Desensitization method for geographically sensitive data

Abstract

The invention discloses a desensitization method for geographic sensitive data. The method comprises the following steps: acquiring an original data set and a priori geographic information graph; according to the priori geographic information map, judging the land utilization type and the geological background unit type of each data acquisition point, and determining the maximum offset based on the corresponding geographic sensitive data; generating a random offset in the maximum offset and in the geological background unit corresponding to the data acquisition point, so as to obtain a desensitization coordinate; applying controlled random disturbance to the geographical sensitive data of each data acquisition point, performing logarithmic transformation, and performing contrast value transformation in combination with a regional background value to obtain contrast value data; and generating an initial grid by combining the desensitization coordinates and the contrast data interpolation, outputting a final grid after multiple times of smooth filtering iteration, and performing hierarchical visual display based on the final grid.

Description

Technical Field

[0001] This invention relates to the field of geographic information technology, and in particular to a method for desensitizing geographically sensitive data. Background Technology

[0002] With the continuous development of environmental survey and monitoring technologies and the increasing demand for open sharing of government data, the secure and efficient release of geospatial sensitive data has become a key issue in the interdisciplinary field of environmental science, geographic information systems, and data security. This type of data is directly linked to core sensitive information such as the location of sensitive facilities and the precise extent of abnormally polluted areas. Directly releasing the raw geographic coordinates and survey and monitoring values ​​could not only lead to the leakage of sensitive information but also easily cause unnecessary public panic, thereby bringing potential security risks and social governance risks.

[0003] Existing technologies for processing geospatial sensitive data have significant limitations, mainly falling into three categories: First, overall obfuscation processing, which involves taking overall values ​​or generally obfuscating the survey and monitoring data within the target area. While this method can reduce the risk of information leakage to some extent, it severely damages the spatial resolution and gradient change characteristics of the data, resulting in the processed data failing to objectively reflect the true regional distribution trend and significantly diminishing its scientific value in scenarios such as environmental analysis and decision support. Second, simple coordinate offset method, which achieves location desensitization by applying a fixed distance or random offset to all data collection points. However, this method is relatively easy to crack, and the offset process lacks targeted constraints, easily disrupting the inherent spatial structure of the data and its intrinsic correlation with geographical elements such as terrain and land use types, seriously affecting the reliability of subsequent data applications. Third, direct data hierarchical display, which involves simply classifying data into levels before releasing it externally without effectively desensitizing the original geographic coordinates. This still cannot avoid the risk of location leakage of sensitive points, resulting in limited security protection. Therefore, there is a need for a desensitization method for geographic sensitive data that can integrate prior geographic knowledge during the desensitization process to effectively protect sensitive information while maximizing the preservation of the scientific value and application reliability of the data. Summary of the Invention

[0004] In view of the above-mentioned prior art, the present invention provides a method for desensitizing geographically sensitive data, which mainly solves the technical problems existing in the background art.

[0005] To achieve the above objectives, the technical solution of this invention is implemented as follows: This invention discloses a method for desensitizing geographically sensitive data, the method comprising the following steps: Obtain the original dataset and the prior geographic information map, which includes a land use type map and a geological background unit map. The original dataset includes the geographic coordinates of each data collection point and the corresponding geographic sensitive data. Based on the prior geographic information map, determine the land use type and geological background unit type of each data collection point, and determine the maximum offset based on its corresponding geographic sensitive data; Random offsets are generated within the geological background unit corresponding to the data acquisition point, based on the maximum offset, to obtain the desensitized coordinates. Controlled random perturbation is applied to the geographic sensitive data of each data collection point, and after logarithmic transformation, contrast transformation is performed in combination with regional background values ​​to obtain contrast data; An initial grid is generated by interpolating the desensitized coordinates and the chromaticity data. After multiple smoothing and filtering iterations, the final grid is output and a hierarchical visualization is performed based on the final grid.

[0006] Optionally, the geographically sensitive data includes environmental monitoring data and pollution distribution survey data. The environmental monitoring data includes at least the content of radionuclides and radiation dose rate, and the pollution distribution survey data includes at least the concentration of heavy metals in the soil, the content of toxic pollutants in the water, and the concentration of air pollutants. Each data collection point collects any type of geographically sensitive data.

