parameter based on the sensitivity level to which each planned land parcel belongs.

[0009] Optionally, the step of performing layered change detection on the corresponding planned land parcels includes: performing feature comparison on the latest two periods of remote sensing images of all planned land parcels, performing physical change screening, and generating preliminary change areas; identifying the analysis strategy parameters with deep analysis instructions, and performing semantic-temporal coupling analysis on the planned land parcels holding such instructions to obtain semantic analysis conclusions; and using the semantic analysis conclusions to verify and filter the preliminary change areas to generate verified change patches.

[0010] Optionally, the semantic-temporal coupling analysis includes: obtaining standard temporal change trajectories corresponding to planning semantics from the structured planning knowledge base; extracting the actual temporal change trajectory of the current planning plot from the preprocessed temporal remote sensing data; performing similarity matching between the actual temporal change trajectory and the standard temporal change trajectory to generate a semantic consistency score; identifying change types based on the semantic consistency score to obtain semantic analysis conclusions.

[0011] Optionally, the method further includes: obtaining manual verification results and generating a verification sample set; comparing the land parcel change monitoring results with the verification sample set, identifying misjudgment cases, and generating a misjudgment analysis report; and adjusting the weights of the evaluation model used to generate the change sensitivity parameter based on the misjudgment analysis report to generate an updated evaluation model.

[0012] Optionally, the preprocessing of the multi-temporal remote sensing images to generate preprocessed time-series remote sensing data includes: performing radiometric calibration and atmospheric correction on the multi-temporal remote sensing images to obtain surface reflectance data; and performing geometric fine correction and image registration on the surface reflectance data to generate preprocessed time-series remote sensing data.

[0013] Optionally, the step of structuring the planned land parcel data to generate a structured planning knowledge base includes: parsing the planned land parcel data for spatial geometric information and attribute information to obtain a set of planning attribute tags; and performing semantic classification and relational modeling on the set of planning attribute tags to generate a structured planning knowledge base.

[0014] Optionally, after generating the land parcel change monitoring results, the method further includes: converting the land parcel change monitoring results into vector data supported by a geographic information system to generate a visual monitoring layer; acquiring planning control lines and basic geographic features, and overlaying the visual monitoring layer with the planning control lines and the basic geographic features to generate a comprehensive monitoring map; providing spatial query and statistical analysis functions based on the comprehensive monitoring map to generate a decision support interface.

[0015] Based on the same inventive concept, this invention also provides a planning land parcel change monitoring system based on time-series remote sensing imagery. The system includes: a data acquisition module for acquiring multi-temporal remote sensing images and planning land parcel data; an image preprocessing module for preprocessing the multi-temporal remote sensing images to generate preprocessed time-series remote sensing data; a planning knowledge base construction module for structuring the planning land parcel data to generate a structured planning knowledge base; a change sensitivity pre-assessment module for performing a change sensitivity pre-assessment on each planning land parcel based on the preprocessed time-series remote sensing data and the structured planning knowledge base, generating change sensitivity parameters for each planning land parcel; a parameter dynamic allocation module for dynamically allocating analysis strategy parameters to each planning land parcel according to the change sensitivity parameters; and a monitoring result output module for performing layered change detection on the corresponding planning land parcels based on the analysis strategy parameters, generating land parcel change monitoring results.

[0016] Compared with the prior art, the present invention has the following advantages: This invention resolves the fundamental contradiction in traditional monitoring methods—the difficulty of balancing high precision and high efficiency—by introducing a dynamic focus decision-making mechanism and a hierarchical change detection engine. Instead of performing brute-force analysis on all plots with uniform intensity, the system intelligently and selectively allocates limited computing resources to the most critical high-risk plots through change sensitivity pre-assessment. This improves overall data processing efficiency while maintaining monitoring accuracy in key areas, enabling highly timely monitoring of large-scale regions.

[0017] This invention endows the monitoring system with the ability to learn from errors and self-evolve by constructing a closed-loop adaptive learning mechanism. When the system makes a misjudgment, it can automatically adjust the parameters of its internal evaluation model by comparing it with external validation data; when encountering unknown new change patterns, it can characterize them and incorporate them into the knowledge base. This self-learning capability ensures that the system can continuously optimize its performance during long-term operation, automatically adapt to the dynamic evolution of planning policies and the surface environment, and enhance the system's robustness and intelligence.

[0018] This invention elevates monitoring capabilities from simple physical change detection to a deeper level of planning semantic understanding and situation assessment by implementing semantic-temporal coupling analysis and linkage with macro-planning indicators. The system can not only identify changes in land parcels, but also understand the specific types of changes and whether they align with planning intentions. Furthermore, through aggregated analysis of regional change trends, it can provide early warnings of the potential impact of the cumulative effects of local changes on macro-planning objectives, offering forward-looking decision support for the scientific management and dynamic adjustment of urban planning.

[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating a method for monitoring changes in planned land parcels based on time-series remote sensing imagery, according to an embodiment of the present invention.

