Yellow River downstream city landscape ecological risk analysis method based on high score data

Abstract

The invention discloses a Yellow River downstream city landscape ecological risk analysis method based on high-score data, and relates to the technical field of resource analysis, and the method comprises the steps: obtaining land utilization classification data, dividing an ecological risk evaluation price network based on the land utilization classification data, carrying out the statistics of landscape pattern indexes in each grid, and carrying out the calculation of the ecological risk evaluation price network. Generating landscape interference degree space distribution data; obtaining multi-dimensional geographical environment factor data of the research area, performing spatial correlation analysis on the multi-dimensional geographical environment factor data and the land utilization classification data, and screening out dominant environment factors; generating landscape vulnerability spatial distribution data based on a spatial association relationship between the dominant environmental factors and the land utilization classification data; according to the method, all technical links are connected in series by taking a uniform grid space as an analysis reference, so that refined space representation and scientific coupling fusion of the landscape interference degree and the landscape vulnerability are realized, and the space accuracy of a landscape ecological risk analysis result is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of resource analysis technology, and in particular to a method for ecological risk analysis of urban landscapes in the lower reaches of the Yellow River based on high-resolution data. Background Technology

[0002] In the analysis of urban landscape ecological risks in the lower reaches of the Yellow River, existing technologies are difficult to adapt to the spatial heterogeneity of topography, hydrology and other factors in the region. The quantification of landscape disturbance lacks rigorous spatial autocorrelation test and outlier correction mechanism. The screening of environmental factors has not achieved refined spatial matching. At the same time, the determination of landscape vulnerability cannot achieve accurate coupling and quantification of land use type and environmental factors. The various analysis links are isolated from each other and lack a unified spatial analysis benchmark.

[0003] Existing technologies, in the entire process of landscape ecological risk analysis, are unable to achieve refined spatial characterization of landscape disturbance and vulnerability, nor have they formed a scientific coupling and integration method. As a result, the final ecological risk analysis results are difficult to accurately reflect the actual spatial differentiation characteristics of landscape ecological risks in the study area. The scientificity, accuracy and reliability of the analysis results are insufficient, and they cannot provide effective technical support for the prevention and control of ecological risks and the formulation of ecological protection plans in cities in the lower reaches of the Yellow River. Summary of the Invention

[0004] This invention provides a method for analyzing the ecological risks of urban landscapes in the lower reaches of the Yellow River based on high-resolution data to address the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides a method for ecological risk analysis of urban landscapes in the lower reaches of the Yellow River based on high-resolution data, comprising: S1. Acquire high-resolution remote sensing images of the study area, interpret the ground features in the high-resolution remote sensing images, and obtain land use classification data; S2. Based on land use classification data, divide the ecological risk assessment grid, statistically analyze the landscape pattern index within each grid, and generate spatial distribution data of landscape disturbance. S3. Obtain multidimensional geographic environmental factor data of the study area, perform spatial correlation analysis between the multidimensional geographic environmental factor data and land use classification data, and screen out the dominant environmental factors; S4. Based on the spatial correlation between dominant environmental factors and land use classification data, generate spatial distribution data of landscape vulnerability; S5. The spatial distribution data of landscape disturbance degree and the spatial distribution data of landscape vulnerability are fused and processed to generate a spatial distribution map of landscape ecological risk.

[0006] Preferably, acquiring high-resolution remote sensing images of the study area includes: Send image orders for the study area to the remote sensing satellite ground receiving station to obtain raw remote sensing image products; Geometric correction is performed on the original remote sensing image products to eliminate geometric distortion and generate spatially registered high-resolution remote sensing images.

[0007] Preferably, the step of interpreting land cover data from high-resolution remote sensing images to obtain land use classification data includes: Multi-scale segmentation is performed on the spatially registered high-resolution remote sensing image to generate an image object layer; Based on a pre-defined sample library of ground features, the image object layer is classified in an object-oriented manner, and land use type attributes are assigned to generate land use classification data.

[0008] Preferably, the step of dividing the ecological risk assessment grid based on land use classification data, statistically analyzing the landscape pattern index within each grid, and generating spatial distribution data of landscape disturbance includes: Based on land use classification data, a rule grid covering the study area was created as an ecological risk assessment grid. Landscape fragmentation, landscape separation, and landscape dominance indices are extracted from each grid cell. The fragmentation, separation, and dominance indices are then fused to generate a quantitative value for the degree of landscape disturbance in each grid cell. Spatial autocorrelation test was performed on the quantitative values ​​of landscape disturbance to identify and mark anomalous grids; The median of the disturbance level quantization value of the abnormal grid is used to replace the disturbance level quantization value of the abnormal grid to generate the corrected landscape disturbance level quantization value. The corrected quantitative values ​​of landscape disturbance from all grid cells are collected to form spatial distribution data of landscape disturbance.