[0007] Optionally, based on prior geographic information maps, determine the land use type and geological background unit type of each data collection point, specifically including: The geographic coordinates of each data collection point are spatially overlaid with the land use type map and geological background unit map in the prior geographic information map; Extract the land use type attribute corresponding to the geographic coordinate from the land use type map to determine the land use type of the data collection point; Extract the geological zoning attributes corresponding to the geographic coordinates from the geological background unit map to determine the geological background unit type to which the data collection point belongs.

[0008] Optionally, before obtaining the original dataset and prior geographic information map, a correspondence table between all land types and all basic maximum offsets is established to obtain a land use type-basic jitter amplitude comparison table.

[0009] Optionally, the maximum offset can be determined based on its corresponding geographic sensitive data, specifically including: Based on the land use type of the data collection point, query the land use type-basic jitter amplitude comparison table to determine the maximum basic offset corresponding to the land use type; Based on any type of geographic sensitive data collected from the data collection point, a scaling factor is set according to the magnitude of the numerical value of that type of geographic sensitive data. Multiply the base maximum offset by the scaling factor to obtain the maximum offset.

[0010] Optionally, a random offset is generated within the geological background unit corresponding to the data acquisition point, based on the maximum offset, to obtain the desensitized coordinates. Specifically, this includes: Using the geographic coordinates of the data collection point as the center, within a spatial range with the maximum offset as the radius, an initial offset is generated through a random function to obtain candidate de-identified coordinates; Determine whether the candidate desensitized coordinates fall within the geological background unit corresponding to the data acquisition point. If the candidate desensitized coordinates are within the geological background unit, then they are determined as the desensitized coordinates of the data acquisition point. If the candidate desensitized coordinates exceed the geological background unit, then the initial offset is discarded and a random offset is regenerated. Repeat the above judgment steps. If no candidate desensitized coordinates that meet the conditions are generated after reaching the preset maximum number of attempts, select the point within the boundary of the geological background unit that is closest to the geographic coordinates as the desensitized coordinates.

[0011] Optionally, the maximum magnitude of the controlled random disturbance is a preset percentage of the geographically sensitive data value, and the higher the data value, the smaller the preset percentage.

[0012] Optionally, after logarithmic transformation, contrast transformation is performed by combining the regional background values ​​to obtain contrast data, specifically including: The controlled random disturbance is superimposed on the geographic sensitive data to obtain intermediate data, and then the intermediate data is subjected to logarithmic transformation. Obtain the regional background value corresponding to the geographically sensitive data type, and calculate the logarithm of the regional background value; The contrast data is obtained by dividing the result of the logarithmic transformation of the intermediate data by the logarithm of the regional background value.

[0013] Optionally, an initial grid is generated by interpolating the desensitized coordinates and contrast data, and the final grid is output after multiple smoothing filtering iterations. Specifically, this includes: Associate the desensitized coordinates of each data collection point with the corresponding contrast data to construct a spatial interpolation dataset. A spatial interpolation algorithm is used to perform interpolation calculations on the interpolation dataset, and an initial grid file is generated according to a preset cell size; The initial grid file is iteratively smoothed using a filtering algorithm with a preset filtering window size. If the frequency distribution difference between the cumulative frequency distribution curves of the two consecutive tests is less than a preset threshold, the iterative smoothing ends, and the filtered grid data is determined as the final grid and output.

[0014] Optionally, hierarchical visualization is performed based on the final grid, specifically including: defining hierarchical thresholds and dividing hierarchical intervals based on the numerical distribution characteristics of the contrast data of each grid cell in the final grid; generating a first type of visualization layer and a second type of visualization layer based on the hierarchical intervals, and displaying the first type of visualization layer or the second type of visualization layer to the user according to different user permissions.

[0015] The beneficial effects of this invention are as follows: The adaptive coordinate desensitization mechanism under triple constraints adopted in this application effectively solves the technical defects of existing simple coordinate offsets that are easy to crack and cross-regional distortion. It dynamically matches the basic jitter amplitude by land use type, adaptively adjusts the scaling coefficient according to the data sensitivity, and uses the boundary of geological background unit as a rigid constraint. While ensuring the safety of the original sensitive points, it maximizes the preservation of the inherent correlation between the data and the geographical environment, and significantly improves the spatial pattern fidelity of the desensitized data. In addition, the dual desensitization strategy of coordinate jitter and controlled data noise addition overcomes the problem of data resolution loss caused by the overall blurring in existing methods. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this utility model, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only preferred embodiments of this utility model. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for desensitizing geographically sensitive data in an embodiment of this application. Detailed Implementation

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. In the following description, the expression "some embodiments" refers to a subset of all possible embodiments; however, it should be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with each other without conflict.