[0022] Figure 2 This is a semantic consistency score recognition change type determination diagram according to an embodiment of the present invention.

[0023] Figure 3 This is a diagram illustrating the sensitivity parameters of planned land parcel changes according to an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of the structure of a planned land parcel change monitoring system based on time-series remote sensing imagery, according to an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Reference Figure 1 One embodiment of the present invention proposes a method for monitoring changes in planned land parcels based on time-series remote sensing images. By using time-series remote sensing images and planned land parcel data, combined with change sensitivity pre-assessment and dynamic analysis strategies, the method can achieve automated and precise change monitoring of planned land parcels.

[0027] The method described in this embodiment specifically includes: S1. Acquire multi-temporal remote sensing images and planned land parcel data; Specifically, the geographical scope and monitoring time period of the monitoring area are first determined. Then, multi-temporal remote sensing images covering the monitoring area are acquired through channels such as remote sensing data receiving stations, commercial satellite data service providers, or public data platforms. Multi-temporal remote sensing images refer to sequences of remote sensing images acquired at different time points, such as monthly, quarterly, or annually. These images should have sufficient spatial resolution to identify changes in land features, and the time span should cover the effective monitoring period of the planned land parcels. The acquired images should contain multiple spectral band information for subsequent feature extraction and change analysis. Simultaneously, planning land parcel data reflecting land use, land nature, and spatial boundaries of the monitoring area needs to be obtained from planning management departments or relevant geographic information system databases. Planning land parcel data is usually stored in GIS vector data format, containing precise spatial geometric information for each planning land parcel, including boundaries, location, and attribute information describing its planned use and functional positioning. The acquired multi-temporal remote sensing images and planning land parcel data are imported into a data processing platform for preliminary organization and management, ensuring spatial and temporal coverage consistency between the two data sources, providing basic input for subsequent preprocessing and analysis. This step, by collecting surface observation information and accurate planning management information at different time points, lays a complete and accurate data foundation for subsequent change monitoring based on planning elements, and achieves a comprehensive capture of the current status and planning requirements of the monitored area.

[0028] S2. Preprocess the multi-temporal remote sensing images to generate preprocessed time-series remote sensing data; Optionally, the preprocessing of the multi-temporal remote sensing images to generate preprocessed time-series remote sensing data includes: Radiometric calibration and atmospheric correction were performed on the multi-temporal remote sensing images to obtain surface reflectance data; Geometric correction and image registration are performed on the surface reflectance data to generate preprocessed time-series remote sensing data.

[0029] Specifically, the multi-temporal remote sensing images are first subjected to radiometric calibration and atmospheric correction to obtain surface reflectance data. Radiometric calibration involves converting the original pixel grayscale values ​​of the remote sensing images... Converted to radiance values ​​received by the sensor The transformation is usually based on a linear model, and the relationship can be expressed as: , in It is the radiance value; It is the original pixel grayscale value, obtained through the output of the remote sensing sensor; It is the sensor's gain parameter; These are the sensor's bias parameters, typically provided by the sensor manufacturer or calibration report. Subsequently, atmospheric correction is performed to eliminate the influence of atmospheric water vapor, aerosols, and other pollutants on the reflected signals from ground objects, thus adjusting the radiance value. Converted to surface reflectance These two steps eliminate the influence of different sensors, times, and atmospheric conditions on image values, resulting in surface reflectance data with unified physical meaning. Next, geometrical correction and image registration are performed on the surface reflectance data to generate preprocessed time-series remote sensing data. Geometrical correction uses ground control points and a high-precision digital elevation model to correct the geometric distortion of the image itself, projecting the image onto a standard geographic coordinate system. Image registration ensures that the pixel positions of the same feature in all multi-temporal remote sensing images correspond accurately at different time points. By aligning the corrected surface reflectance data, spatially accurate and temporally consistent preprocessed time-series remote sensing data is finally generated.

[0030] For example, two time phases were obtained for a certain planning area. and Remote sensing image preprocessing generates preprocessed time-series remote sensing data. Radiometric calibration and atmospheric correction example: Obtaining the gain of a specific band from the calibration report of a satellite remote sensing sensor. bias Select The original grayscale value of a pixel of a certain feature in the image. The value is 100. Substitute this value into the formula to calculate its radiance value. for = Then, an atmospheric correction model is applied to this radiance value to eliminate atmospheric effects, and it is then converted into surface reflectance data, for example... Geometric fine correction and image registration verification example: using the corrected image... Using surface reflectance data from the imagery as a baseline, multiple precise ground control points were used to... Image registration is performed using surface reflectance data from the images. To verify the registration effect, a stable ground feature point is selected, such as the northwest corner of a building, and its position within the image is checked. The coordinates on the image are (500, 800), after registration The coordinates of this feature point on the image should also be (500, 800), and the pixel position deviation should be controlled within one pixel, so as to finally obtain spatially aligned and physically meaningful preprocessed time-series remote sensing data.