[0009] Preferably, the step of performing spatial autocorrelation testing on the quantification value of landscape disturbance degree to identify and mark anomalous grids includes: The global Moran index was used to measure the quantitative values ​​of landscape disturbance in all grid cells to obtain the global Moran index value. A significance test is performed on the global Moran index value to obtain the corresponding significance test value. The significance test value is compared with the preset significance level threshold. If the significance test value is less than the preset significance level threshold, it is determined that the quantitative value of landscape disturbance degree has significant spatial clustering characteristics. For landscape disturbances that are determined to have significant spatial clustering characteristics, the local Moran index of each grid is measured to obtain the local Moran index value and the corresponding local significance test value of each grid. Grids with negative local Moran index values ​​and local significance test values ​​less than a preset significance level threshold are selected and marked as spatially anomalous grids.

[0010] Preferably, the step of acquiring multidimensional geographical environmental factor data of the study area, performing spatial correlation analysis between the multidimensional geographical environmental factor data and land use classification data, and screening out the dominant environmental factors includes: Spatial overlay analysis was performed on multidimensional geographic environmental factor data and land use classification data to extract environmental factor values ​​corresponding to each land use type and generate environmental factor value sets. Geographically weighted principal component decomposition is performed on the environmental factor value set. The eigenvectors of the local covariance matrix are solved at each grid center point to generate the local factor loading coefficients at each spatial location. Spatial interpolation is performed on the local factor loading coefficients to generate a continuously distributed factor loading surface; On the factor loading surface, the loading coefficients of each environmental factor are compared pixel by pixel. The environmental factor with the largest loading coefficient at each pixel position is selected as the dominant environmental factor of that pixel, forming a set of dominant environmental factors for spatial variability.

[0011] Preferably, the step of performing geographically weighted principal component decomposition on the environmental factor value set, solving for the eigenvectors of the local covariance matrix at each grid center point, and generating local factor loading coefficients at each spatial location includes: The environmental factor value set is standardized to generate a standardized environmental factor matrix. Construct a spatial weight matrix, and use the spatial weight matrix to weight the environmental factor values ​​for each grid center point to generate a locally weighted environmental factor matrix; Eigenvalue decomposition is performed on the locally weighted environmental factor matrix, and its principal component eigenvectors are extracted as the local factor loading coefficients of the grid center point.

[0012] Preferably, the generation of spatial distribution data of landscape vulnerability based on the spatial correlation between dominant environmental factors and land use classification data includes: Extract the numerical distribution range of dominant environmental factors across different land use types; The numerical distribution range is input into the preset ecological response coupling processor. The ecological response coupling processor has a built-in interaction matrix between land use type and environmental factors. The interaction matrix is ​​used to map the numerical distribution range and output the ecological response level of each land use type under different combinations of environmental factors. Fill in the ecological response level into the corresponding spatial unit in the land use classification data; The ecological response levels of all spatial units are aggregated to generate spatial distribution data of landscape vulnerability.

[0013] Preferably, the ecological response coupling processor has a built-in interaction matrix between land use type and environmental factors. This interaction matrix is ​​used to map the numerical distribution range, outputting the ecological response level of each land use type under different combinations of environmental factors, including: Read the pre-set interaction matrix between land use type and environmental factors. The interaction matrix records the sensitivity coefficient and tolerance threshold of each land use type to each dominant environmental factor. The dominant environmental factor values ​​of each spatial unit are compared with the tolerance threshold to determine the stress level of each environmental factor in that unit. Based on the sensitivity coefficients matched to the stress levels, the sensitivity coefficients of each environmental factor are fused to generate the ecological response level of the unit.