[0019] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described in order to avoid obscuring the invention.

[0020] It should be understood that the present invention can be embodied in various forms and should not be construed as being limited to the embodiments set forth herein. Rather, providing these embodiments will make the disclosure thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Furthermore, the terminology used herein is intended only to describe particular embodiments and is not intended to limit the invention. When used herein, the singular forms “a,” “an,” and “the” are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “compose” and / or “comprising,” when used in this specification, identify the presence of the stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups. When used herein, the term “and / or” includes any and all combinations of the associated listed items.

[0021] It should also be noted that when an element is referred to as being "fixed to" another element, it can be directly attached to the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "inner," "outer," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only possible implementation.

[0022] To fully understand this invention, a detailed structure will be presented in the following description to illustrate the technical solution proposed by this invention. Optional embodiments of the invention are described in detail below; however, in addition to these detailed descriptions, the invention may have other embodiments.

[0023] Please refer to the attached document. Figure 1 This application provides a method for desensitizing geographically sensitive data, the method comprising the following steps: S1. Obtain the original dataset and the prior geographic information map, which includes a land use type map and a geological background unit map. The original dataset includes the geographic coordinates of each data collection point and the corresponding geographic sensitive data.

[0024] In this embodiment, the original dataset contains information from multiple data collection points covering key locations within the study area. Each data collection point must record precise geographic coordinates, typically in latitude and longitude format (e.g., WGS84 coordinate system) to ensure accurate mapping to geographic space. Simultaneously, each data collection point must correspond to a unique type of geographically sensitive data. This type of data consists of raw, unsensitized quantitative values, specifically including two main categories: environmental monitoring data and pollution distribution survey data. Environmental monitoring data must at least cover the concentration of radioactive nuclides (e.g., uranium, thorium, radium, etc., in Bq / kg units) and radiation dose rate (units of Bq / kg). The pollution distribution survey data should include at least the soil heavy metal concentration (e.g., lead, mercury, cadmium, etc., in mg / kg), the water toxic pollutant content (e.g., benzene series compounds, chemical oxygen demand, etc., in mg / L), and the air pollutant concentration (e.g., PM2.5, sulfur dioxide, etc., in μg / m³). Each data collection point should be associated with only one type of geographically sensitive data. For example, data collection point A should only collect radiation dose rate, and data collection point B should only collect soil heavy metal concentration. The specific type of data to be collected should be clearly determined according to the survey and monitoring objectives of the study area to avoid a single collection point being associated with multiple types of geographically sensitive data at the same time.

[0025] While acquiring the original dataset, it is necessary to simultaneously collect prior geographic information maps that highly match the spatial scope and research objectives of the dataset, ensuring that the spatial reference systems of the two types of data are completely identical. The prior geographic information maps specifically include land use type maps and geological background unit maps. The land use type maps must clearly delineate various land use attributes within the study area, covering at least core types such as urban and rural settlements, cultivated land, forest land, grassland, water bodies, and unused land. The scale of these maps should be reasonably selected based on the density of the original data collection points. For example, a 1:100,000 land use vector map of a certain region can be collected (including urban and rural areas, cultivated land, forest land, grassland, water bodies, and unused land types, ensuring accurate matching of the human settlement characteristics of each data collection point). Geological background unit maps need to be drawn based on regional geological survey results. Different geographical background units can be divided according to standards such as lithological characteristics and structural units. For example, a 1:500,000 geological lithological zoning vector map of a certain region can be collected, and each unit can be assigned a unique identification code, such as "G1" for lithological unit. After obtaining the above two types of layers, they need to be standardized and preprocessed, including unifying vector data formats, correcting boundary topological relationships, and normalizing attribute fields. Simultaneously, the original dataset needs to be cleaned to remove outliers and missing values ​​caused by collection errors, ensuring that all data and layers meet the requirements for subsequent land use type determination, geological background unit matching, and basic jitter amplitude calculation.