[0031] S3. Perform structured processing on the planned land parcel data to generate a structured planning knowledge base; Optionally, the step of structuring the planned land parcel data to generate a structured planning knowledge base includes: Spatial geometric information and attribute information are parsed from the planned land parcel data to obtain a set of planning attribute tags; Semantic classification and relational modeling are performed on the planning attribute tag set to generate a structured planning knowledge base.

[0032] Specifically, the first step involves analyzing the spatial geometric information and attribute information of the planned land parcel data to obtain a set of planning attribute tags. Spatial geometric information analysis involves reading the planned land parcel data in vector data format from a geographic information system (GIS) to extract the precise boundary coordinates, centroid location, and area of ​​each planned land parcel. Attribute information analysis involves reading the associated attribute tables in the planned land parcel data to extract the planned use code and control indicators, such as plot ratio, building height, and effective date—non-spatial descriptive information. These spatial and attribute information are combined to form a complete set of planning attribute tags describing the planning requirements of each land parcel. Secondly, the planning attribute tag set is semantically classified and its relationships modeled to generate a structured planning knowledge base. Semantic classification standardizes and categorizes the planned use codes and control indicators in the planning attribute tag set, converting heterogeneous planning descriptions into unified planning semantics; for example, various specific residential land codes are uniformly categorized as "residential construction." Relationship modeling defines and establishes relationships between land parcels, between planning semantics, and between land parcels and policy texts based on the spatial relationships between land parcels and the logical relationships in planning requirements. For example, spatial adjacency relationships between adjacent plots can be established, and these classifications and relationships can be stored in a knowledge graph or database that can be efficiently queried and reasoned by computers, ultimately generating a structured planning knowledge base.

[0033] For example, for a numbered The planned land parcel data is analyzed. The planned land parcel data shows: Spatial geometric information: land parcel Area Square meters, coordinate boundaries defined. Attribute information: Planning code is... , representing park green space, effective date is By analyzing spatial geometric information and attribute information, a set of planning attribute labels is obtained. Semantic classification and association modeling: Semantic classification: This involves classifying the planning code... Mapped or categorized into a unified planning semantic: "ecological protection green space". Relationship modeling: identifying land parcels. Another plot of land adjacent to it The spatial relationship for the planned "residential construction" site is adjacent. Ultimately, this is designated as a plot of land in the structured planning knowledge base. Knowledge entries have been created: This knowledge base entry not only clarifies the land parcel The planning requirement to maintain the green space status also foreshadows the risk of encroachment from adjacent construction land through correlation. This structured knowledge facilitates the system to accurately calculate policy sensitivity parameters and neighborhood correlation parameters in subsequent change sensitivity pre-assessment, and provides a standard temporal change trajectory of "ecological protection green space" for semantic-temporal coupling analysis.

[0034] S4. Based on the preprocessed time-series remote sensing data and the structured planning knowledge base, perform a change sensitivity pre-assessment for each planning plot and generate change sensitivity parameters for each planning plot; Optionally, the step of performing a change sensitivity pre-assessment for each planned land parcel and generating change sensitivity parameters for each planned land parcel includes: Historical fluctuation characteristics of each planned land parcel are extracted from the preprocessed time-series remote sensing data to generate historical stability parameters. The policy texts associated with the planned land parcels are extracted from the structured planning knowledge base, and the regulatory intensity of the policy texts is analyzed to generate policy sensitivity parameters. Obtain verified change information of adjacent planned land parcels, calculate the spatial correlation between the current planned land parcel and the adjacent planned land parcels, and generate neighborhood correlation parameters; The change sensitivity parameter is calculated by integrating the historical stability parameter, the policy sensitivity parameter, and the neighborhood correlation parameter.

[0035] Specifically, the process begins by extracting image data of the designated planning plot from preprocessed time-series remote sensing data. Time series data of spectral indicators such as the normalized difference vegetation index (NDVI) or building index of pixels within the plot are calculated, followed by the coefficient of variation or standard deviation. This quantifies the fluctuations in the plot's surface physical state under uninterrupted or natural evolution, generating a historical stability parameter reflecting the plot's inherent stability. A higher historical stability parameter value indicates significant fluctuations in surface characteristics and lower stability over past periods. Next, a structured planning knowledge base is accessed to extract spatially associated planning policy texts related to the current planning plot. Natural language processing (NLP) techniques are used to semantically analyze these policy texts, identifying and quantifying the intensity of regulatory directives. For example, a regulatory intensity dictionary is constructed, containing keywords such as "key development," "permitted construction," "restricted renovation," and "prohibited occupation," with each word assigned a corresponding numerical weight. By statistically analyzing the regulatory terms and their weights appearing in the policy texts associated with the current plot, a weighted sum is obtained to obtain a policy sensitivity parameter. A higher parameter indicates greater pressure for change under planning guidance or constraints on the plot. Simultaneously, verified change information of adjacent planned plots surrounding the current planned plot, which has been manually checked or model-verified in the previous monitoring period, is obtained. Based on the spatial proximity analysis function of the Geographic Information System (GIS), the spatial correlation between the current plot and these changed adjacent plots is calculated, for example, using the inverse distance weighting method, where the closer the changed plot is, the greater its influence on the current plot. The influence of all adjacent changed plots is summed to generate a neighborhood correlation parameter, which characterizes the spatial spillover effect or driving force of surrounding construction activities on the current plot. Finally, to comprehensively consider the impact of the above three dimensions, the historical stability parameter, policy sensitivity parameter, and neighborhood correlation parameter need to be fused to calculate the final change sensitivity parameter. Since these three parameters have different dimensions and value ranges, they are first normalized to map them into a unified numerical range. Then, they are fused using a weighted summation method to obtain the change sensitivity parameter, the calculation formula of which is: , in It is a historical stability parameter. It is a policy sensitivity parameter. These are neighborhood correlation parameters; they are all dimensionless numerical values, and In this model, the inverse value should be taken to reflect historical instability, or it should be given a negative weight. These are preset weighting coefficients, whose values ​​are derived from the evaluation model to ensure... This multi-dimensional parameter fusion method can comprehensively assess the risk level of potential changes in each planned plot, providing a scientific basis for subsequent monitoring resource allocation.