[0014] Preferably, the step of fusing spatial distribution data of landscape disturbance degree with spatial distribution data of landscape vulnerability to generate a spatial distribution map of landscape ecological risk includes: The grid-by-grid method retrieves the quantitative values ​​of disturbance intensity from the spatial distribution data of landscape disturbance intensity and the ecological response levels from the spatial distribution data of landscape vulnerability. The quantified value of the disturbance degree is normalized and aligned with the ecological response level to generate aligned disturbance degree contribution value and vulnerability degree contribution value. The ecological risk index of the grid is generated by multiplying and fusing the disturbance contribution value and the vulnerability contribution value. The ecological risk indices of all grid cells are collected and arranged according to their spatial location to generate a spatial distribution map of landscape ecological risk.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. Based on high-resolution remote sensing data, a full-process analysis system for landscape ecological risk was constructed to adapt to the spatial heterogeneity of cities in the lower reaches of the Yellow River. Using a unified grid space as the analysis benchmark, various technical links were connected, realizing the refined spatial representation and scientific coupling and integration of landscape disturbance degree and landscape vulnerability. The system accurately captured the differences in landscape ecological risk in different spatial locations in the study area, significantly improving the spatial accuracy of landscape ecological risk analysis results. It can truly reflect the actual spatial differentiation characteristics of urban landscape ecological risk in the lower reaches of the Yellow River, and provide accurate spatial data support for regional ecological risk prevention and control and ecological protection planning.

[0016] 2. High-precision land use classification data is obtained through precise processing and interpretation of high-resolution remote sensing images. Rigorous spatial autocorrelation testing and outlier correction of landscape disturbance quantification values ​​are performed using the Moran index. Geographically weighted principal component decomposition is combined to achieve refined spatial screening of dominant environmental factors. Simultaneously, precise coupling quantification of landscape vulnerability is achieved based on the interaction matrix between land use type and environmental factors. Furthermore, scientific coupling of two types of risk factors is achieved through normalization alignment and product fusion. These multi-dimensional improvements enhance the reliability of the analysis data and the scientific rigor of the analysis methods, resulting in a final landscape ecological risk spatial distribution map that is both accurate and practical, providing effective technical support for the refined management of urban landscape ecological risks in the lower reaches of the Yellow River. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data, as provided in an embodiment of the present invention. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] 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, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1, referring to Figure 1 The diagram shown is a flowchart illustrating a method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data, according to an embodiment of the present invention. In this embodiment, the method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data includes: S1. Acquire high-resolution remote sensing images of the study area, interpret the ground features in the high-resolution remote sensing images, and obtain land use classification data; S2. Based on land use classification data, divide the ecological risk assessment grid, statistically analyze the landscape pattern index within each grid, and generate spatial distribution data of landscape disturbance. S3. Obtain multidimensional geographic environmental factor data of the study area, perform spatial correlation analysis between the multidimensional geographic environmental factor data and land use classification data, and screen out the dominant environmental factors; S4. Based on the spatial correlation between dominant environmental factors and land use classification data, generate spatial distribution data of landscape vulnerability; S5. The spatial distribution data of landscape disturbance degree and the spatial distribution data of landscape vulnerability are fused and processed to generate a spatial distribution map of landscape ecological risk.

[0021] In embodiments of the present invention, acquiring high-resolution remote sensing images of the study area includes: Send image orders for the study area to the remote sensing satellite ground receiving station to obtain raw remote sensing image products; Geometric correction is performed on the original remote sensing image products to eliminate geometric distortion and generate spatially registered high-resolution remote sensing images.

[0022] In embodiments of the present invention, land use classification data is obtained by interpreting ground features from high-resolution remote sensing images, including: Multi-scale segmentation is performed on the spatially registered high-resolution remote sensing image to generate an image object layer; Based on a pre-defined sample library of ground features, the image object layer is classified in an object-oriented manner, and land use type attributes are assigned to generate land use classification data.

[0023] Specifically, an image order covering the urban study area in the lower reaches of the Yellow River is submitted to the Gaofen remote sensing satellite ground receiving station, and raw remote sensing image products containing information on the surface reflection, texture and spatial location of the study area are received.

[0024] Furthermore, the coordinates of ground-measured control points evenly distributed within the study area are compared point by point with the corresponding pixel coordinates in the original remote sensing image. By performing translation, rotation, and scaling operations on the pixel coordinates, geometric distortion of the image is eliminated, generating a high-resolution remote sensing image after spatial registration that accurately matches the actual geographic space of the study area.

[0025] Furthermore, based on the main land cover types in the study area, the segmentation scale was set, and the spatially registered high-resolution remote sensing images were segmented into multi-scale segments by region, dividing the images into image object layers with clear boundaries and consistent internal spectral characteristics.

[0026] Finally, a pre-set land feature interpretation sample library is retrieved, which contains typical spectral and texture feature samples of various land features in the study area. Each object in the image object layer is classified and determined in an object-oriented manner, and a corresponding land use type attribute is assigned to each object, ultimately generating complete land use classification data.