[0026] S2. Based on the prior geographic information map, determine the land use type and geological background unit type of each data collection point, and determine the maximum offset based on its corresponding geographic sensitive data. S3. Generate a random offset within the geological background unit corresponding to the data acquisition point, based on the maximum offset, to obtain the desensitized coordinates. The core of steps S2 and S3 is to each original coordinate point ( Generate a new, desensitized coordinate system. This process is controlled by the following three constraints to ensure that the desensitized data points maintain scientific consistency with the original environmental background even after the location shift.

[0027] The first constraint is the intensity constraint of human activities, that is, based on the original coordinate points ( The basic jitter amplitude is determined by the land use type (e.g., urban planning area, rural settlement, farmland, forest land, etc.) where the land is located. The principle is: the denser the human settlement, the smaller the amplitude of the foundation shaking, so as to protect the precise location of infrastructure and sensitive areas; conversely, in areas such as open fields, the amplitude of the foundation shaking can be increased.

[0028] In practice, the geographic coordinates of each data collection point are used as the spatial query benchmark, and point-by-point spatial overlay matching is performed with the land use type map and geological background unit map in the prior geographic information map. During the overlay with the land use type map, the system extracts the land use type attribute information corresponding to the coordinate point through spatial location association. Based on the preset land use classification system (such as urban areas, rural settlements, cultivated land, forest land, grassland, water areas, unused land, etc.), the specific land use type of each data collection point is determined. For example, if the attribute extracted after overlaying the coordinates of a collection point is "forest land," then its land use type is determined to be forest land. Similarly, during the overlay with the geological background unit map, based on spatial location association, the geological zoning attribute corresponding to the coordinate point is extracted, including core information such as lithological zoning and tectonic unit zoning. Combined with the coding or classification standards of the geological background unit, the geological background unit type to which the data collection point belongs is determined. For example, if the lithological unit code extracted after overlaying a collection point is "G1," then its geological background unit type is determined to be the lithological zoning corresponding to that code.

[0029] The second constraint is the data sensitivity constraint, based on the geographic sensitivity data of that point. The sensitivity reflected (such as the level of radioactivity) modulates the base fluctuation amplitude. The principle is: the higher the investigated or monitored value (the more sensitive), the final fluctuation amplitude should be further reduced based on the previous constraints, so as to achieve key protection for high-value anomalies.

[0030] In practice, based on the determined land use type, a pre-set land use type-basic jitter amplitude comparison table is consulted. The table is formulated with the intensity of regional human activities as the core basis and follows the principle that "the denser the human settlement, the smaller the basic jitter amplitude". For example, the table can be set to a basic jitter amplitude of 40 meters for urban areas, 80 meters for rural settlements, 150 meters for forest and grassland, and 250 meters for unused land. The basic jitter amplitude of the data collection point is directly obtained by matching the land use type.

[0031] Based on the original geographic sensitive data values ​​corresponding to the data collection point, an appropriate scaling factor α (0 < α ≤ 1) is determined. The scaling factor is set according to the core rule that "the higher the geographic sensitive data value, the higher the sensitivity level, and the smaller the corresponding scaling factor." For example, if the radiation dose rate in the preset geographic sensitive data is greater than 150 nGy / h, the sensitivity level is high, and the corresponding scaling factor α = 0.6; if the dose rate does not exceed 150 nGy / h, the sensitivity level is low, and the corresponding scaling factor α = 1.0. The specific numerical range and the correspondence between the scaling factor can be adjusted according to the security requirements and data characteristics of the actual application scenario. The acquired basic jitter amplitude is multiplied by the corresponding scaling factor, i.e. The maximum offset of the coordinate jitter of the data acquisition point is obtained. .

[0032] The third constraint is the geological background boundary constraint, which is a crucial step in preventing desensitization distortion. (Definition point) The geological background unit to which it belongs is In generating candidate desensitized coordinates At that time, it is necessary to ensure that the geological unit into which the candidate point falls must be correct. With the original geological unit The same applies. If a candidate point crosses into other geological units, the jitter is considered invalid, and a new random offset needs to be generated until the geological unit consistency condition is met, or after reaching the preset maximum number of attempts, the nearest point within the adjacent boundary is used.