[0036] For example, for a numbered A pre-assessment of the change sensitivity of planned land parcels is conducted. Land parcels are extracted from preprocessed time-series remote sensing data. The historical fluctuation characteristics are calculated and standardized to obtain the historical stability parameters. Parcel analysis from structured planning knowledge base The planning policy text has a moderately high level of regulation, generating policy sensitivity parameters. Acquire land parcels Verified changes to five adjacent planned land parcels were analyzed, with two of these parcels identified as having undergone illegal changes through manual verification. Spatial correlation was calculated and standardized to generate neighborhood correlation parameters. Sensitivity parameters Calculation: The weights of the evaluation model are configured as follows The three parameters are then fused and calculated: The final plot of land The sensitivity parameter for the change is 0.48. Based on this value, the system will classify it into the corresponding sensitivity level and assign the corresponding analysis strategy parameters.

[0037] S5. Based on the aforementioned change sensitivity parameters, dynamically allocate analysis strategy parameters to each planned plot; Optionally, the dynamic allocation of analysis strategy parameters for each planned land parcel includes: Based on the values ​​of the aforementioned sensitivity parameters, each planned plot of land is classified into different sensitivity levels; Obtain the analysis time window parameters and analysis intensity level parameters corresponding to each of the sensitivity levels; Based on the sensitivity level of each planned land parcel, the corresponding analysis time window parameters and analysis intensity level parameters are combined to generate analysis strategy parameters.

[0038] Specifically, firstly, each planned land parcel is classified into different sensitivity levels based on the value of the change sensitivity parameter. The change sensitivity parameter, obtained through pre-assessment, measures the likelihood and importance of changes occurring to the planned land parcel. Next, the analysis time window parameter and analysis intensity level parameter corresponding to each sensitivity level are obtained. The analysis time window parameter defines the time period to be monitored, such as the latest period, the two most recent periods, or the entire historical time series; the analysis intensity level parameter defines the granularity of change detection, such as performing only a rapid screening of physical changes or requiring semantic-temporal coupled analysis. Finally, based on the sensitivity level of each planned land parcel, the corresponding analysis time window parameter and analysis intensity level parameter are combined to generate analysis strategy parameters. This dynamic allocation strategy allows limited analytical resources to be concentrated on land parcels with high change sensitivity, improving monitoring efficiency and accuracy.

[0039] For example, the sensitivity parameter is a value between 0 and 100, where 0 represents the lowest sensitivity and 100 represents the highest sensitivity. Sensitivity levels can be divided into three levels: low sensitivity level, parameter value... This corresponds to a combination of the analysis time window parameter being "the latest two periods" and the analysis intensity level parameter being "rapid screening." Medium sensitivity level, parameter value... This corresponds to a combination of the analysis time window parameter "latest four periods" and the analysis intensity level parameter "rapid screening + local depth analysis". High sensitivity level, parameter value... The corresponding analysis time window parameter is a combination of "all historical time series" and the analysis intensity level parameter is a combination of "rapid screening + comprehensive in-depth analysis". Now, let's take a planned plot of land as an example. Reproduction: Plot After a preliminary assessment of change sensitivity, the calculated change sensitivity parameter value was 85. Based on this value range, the land parcel... The site was classified as high-sensitivity. Referring to the preset strategy table, the analysis time window parameter for high-sensitivity was "All Historical Time Series," and the analysis intensity level parameter was "Fast Screening + Comprehensive In-Depth Analysis." These two parameters were combined to generate the land parcel. The analysis strategy parameters. When subsequent stratified change detection is performed, for the plot... The system will analyze its entire historical time-series remote sensing data and perform semantic-temporal coupling analysis based on the fast screening, thereby detecting its changes in a more detailed and comprehensive manner.

[0040] S6. Based on the analysis strategy parameters, perform stratified change detection on the corresponding planned plots and generate plot change monitoring results.