[0027] Overall, this scheme acquires high-resolution remote sensing images of the study area and performs geometric correction processing to eliminate image geometric distortion and achieve accurate registration with the actual geographic space. Then, the spatially registered images are segmented at multiple scales to generate image object layers. Combined with a pre-set land cover interpretation sample library, the image object layers are classified in an object-oriented manner and assigned land use type attributes, ultimately obtaining land use classification data.

[0028] Overall, this scheme ensures that the image interpretation unit closely matches the spatial morphology and spectral characteristics of the actual ground features, breaking through the limitations of traditional pixel-level interpretation. At the same time, the high precision of the high-resolution image combined with the spatial registration effect of geometric precision correction significantly improves the accuracy of ground feature interpretation.

[0029] Furthermore, the exclusive matching of the pre-set sample library effectively reduces interpretation errors. The resulting land use classification data can reflect the spatial pattern of land use in the study area in a refined and accurate manner, providing high-precision and highly matched basic spatial data for subsequent landscape ecological risk analysis, thus ensuring the scientific nature and accuracy of subsequent analysis work from the source.

[0030] In embodiments of the present invention, ecological risk assessment grids are divided based on land use classification data, and landscape pattern indices within each grid are statistically analyzed to generate spatial distribution data of landscape disturbance, including: Based on land use classification data, a rule grid covering the study area was created as an ecological risk assessment grid. Landscape fragmentation, landscape separation, and landscape dominance indices are extracted from each grid cell. The fragmentation, separation, and dominance indices are then fused to generate a quantitative value for the degree of landscape disturbance in each grid cell. Spatial autocorrelation test was performed on the quantitative values ​​of landscape disturbance to identify and mark anomalous grids; The median of the disturbance level quantization value of the abnormal grid is used to replace the disturbance level quantization value of the abnormal grid to generate the corrected landscape disturbance level quantization value. The corrected quantitative values ​​of landscape disturbance from all grid cells are collected to form spatial distribution data of landscape disturbance.

[0031] In embodiments of the present invention, spatial autocorrelation testing is performed on the quantified values ​​of landscape disturbance to identify and mark anomalous grids, including: The global Moran index was used to measure the quantitative values ​​of landscape disturbance in all grid cells to obtain the global Moran index value. A significance test is performed on the global Moran index value to obtain the corresponding significance test value. The significance test value is compared with the preset significance level threshold. If the significance test value is less than the preset significance level threshold, it is determined that the quantitative value of landscape disturbance degree has significant spatial clustering characteristics. For landscape disturbances that are determined to have significant spatial clustering characteristics, the local Moran index of each grid is measured to obtain the local Moran index value and the corresponding local significance test value of each grid. Grids with negative local Moran index values ​​and local significance test values ​​less than a preset significance level threshold are selected and marked as spatially anomalous grids.

[0032] Specifically, based on the clearly defined spatial boundaries of the study area using land use classification data, square grids are uniformly divided along the latitude and longitude of the study area to ensure that each grid can completely cover the land use patches within the corresponding area, and that the grid edges are precisely connected to the boundaries of the study area, ultimately forming a regular grid that covers the entire study area. This regular grid is the ecological risk assessment grid.

[0033] Furthermore, each grid in the ecological risk assessment grid is examined one by one to identify patches of all land use types within each grid. Landscape fragmentation is determined by statistically analyzing the number and degree of fragmentation of patches within the grid. Landscape separation is determined by observing the dispersion distance of different types of patches within the grid. Landscape dominance is determined by comparing the area ratio of a certain type of patch with all patches within the grid. The results of the three indices are superimposed and integrated to obtain a unique quantitative value of the degree of landscape disturbance for each grid.

[0034] Furthermore, the landscape disturbance degree quantification values ​​of all grids in the ecological risk assessment grid of the urban study area in the lower reaches of the Yellow River were traversed. First, the total number of grids was counted and the average value of all grid quantification values ​​was calculated. Then, a grid spatial adjacency weight matrix was constructed. Based on this matrix, the adjacent grids of each grid were determined. Combining the difference between the grid quantification value and the average value, and the adjacency weight matrix, the global Moran index was measured, and the global Moran index value corresponding to the landscape disturbance degree quantification value of the study area was calculated.

[0035] Furthermore, a statistical significance test is conducted based on the calculated global Moran index value. The test distribution is generated through random permutation, and the corresponding significance test value is calculated. A significance level threshold is pre-set in conjunction with the actual needs of urban landscape ecological risk analysis in the lower reaches of the Yellow River. The significance test value is compared with the preset significance level threshold. If the significance test value is less than the preset significance level threshold, it is determined that the quantitative value of landscape disturbance degree has significant spatial clustering characteristics.