[0033] At each original coordinate point ( Generate a new, desensitized coordinate system. In the specific implementation process, the original geographic coordinates of the data collection points are used. Centered on the previously calculated maximum offset, A circular spatial range is defined for the radius; this range serves as the effective boundary for coordinate jitter, ensuring that the offset amplitude remains within a safe and controllable range. A candidate point is then selected within this circular spatial range using a preset random function, such as a normally distributed random function. As candidate de-identification coordinates, a normally distributed random function is selected here. This ensures that the candidate de-identification coordinates exhibit reasonable distribution characteristics within the effective range, guaranteeing the randomness of de-identification to enhance security while avoiding excessive concentration or dispersion of offsets, thus maintaining the overall spatial pattern of the data.

[0034] After generating candidate anonymized coordinates, a geological background unit attribution verification is performed to determine whether the candidate coordinates fall within the geological background unit corresponding to the data collection point. The verification process is achieved through spatial location matching, performing spatial overlay analysis between the candidate anonymized coordinates and the boundary vector data of the geological background unit to determine the candidate points. Whether it is within the unit boundary. If the verification result shows that the candidate desensitized coordinates are within the geological background unit, then the candidate coordinates are directly determined as the final desensitized coordinates of the data collection point, and the coordinate desensitization is completed; if the verification result shows that the candidate desensitized coordinates exceed the boundary of the original geological background unit, it means that the offset will cause the data to deviate from the inherent geological environment, which may lead to misinterpretation of subsequent data. Therefore, the initial offset should be discarded, and a new offset and corresponding candidate desensitized coordinates should be generated again through a random function, and the above verification steps should be performed again.

[0035] To avoid infinite loops caused by extreme spatial conditions, such as extremely limited available space within the maximum offset range of the original geological background unit, a maximum number of attempts (e.g., 10) needs to be preset. This number can be flexibly adjusted according to the complexity of the geological environment and data security requirements of the actual application scenario. If no candidate desensitized coordinates that meet the conditions are generated after reaching the preset maximum number of attempts, the shortest distance from the original geographic coordinates to the boundary of its geological background unit is calculated, and the point within that boundary that is closest to the original geographic coordinates is selected as the final desensitized coordinates.

[0036] For example, for point P (located in "woodland", belonging to lithological unit "G1", Data = 120 nGy / h), calculate the maximum offset. =150 meters * 1.0 = 150 meters. A new coordinate is randomly generated within a 150-meter radius around it, and limited to the lithological unit "G1".

[0037] This process is repeated for all data collection points to generate a de-identified coordinate set.

[0038] S4. Apply controlled random perturbation to the geographic sensitive data of each data collection point, and after logarithmic transformation, combine it with the regional background value to perform contrast transformation to obtain contrast data; The core of this step is to achieve data anonymization. The principle of data anonymization is that "the degree of data sensitivity is negatively correlated with the amplitude of disturbance". That is, the higher the original value of the geographically sensitive data, the higher the sensitivity level, and the smaller the corresponding random disturbance amplitude, so as to avoid high-value data losing its scientific reference value due to excessive disturbance; the lower the original value, the lower the sensitivity level, and the disturbance amplitude can be appropriately relaxed to improve the security of anonymization.

[0039] Specifically, a controlled random perturbation is applied to the geographically sensitive data collected at the data collection points to generate intermediate data. Subsequently, on Perform a logarithmic transformation to obtain the result of the logarithmic transformation. Because geographically sensitive data typically exhibits a skewed distribution, with high-value data being relatively concentrated and highly dispersed, direct use in subsequent analysis is susceptible to the influence of extreme values. Furthermore, the absolute error caused by random disturbances is more significant in high-value regions, requiring correction through logarithmic transformation. This embodiment preferably employs a base-10 logarithmic transformation. This transformation effectively compresses the numerical range of high-value data, reduces the impact of extreme values ​​on the overall data distribution, makes skewed data closer to a normal distribution, and improves the stability of subsequent data analysis. Simultaneously, the logarithmic transformation converts the absolute error caused by disturbances into a relative error, significantly suppressing the absolute influence of noise in high-value regions and ensuring that the relative trend of data change remains intact.