[0041] Optionally, the step of performing stratified change detection on the corresponding planned land parcels includes: Feature comparison is performed on the latest two phases of remote sensing images of all planned plots to conduct a rapid screening of physical changes and generate preliminary changed areas; The analysis strategy parameters with deep analysis instructions are identified, and semantic-temporal coupling analysis is performed on the planning plots holding the instructions to obtain semantic analysis conclusions. Using the semantic analysis results, the preliminary change areas are verified and filtered to generate verified change patches.

[0042] Specifically, the system first compares the features of the two most recent remote sensing images of all planned land parcels to perform a rapid physical change screening, generating preliminary change areas. This rapid physical change screening identifies areas of land cover change by comparing the differences in spectral, textural, and other physical features between the two images. For example, image interpolation or change vector analysis can be used to calculate the change magnitude of each pixel. When the change magnitude of a certain area within a parcel exceeds a preset threshold, it is determined to be a preliminary change area. Next, the system identifies analysis strategy parameters with depth analysis instructions and performs semantic-temporal coupling analysis on the planned land parcels holding these instructions, obtaining semantic analysis conclusions. Depth analysis instructions are dynamically assigned based on change sensitivity parameters, indicating that the parcel requires more refined and accurate change detection. The semantic-temporal coupling analysis combines the planning semantic information of the planned land parcels with their historical temporal remote sensing image data to analyze the rationality, trend, and type of land parcel changes, generating semantic analysis conclusions on whether the changes conform to the planning. Finally, using the semantic analysis conclusions, the preliminary change areas are verified and filtered, generating verified change patches.

[0043] For example, during a certain monitoring period, the system acquires the latest two periods of remote sensing images for all planned land parcels. Physical change rapid screening: Feature comparison is performed on the latest two periods of images for all land parcels to calculate the magnitude of change. Land parcel If the change in a certain area exceeds a threshold, it is quickly identified as a preliminary area of ​​change. Command recognition and deep analysis: The system checks land parcels. The analysis strategy parameters revealed that it possesses deep analysis commands. The system analyzes the land parcels. Semantic-temporal coupling analysis was performed. The planned use of this plot of land is "arable land." The semantic-temporal coupling analysis revealed that the actual temporal change trajectory of the initial change area showed a change from vegetation cover to bare soil, which is consistent with the standard temporal change trajectory of "arable land" during the crop harvest season. Therefore, the semantic analysis conclusion was "normal seasonal change." Verification and filtering: The system used the semantic analysis conclusion of "normal seasonal change" to filter the initial change area. This area was determined to be a pseudo-change and will not be retained as a final verified change patch. In contrast, the plot... A preliminary change area was also detected, with its semantic-temporal coupling analysis indicating a change from woodland to buildings. This is significantly inconsistent with the planning semantics of the site as an "ecological protection zone," leading to a semantic analysis conclusion of "suspected illegal construction." This conclusion will be used to verify and preserve this preliminary change area, ultimately generating a verified change patch. This hierarchical detection method achieves rapid identification of potential changes over a wide area through physical change screening, and then performs high-precision verification and filtering through targeted semantic-temporal coupling analysis, effectively identifying real changes, especially those with planning significance, thereby improving the efficiency and reliability of change detection.

[0044] Optionally, the execution of semantic-temporal coupling analysis includes: Obtain the standard temporal change trajectory corresponding to the planning semantics from the structured planning knowledge base; Extract the actual temporal change trajectory of the current planned land parcel from the preprocessed temporal remote sensing data; The actual temporal change trajectory is matched with the standard temporal change trajectory to generate a semantic consistency score. Based on the semantic consistency score, the change type is identified, and semantic analysis conclusions are obtained.

[0045] Specifically, the first step is to obtain standard temporal change trajectories corresponding to the planning semantics from the structured planning knowledge base. Planning semantics refers to the specific functions and requirements assigned to a planned land parcel, such as "farmland protection," "ecological green space," or "residential construction." Standard temporal change trajectories are pre-constructed based on historical experience and planning requirements, describing typical, compliant, or expected patterns of land cover changes over time under a given planning semantic. Next, the actual temporal change trajectory of the current planned land parcel is extracted from the preprocessed temporal remote sensing data. The actual temporal change trajectory is the sequence of land cover change characteristics for that land parcel from the start of monitoring to the latest period, which can be analyzed by examining the normalized vegetation index of pixels within the land parcel. The semantic consistency score is obtained by analyzing changes in factors such as building indices over time. Then, the actual time-series change trajectory is compared with the standard time-series change trajectory to generate a semantic consistency score. Similarity matching can be quantified using time series analysis methods, such as dynamic time warping algorithms or Pearson correlation coefficients. The calculation can be expressed as: , in, Represents a similarity matching function. It represents the actual temporal change trajectory, and its value originates from the parcel feature sequence extracted from the preprocessed temporal remote sensing data. This is a standard time-series change trajectory, whose values ​​are derived from a structured planning knowledge base. The semantic consistency score measures the degree of similarity between the actual change pattern and the compliant change pattern. Finally, based on the semantic consistency score, change types are identified, and semantic analysis conclusions are obtained. The semantic consistency score determines the change type as follows: Figure 2 As shown, by comparing the semantic consistency score with a preset threshold, it is possible to determine whether the current land parcel's change is a compliant change, a low-risk change, or a suspected non-compliant change, thereby generating the final semantic analysis conclusion.