[0036] Furthermore, for the quantitative values ​​of landscape disturbance degree that are determined to have significant spatial clustering characteristics, local Moran index measurements were carried out for each grid of the ecological risk assessment grid in the study area. Taking a single grid as the core, the local Moran index value corresponding to each grid was calculated by combining the quantitative values ​​of landscape disturbance degree of the grid and the surrounding neighboring grids, the difference between the two and the overall average value, and the spatial adjacency weight. At the same time, the same random permutation method was used to carry out statistical significance tests on each local Moran index value to obtain the local significance test value corresponding to each grid.

[0037] Furthermore, the local Moran's index values ​​and corresponding local significance test values ​​of all grids were analyzed, and the two numerical judgment criteria were checked simultaneously. Grids with negative local Moran's index values ​​and local significance test values ​​less than the preset significance level threshold were screened out and uniformly marked as spatially anomalous grids.

[0038] Furthermore, for each spatially anomalous grid, eight neighboring grids are selected as neighborhood grids. The landscape disturbance degree quantification values ​​of these neighborhood grids are collected and arranged in order of magnitude. The value in the middle position after arrangement is taken as the median. The original landscape disturbance degree quantification value of the spatially anomalous grid is replaced with this median to generate the corrected landscape disturbance degree quantification value.

[0039] Finally, the corrected landscape disturbance quantification values ​​of all grids were arranged one by one according to the actual spatial location of each grid in the study area of ​​the lower Yellow River city, and the quantification results of all spatial locations were fully collected to form the spatial distribution data of landscape disturbance.

[0040] In summary, this scheme clarifies the core elements and operational methods of Moran's index calculation, making spatial autocorrelation testing feasible and avoiding result bias caused by ambiguity in testing methods. The stratified significance test of global and local data, combined with the test method of random permutation, makes the determination of spatial clustering characteristics more statistically rigorous and effectively eliminates the interference of random distribution.

[0041] The dual screening criteria for outlier grids can accurately identify outlier grids that contradict the surrounding spatial characteristics, avoiding misjudgments caused by single numerical judgments. Median correction further eliminates the impact of outliers on the overall data. The resulting spatial distribution data of landscape disturbance not only fits the actual spatial distribution characteristics of urban landscape disturbance in the lower reaches of the Yellow River, but also has both data accuracy and spatial rationality. This provides highly reliable quantitative data support for subsequent coupled analysis of landscape ecological risks, and greatly improves the scientific nature of the overall analysis.

[0042] In embodiments of the present invention, multidimensional geographical environmental factor data of the study area are obtained, and spatial correlation analysis is performed between the multidimensional geographical environmental factor data and land use classification data to screen out the dominant environmental factors, including: Spatial overlay analysis was performed on multidimensional geographic environmental factor data and land use classification data to extract environmental factor values ​​corresponding to each land use type and generate environmental factor value sets. Geographically weighted principal component decomposition is performed on the environmental factor value set. The eigenvectors of the local covariance matrix are solved at each grid center point to generate the local factor loading coefficients at each spatial location. Spatial interpolation is performed on the local factor loading coefficients to generate a continuously distributed factor loading surface; On the factor loading surface, the loading coefficients of each environmental factor are compared pixel by pixel. The environmental factor with the largest loading coefficient at each pixel position is selected as the dominant environmental factor of that pixel, forming a set of dominant environmental factors for spatial variability.

[0043] In embodiments of the present invention, a geographically weighted principal component decomposition is performed on the environmental factor value set, and the eigenvectors of the local covariance matrix are solved at each grid center point to generate local factor loading coefficients at each spatial location, including: The environmental factor value set is standardized to generate a standardized environmental factor matrix. Construct a spatial weight matrix, and use the spatial weight matrix to weight the environmental factor values ​​for each grid center point to generate a locally weighted environmental factor matrix; Eigenvalue decomposition is performed on the locally weighted environmental factor matrix, and its principal component eigenvectors are extracted as the local factor loading coefficients of the grid center point.

[0044] The formula for calculating the local factor loading coefficient is as follows: ; In the formula, Indicates the first The local factor loading coefficient vector of each grid is used to measure the contribution of each environmental factor to the variation in the local space; X represents the standardized environmental factor matrix. Indicates the first A spatial weighted diagonal matrix is ​​constructed with the center point of each grid as the core. Represents the locally weighted covariance matrix; This represents the eigenvector corresponding to the largest eigenvalue after performing eigenvalue decomposition on the matrix.