[0040] Furthermore, the background value corresponding to the geographically sensitive data type is obtained. This background value is determined based on the overall data characteristics of the study area, using statistical analysis methods such as the mean, median, or 95th percentile of the entire data set, to ensure that the background value objectively reflects the normal environmental level of the study area. Subsequently, the logarithm of the regional background value is calculated. The dataset after logarithmic transformation Logarithm of the background value Perform division to obtain the contrast value data.

[0041] This transformation process normalizes all data to the regional background baseline, so that the value of the contrast data directly reflects the degree of deviation of the corresponding point from the background level. For example, a contrast value greater than 1 indicates that the data at that point is higher than the regional background level, and the larger the value, the more significant the abnormality. A contrast value equal to 1 indicates that it is consistent with the background level; and a value less than 1 indicates that it is lower than the background level. This clearly highlights the abnormal distribution characteristics of the data and provides an intuitive basis for the identification of abnormal areas after subsequent gridding.

[0042] In some implementations, the maximum amplitude of the controlled random perturbation is a preset percentage of the geographically sensitive data value, and the higher the data value, the smaller the preset percentage. For example, for common geographically sensitive data such as radiation dose rate and soil heavy metal concentration, the preset normal perturbation amplitude can be ±10% to ±15% of the original data value; for highly sensitive data where the original value exceeds a preset high value threshold, such as radiation dose rate >180 nGy / h, the perturbation amplitude is tightened to ±5% to ±8% to strictly control the distortion of high-value data.

[0043] S5. The initial grid is generated by interpolating the desensitized coordinates and the chromaticity data. After multiple smoothing and filtering iterations, the final grid is output and a hierarchical visualization is performed based on the final grid.

[0044] Specifically, the desensitized coordinates ( Spatial interpolation is performed between the grid data and the contrast data. The primary spatial interpolation algorithm is ordinary kriging, which constructs a variogram model by analyzing the spatial variation characteristics between data collection points. This effectively quantifies the uncertainty of spatial interpolation, improving the scientific rigor and accuracy of the grid data. During interpolation, the grid cell size should be reasonably set according to the study area, data collection point density, and actual application requirements. For example, a size of 50m × 50m is used to ensure the grid can accurately depict the spatial variation trend of the data while avoiding data redundancy or noise amplification due to excessively small cells. Through interpolation, the discrete contrast data is transformed into a continuous initial grid file. This grid file completely covers the data collection area, with each cell corresponding to a unique contrast data point, intuitively presenting the spatial distribution pattern of the data. After being transformed into a continuous initial grid file, frequency statistics are performed on the grid data to generate a cumulative frequency distribution map and corresponding cumulative frequency distribution curve for the initial grid. This curve objectively reflects the numerical distribution characteristics of the grid data.

[0045] After the initial grid is generated, iterative smoothing filtering removes random noise and local singularities while preserving the core trend characteristics of the data. Moving average filtering or median filtering is preferred. The filtering window size can be set according to the grid cell size and data characteristics, for example, a 5×5 window, ensuring that the filtering process effectively suppresses noise without disrupting the macroscopic spatial structure of the data. After each filtering process, a global frequency count is performed on the current grid data to generate a smoothed cumulative frequency distribution curve. The difference in cumulative frequency distribution is measured by calculating the Kolmogorov–Smirnov statistic D value or the area difference between the curves after two smoothing processes, determining whether the data is stabilizing. If the difference is less than a preset threshold, it indicates that most of the random noise in the grid data has been removed, and it can stably reflect the regional trend. At this point, filtering is stopped, and the current grid is designated as the final grid file. If the difference does not meet the threshold, the filtering and difference calculation steps are repeated until the termination condition is met or the preset maximum number of iterations is reached, avoiding the impact of infinite iteration on processing efficiency.

[0046] In some implementations, hierarchical visualization is performed based on the final grid, specifically including: defining a hierarchical threshold and dividing the grid into hierarchical intervals based on the numerical distribution characteristics of the contrast data of each grid cell in the final grid; generating a first type of visualization layer and a second type of visualization layer based on the hierarchical intervals, and displaying the first type of visualization layer or the second type of visualization layer to the user according to different user permissions.