[0046] For example, for a planned plot of land Performing semantic-temporal coupling analysis, the planning semantics of this plot of land is "ecological protection green space". Obtaining the standard trajectory: The standard temporal variation trajectory corresponding to "ecological protection green space" is obtained from the structured planning knowledge base. This trajectory is represented by the normalized difference in vegetation index. average It has remained above 0.6 for a long time with little fluctuation. During four consecutive monitoring periods Value Extracting actual trajectories: Extracting land parcels from preprocessed time-series remote sensing data. The actual time-series change trajectory during the four monitoring periods in the same period ,by Sequence representation. Similarity matching: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] and Perform similarity matching and calculate semantic consistency score. Pearson correlation coefficient was used as... A function is used to calculate similarity. Case 1: Compliance: If the land parcels... It maintained a high vegetation cover .calculate and The Pearson correlation coefficient is used to obtain the semantic consistency score. Case 2: Suspected violation: If the land parcel Vegetation destruction and construction occurred, .calculate and The Pearson correlation coefficient is used to obtain the semantic consistency score. Identify change type: The preset compliance threshold is 0.70. For Case 1... Because the score was far above the threshold, the change type was identified as "ecological stability," and the semantic analysis conclusion was "normal change, in line with planning." For Case Two... Since the score was far below the threshold, the negative correlation indicated that the change trend was opposite. The change type was identified as "sharp reduction of vegetation", and the semantic analysis conclusion was "suspected illegal destruction of green space".

[0047] Optionally, after generating the land parcel change monitoring results, the method further includes: The monitoring results of the land parcel changes are converted into vector data supported by the geographic information system to generate a visual monitoring layer; The planning control line and basic geographic features are obtained, and the visualization monitoring layer is overlaid with the planning control line and the basic geographic features to generate a comprehensive monitoring map; Based on the comprehensive monitoring map, spatial query and statistical analysis functions are provided, and a decision support interface is generated.

[0048] Specifically, the land parcel change monitoring results are first converted into vector data supported by a Geographic Information System (GIS) to generate a visual monitoring layer. The land parcel change monitoring results are the output of the layered change detection step, containing the geometric location and change attribute information of the changed land features. This step involves extracting the geometric information of these changed areas and attaching their change attributes, such as change type and change area, to the corresponding geometric figures, forming vector format data that meets GIS requirements. This vector data is the visual monitoring layer. Next, planning control lines and basic geographic features are acquired, and the visual monitoring layer is overlaid with the planning control lines and basic geographic features to generate a comprehensive monitoring map. Planning control lines refer to legally binding spatial boundaries, such as ecological protection red lines and permanent basic farmland control lines. Basic geographic features include background geographic information such as roads, water systems, and administrative divisions. Through spatial overlay analysis technology, the visual monitoring layer is precisely placed on the geographic background of the planning control lines and basic geographic features, generating a comprehensive monitoring map that integrates change information, planning requirements, and the geographic environment. Finally, spatial query and statistical analysis functions are provided based on the comprehensive monitoring map, generating a decision support interface. The spatial query function allows users to immediately obtain detailed monitoring and planning attributes by selecting areas or patches on the map. The statistical analysis function can quickly summarize and calculate the number, total area, and conflict area with planning control lines of specific areas or types of changed patches. By integrating these functions into an interactive visualization interface, a decision support interface is generated. This step transforms abstract monitoring results into intuitive and operable geographic information products, greatly facilitating planning managers to quickly locate changes, trace their attributes, and quantitatively assess them, thereby improving the efficiency and accuracy of management decisions.

[0049] For example, the land change monitoring results identified a 500-square-meter area with a change type of "suspected illegal land occupation and construction". Convert to vector data: convert the region Extract the boundary coordinates into a polygon vector object and then extract its attribute information, including the area. Square meters, change type added, generating a visual monitoring layer. Overlay to generate a map: The system obtains the "river blue line" from the planning control line and road and building layers from the basic geographic features. Through spatial overlay analysis, change areas are identified. The 300 square meter area overlaps spatially with the "river blue line". The system will... Displayed alongside river blue lines and roads, generating a comprehensive monitoring map. Function query interface: On the decision support interface, administrators click on the changed area. The system immediately displayed its detailed attributes through the spatial query function: the total changed area is 500 square meters, and the changed area within the river channel blue line is 300 square meters, suspected of being illegal land occupation and construction. At the same time, the statistical analysis function automatically summarized the total area of ​​all suspected illegal constructions within the "river channel blue line" during this period, which is 300 square meters.