[0047] Specifically, for the first type of users, such as functional departments and professional analysts, a first type of visualization layer is generated. This involves first determining the grading threshold by combining the gradient changes of the cumulative frequency distribution curve with equidistant grading or natural breakpoint grading methods to ensure that the numerical span of each interval is uniform or conforms to the natural clustering characteristics of the data. Then, the cumulative frequency distribution of the contrast value data of all pixels in the final grid is calculated, and the contrast value corresponding to the preset quantile is selected as the anomaly threshold T. For example, if the cumulative frequency curve of the contrast value data shows obvious distribution inflection points at gradient positions such as 0.2, 0.4, and 0.6, and the overall data dispersion is moderate, a threshold of 0 can be set. 2. A hierarchical interval is defined, delineating multiple consecutive hierarchical intervals from the minimum value to the anomaly threshold. Each interval contains an appropriate number of grid cells. For each grid cell, a detailed contour map or a hierarchical color map with continuous color levels is drawn to clearly present the spatial variation details and gradient differences of the data. All continuous areas with contrast values ​​greater than T are identified as anomaly areas, filled and rendered with a uniform color or pattern, and the fine numerical gradients within these areas are no longer displayed. Neutral descriptions such as "area of ​​special concern" are added. While providing core trend information to professional users, the precise details of anomaly areas are masked, meeting data confidentiality requirements.

[0048] For example, based on the final grid, fine contour lines are drawn at intervals of A=0.2 and rendered with continuous color levels. Then, the cumulative frequency of all pixels in the final grid is calculated, and the 95th percentile value is taken as the chromaticity value T=1.8. In the final grid map, all continuous areas with chromaticity values ​​>1.8 are filled with uniform dark red blocks and marked as "areas of particular interest", thus generating the first visualization layer.

[0049] For the second type of users, such as the general public, a second type of visualization layer is generated. This layer employs a wide-interval grading strategy, simplifying data representation by broadening the grading range and focusing on core distribution intervals. For example, referencing the contrast values ​​corresponding to key quantiles such as 20%, 40%, 80%, and 95% in the cumulative frequency curve, approximately five grading intervals are defined. This ensures that each interval covers a certain proportion of grid cells, avoiding difficulties in public understanding due to overly fine intervals. Specifically, thresholds of 0.2, 0.6, 1.2, and 2.0 can be set, corresponding to five intervals: "Background Area," "High Background Area," "Generally High Background Area," "High Background Area," and "Significantly High Background Area." This not only covers the contrast value range of the vast majority of grid cells but also weakens subtle differences through the wide-interval design, and avoids using warning terms such as "pollution," "abnormal," and "dangerous" that could easily cause public panic or misunderstanding.

[0050] For example, based on the final grid, the grading intervals are redefined: A < 0.2 is Grade I, 0.2 ≤ A < 0.6 is Grade II, 0.6 ≤ A < 1.2 is Grade III, 1.2 ≤ A < 2.0 is Grade IV, and A ≥ 2.0 is Grade V; correspondingly named: Grade I - Level Background, Grade II - Slightly High Background, Grade III - Relatively High Background, Grade IV - High Background, and Grade V - Significantly High Background. A soft gradient of blue-light green-yellow-orange-red is used to generate a second visualization layer for external display.

[0051] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for desensitizing geographically sensitive data, characterized in that, The method includes the following steps: Obtain the original dataset and the prior geographic information map, which includes a land use type map and a geological background unit map. The original dataset includes the geographic coordinates of each data collection point and the corresponding geographic sensitive data. Based on the prior geographic information map, determine the land use type and geological background unit type of each data collection point, and determine the maximum offset based on its corresponding geographic sensitive data; Random offsets are generated within the geological background unit corresponding to the data acquisition point, based on the maximum offset, to obtain the desensitized coordinates. Controlled random perturbation is applied to the geographic sensitive data of each data collection point, and after logarithmic transformation, contrast transformation is performed in combination with regional background values ​​to obtain contrast data; An initial grid is generated by interpolating the desensitized coordinates and the contrast data. After multiple smoothing and filtering iterations, the final grid is output and a hierarchical visualization is performed based on the final grid.

2. The method for desensitizing geographically sensitive data according to claim 1, characterized in that, Geographically sensitive data includes environmental monitoring data and pollution distribution survey data. The environmental monitoring data includes at least the content of radionuclides and radiation dose rate, and the pollution distribution survey data includes at least the concentration of heavy metals in soil, the content of toxic pollutants in water, and the concentration of air pollutants. Each data collection point collects any type of geographically sensitive data.

3. The method for desensitizing geographically sensitive data according to claim 1, characterized in that, Based on the prior geographic information map, determine the land use type and geological background unit type of each data collection point, specifically including: The geographic coordinates of each data collection point are spatially overlaid with the land use type map and geological background unit map in the prior geographic information map; Extract the land use type attribute corresponding to the geographic coordinate from the land use type map to determine the land use type of the data collection point; Extract the geological zoning attributes corresponding to the geographic coordinates from the geological background unit map to determine the geological background unit type to which the data collection point belongs.

4. The method for desensitizing geographically sensitive data according to claim 3, characterized in that, Before acquiring the original dataset and prior geographic information map, a correspondence table between all land types and all basic maximum offsets is established to obtain a land use type-basic jitter amplitude comparison table.

5. A method for desensitizing geographically sensitive data according to claim 4, characterized in that, The maximum offset is determined based on its corresponding geographic sensitive data, specifically including: Based on the land use type of the data collection point, query the land use type-basic jitter amplitude comparison table to determine the maximum basic offset corresponding to the land use type; Based on any type of geographic sensitive data collected from the data collection point, a scaling factor is set according to the magnitude of the numerical value of that type of geographic sensitive data. Multiply the base maximum offset by the scaling factor to obtain the maximum offset.

6. A method for desensitizing geographically sensitive data according to claim 5, characterized in that, Random offsets are generated within the geological background unit corresponding to the data acquisition point, based on the maximum offset, to obtain desensitized coordinates. Specifically, this includes: Using the geographic coordinates of the data collection point as the center, within a spatial range with the maximum offset as the radius, an initial offset is generated through a random function to obtain candidate de-identified coordinates; Determine whether the candidate desensitized coordinates fall within the geological background unit corresponding to the data acquisition point. If the candidate desensitized coordinates are within the geological background unit, then they are determined as the desensitized coordinates of the data acquisition point. If the candidate desensitized coordinates exceed the geological background unit, then the initial offset is discarded and a random offset is regenerated. Repeat the above judgment steps. If no candidate desensitized coordinates that meet the conditions are generated after reaching the preset maximum number of attempts, select the point within the boundary of the geological background unit that is closest to the geographic coordinates as the desensitized coordinates.

7. A method for desensitizing geographically sensitive data according to claim 6, characterized in that, The maximum magnitude of the controlled random disturbance is a preset percentage of the geographically sensitive data value, and the higher the data value, the smaller the preset percentage.

8. A method for desensitizing geographically sensitive data according to claim 7, characterized in that, After logarithmic transformation, contrast transformation is performed using the regional background values ​​to obtain contrast data, specifically including: The controlled random disturbance is superimposed on the geographic sensitive data to obtain intermediate data, and then the intermediate data is subjected to logarithmic transformation. Obtain the regional background value corresponding to the geographically sensitive data type, and calculate the logarithm of the regional background value; The contrast data is obtained by dividing the result of the logarithmic transformation of the intermediate data by the logarithm of the regional background value.

9. A method for desensitizing geographically sensitive data according to claim 8, characterized in that, An initial mesh is generated by interpolating the desensitized coordinates and contrast data. The final mesh is then output after multiple smoothing and filtering iterations. Specifically, this includes: Associate the desensitized coordinates of each data collection point with the corresponding contrast data to construct a spatial interpolation dataset. A spatial interpolation algorithm is used to perform interpolation calculations on the interpolation dataset, and an initial grid file is generated according to a preset cell size; The initial grid file is iteratively smoothed using a filtering algorithm with a preset filtering window size. If the frequency distribution difference between the cumulative frequency distribution curves of the two consecutive tests is less than a preset threshold, the iterative smoothing ends, and the filtered grid data is determined as the final grid and output.

10. A method for desensitizing geographically sensitive data according to claim 9, characterized in that, The final grid is used for hierarchical visualization, which includes: defining hierarchical thresholds and dividing the grid into hierarchical intervals based on the numerical distribution characteristics of the contrast data of each grid cell in the final grid; generating a first type of visualization layer and a second type of visualization layer based on the hierarchical intervals; and displaying the first type of visualization layer or the second type of visualization layer to the user according to different user permissions.

Images