[0050] Optionally, the method further includes: Obtain the results of manual verification and generate a verification sample set; The monitoring results of the changes in the land parcels are compared with the verification sample set to identify misjudgment cases and generate a misjudgment analysis report; Based on the misjudgment analysis report, the weights of the evaluation model used to generate the change sensitivity parameter are adjusted to generate an updated evaluation model.

[0051] Specifically, the first step involves obtaining manual verification results and generating a verification sample set. This step involves manually verifying the changed land parcels identified by the system after the system automatically generates the monitoring results. This is done through methods such as on-site inspections, high-resolution UAV imagery, or feedback from authoritative departments, to determine whether the changes are genuine and accurate in their nature. These manually verified change information are then formatted, such as "genuine violation change," "genuine compliance change," or "false change / false alarm," to create a verification sample set containing fields such as spatial location, time, and the true value of the change. Next, the land parcel change monitoring results are compared with the verification sample set to identify misjudgment cases and generate a misjudgment analysis report. By comparing the land parcel change monitoring results output by the system with the true values ​​in the verification sample set, two main types of misjudgments can be identified: one is missed reporting, i.e., genuine changes that the system did not detect; the other is false alarms, i.e., changes detected by the system but manually verified as false changes or incorrect change types. Based on these misjudgment cases, their characteristics, frequency of occurrence, and possible causes are statistically analyzed, and a misjudgment analysis report is compiled. Finally, based on the misjudgment analysis report, the weights of the evaluation model used to generate the change sensitivity parameters are adjusted to generate an updated evaluation model. The change sensitivity parameters are calculated by fusing historical stability parameters, policy sensitivity parameters, and neighborhood correlation parameters; the fusing process depends on the weight configuration of each parameter. Based on the systematic biases revealed in the misjudgment analysis report, the weights of these input parameters are adjusted. This can be done to correct the predictive power of the evaluation model. For example, if a large number of underreporting events are found in highly policy-sensitive areas, the weight of the policy sensitivity parameter can be appropriately increased. New evaluation model weights The adjustment logic is as follows: , in These are the weights before adjustment. The adjustment amount is based on misjudgment analysis to ensure... By iteratively adjusting the weights of the evaluation model, an updated evaluation model can be generated, thereby improving the accuracy of the prediction of the sensitivity parameters for the next cycle.

[0052] For example, the initial evaluation model weights used to generate the change sensitivity parameters are configured as follows: historical stability parameter weights Policy sensitivity parameter weights Neighborhood correlation parameter weights Initial total weight The parameters for sensitivity to changes in planned land parcels are as follows: Figure 3 As shown. Obtaining manual verification results: In the land change monitoring results output by the system, the land parcels... It was identified as a "compliance change under low-sensitivity circumstances." Upon manual verification, the land parcel was found to have... In reality, it was "illegal construction within a highly policy-sensitive area." This verification result was recorded in the verification sample set. Comparison and Reporting: The system will compare and report the land parcels. The system results were compared with the verified true values, identifying it as a case of missed reporting. The misjudgment analysis report indicated that although the land parcel... The policy sensitivity parameter is high, but due to the weight of the historical stability parameter... An excessively high sensitivity level leads to an underestimation of the change sensitivity parameter in the final calculation. The system incorrectly assigns it to a low sensitivity level, resulting in weak change detection and ultimately, missed detections. To address this, the system adjusts the assessment model weights: based on the misjudgment analysis report, it decides to reduce the influence of the historical stability parameter in the assessment model while increasing the influence of the policy sensitivity parameter. This reduces... ,improve Keep The adjusted and updated evaluation model weights are as follows: , , Adjusted total weight .

[0053] Based on the same inventive concept, such as Figure 4 As shown, the present invention also provides a planning plot change monitoring system based on time-series remote sensing imagery, the system comprising: The data acquisition module is used to acquire multi-temporal remote sensing images and planned land parcel data; The image preprocessing module is used to preprocess the multi-temporal remote sensing images to generate preprocessed time-series remote sensing data. The planning knowledge base construction module is used to perform structured processing on the planning plot data to generate a structured planning knowledge base; The change sensitivity pre-assessment module is used to perform a change sensitivity pre-assessment on each planning plot based on the preprocessed time-series remote sensing data and the structured planning knowledge base, and generate change sensitivity parameters for each planning plot. The parameter dynamic allocation module is used to dynamically allocate analysis strategy parameters to each planned plot based on the change sensitivity parameters. The monitoring result output module is used to perform stratified change detection on the corresponding planned plots based on the analysis strategy parameters, and generate plot change monitoring results.

[0054] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0055] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A method for monitoring changes in planned land parcels based on time-series remote sensing imagery, characterized in that, The method includes: Acquire multi-temporal remote sensing images and planned land parcel data; The multi-temporal remote sensing images are preprocessed to generate preprocessed time-series remote sensing data. The planned land parcel data is processed in a structured manner to generate a structured planning knowledge base; Based on the preprocessed time-series remote sensing data and the structured planning knowledge base, a change sensitivity pre-assessment is performed on each planning plot, generating change sensitivity parameters for each planning plot; Based on the aforementioned change sensitivity parameters, analysis strategy parameters are dynamically assigned to each planned land parcel; Based on the analysis strategy parameters, stratified change detection is performed on the corresponding planned land parcels to generate land parcel change monitoring results.

2. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 1, characterized in that, The process of conducting a change sensitivity pre-assessment for each planned land parcel and generating change sensitivity parameters for each planned land parcel includes: Historical fluctuation characteristics of each planned land parcel are extracted from the preprocessed time-series remote sensing data to generate historical stability parameters. The policy texts associated with the planned land parcels are extracted from the structured planning knowledge base, and the regulatory intensity of the policy texts is analyzed to generate policy sensitivity parameters. Obtain verified change information of adjacent planned land parcels, calculate the spatial correlation between the current planned land parcel and the adjacent planned land parcels, and generate neighborhood correlation parameters; The change sensitivity parameter is calculated by integrating the historical stability parameter, the policy sensitivity parameter, and the neighborhood correlation parameter.

3. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 1, characterized in that, The dynamic allocation analysis strategy parameters for each planned land parcel include: Based on the values ​​of the aforementioned sensitivity parameters, each planned plot of land is classified into different sensitivity levels; Obtain the analysis time window parameters and analysis intensity level parameters corresponding to each of the sensitivity levels; Based on the sensitivity level of each planned land parcel, the corresponding analysis time window parameters and analysis intensity level parameters are combined to generate analysis strategy parameters.

4. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 1, characterized in that, The step of performing stratified change detection on the corresponding planned land parcels includes: Feature comparison is performed on the latest two phases of remote sensing images of all planned plots to conduct a rapid screening of physical changes and generate preliminary changed areas; The analysis strategy parameters with deep analysis instructions are identified, and semantic-temporal coupling analysis is performed on the planning plots holding the instructions to obtain semantic analysis conclusions. Using the semantic analysis results, the preliminary change areas are verified and filtered to generate verified change patches.

5. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 4, characterized in that, The execution semantic-temporal coupling analysis includes: Obtain the standard temporal change trajectory corresponding to the planning semantics from the structured planning knowledge base; Extract the actual temporal change trajectory of the current planned land parcel from the preprocessed temporal remote sensing data; The actual temporal change trajectory is matched with the standard temporal change trajectory to generate a semantic consistency score. Based on the semantic consistency score, the change type is identified, and semantic analysis conclusions are obtained.

6. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 1, characterized in that, The method further includes: Obtain the results of manual verification and generate a verification sample set; The monitoring results of the changes in the land parcels are compared with the verification sample set to identify misjudgment cases and generate a misjudgment analysis report; Based on the misjudgment analysis report, the weights of the evaluation model used to generate the change sensitivity parameter are adjusted to generate an updated evaluation model.

7. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 1, characterized in that, The preprocessing of the multi-temporal remote sensing images to generate preprocessed time-series remote sensing data includes: Radiometric calibration and atmospheric correction were performed on the multi-temporal remote sensing images to obtain surface reflectance data; Geometric correction and image registration are performed on the surface reflectance data to generate preprocessed time-series remote sensing data.

8. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 1, characterized in that, The step of structuring the planned land parcel data to generate a structured planning knowledge base includes: Spatial geometric information and attribute information are parsed from the planned land parcel data to obtain a set of planning attribute tags; Semantic classification and relational modeling are performed on the planning attribute tag set to generate a structured planning knowledge base.

9. The method for monitoring changes in planned land parcels based on time-series remote sensing imagery according to claim 1, characterized in that, The generation of land parcel change monitoring results also includes: The monitoring results of the land parcel changes are converted into vector data supported by the geographic information system to generate a visual monitoring layer; The planning control line and basic geographic features are obtained, and the visualization monitoring layer is overlaid with the planning control line and the basic geographic features to generate a comprehensive monitoring map; Based on the comprehensive monitoring map, spatial query and statistical analysis functions are provided, and a decision support interface is generated.

10. A planning land parcel change monitoring system based on time-series remote sensing imagery, applied to the planning land parcel change monitoring method based on time-series remote sensing imagery as described in any one of claims 1-9, characterized in that, The system includes: The data acquisition module is used to acquire multi-temporal remote sensing images and planned land parcel data; The image preprocessing module is used to preprocess the multi-temporal remote sensing images to generate preprocessed time-series remote sensing data. The planning knowledge base construction module is used to perform structured processing on the planning plot data to generate a structured planning knowledge base; The change sensitivity pre-assessment module is used to perform a change sensitivity pre-assessment on each planning plot based on the preprocessed time-series remote sensing data and the structured planning knowledge base, and generate change sensitivity parameters for each planning plot. The parameter dynamic allocation module is used to dynamically allocate analysis strategy parameters to each planned plot based on the change sensitivity parameters. The monitoring result output module is used to perform stratified change detection on the corresponding planned plots based on the analysis strategy parameters, and generate plot change monitoring results.

Images