[0045] Specifically, data on four types of geographical environmental factors—topography, elevation, soil texture, river network density, and precipitation distribution—were collected from the urban study area in the lower reaches of the Yellow River. This ensured that the spatial scope of each type of data was completely consistent with the land use classification data, forming a complete multidimensional geographical environmental factor data set.

[0046] Furthermore, multidimensional geographic environmental factor data and land use classification data are spatially overlaid to ensure precise spatial correspondence between the two types of data. Each land use type is investigated one by one, and the specific values ​​of various environmental factors corresponding to all spatial units of that type are extracted and integrated to generate an environmental factor value set.

[0047] Furthermore, take all values ​​of each environmental factor in the environmental factor value set, subtract the minimum value from the maximum value of the factor to obtain the difference, then subtract the minimum value of the factor from each value and divide by the difference to complete the standardization transformation of all values ​​and generate a standardized environmental factor matrix.

[0048] Furthermore, taking the center point of each grid as the core, weights are set according to the distance between the surrounding grids and the center point. The closer the distance, the greater the weight; the farther the distance, the smaller the weight. The weight values ​​are arranged according to the grid position to form a spatial weight diagonal matrix. This matrix is ​​multiplied element-wise with the standardized environmental factor matrix to complete the weighting process and generate a local weighted environmental factor matrix.

[0049] Furthermore, eigenvalue decomposition is performed on the local weighted environmental factor matrix to extract the eigenvector corresponding to the largest eigenvalue in the matrix. This eigenvector is the local factor loading coefficient of the grid center point. The local factor loading coefficients of all grid center points are obtained one by one in this way.

[0050] Furthermore, local factor loading coefficients at the center points of adjacent grids are selected, and the loading coefficients of the blank areas between grids are supplemented by the linear transition of adjacent values ​​to achieve spatial interpolation filling, ultimately generating a continuously distributed factor loading surface.

[0051] Finally, the factor loading surface is checked pixel by pixel, and the loading coefficients of various environmental factors at each pixel location are compared. The environmental factor with the largest loading coefficient is selected as the dominant environmental factor of that pixel. The dominant environmental factors of all pixels are integrated to form a set of dominant environmental factors for spatial variation.

[0052] In summary, this scheme accurately captures the differences in the contribution of environmental factors to variations in different spatial locations, abandons the homogenization processing mode of traditional global analysis, fills the numerical gaps between grids with spatial interpolation, and achieves refined spatial matching of dominant environmental factors through pixel-by-pixel screening. The generated set of dominant environmental factors can truly reflect the core ecological influencing factors in different locations of the study area, effectively eliminate the interference of non-dominant factors, make subsequent landscape vulnerability analysis more targeted, and significantly improve the spatial refinement level of the overall analysis.

[0053] In embodiments of the present invention, spatial distribution data of landscape vulnerability is generated based on the spatial correlation between dominant environmental factors and land use classification data, including: Extract the numerical distribution range of dominant environmental factors across different land use types; The numerical distribution range is input into the preset ecological response coupling processor. The ecological response coupling processor has a built-in interaction matrix between land use type and environmental factors. The interaction matrix is ​​used to map the numerical distribution range and output the ecological response level of each land use type under different combinations of environmental factors. Fill in the ecological response level into the corresponding spatial unit in the land use classification data; The ecological response levels of all spatial units are aggregated to generate spatial distribution data of landscape vulnerability.

[0054] In embodiments of the present invention, the ecological response coupling processor has a built-in interaction matrix between land use type and environmental factors. The interaction matrix is ​​used to map the numerical distribution range, outputting the ecological response level of each land use type under different combinations of environmental factors, including: Read the pre-set interaction matrix between land use type and environmental factors. The interaction matrix records the sensitivity coefficient and tolerance threshold of each land use type to each dominant environmental factor. The dominant environmental factor values ​​of each spatial unit are compared with the tolerance threshold to determine the stress level of each environmental factor in that unit. Based on the sensitivity coefficients matched to the stress levels, the sensitivity coefficients of each environmental factor are fused to generate the ecological response level of the unit.

[0055] The formula for calculating the ecological response level is as follows: ; In the formula, The ecological response level of the k-th spatial unit is dimensionless and can be further mapped to a level identifier; M represents the total number of dominant environmental factors. This represents the value of the k-th unit on the m-th dominant environmental factor; This represents the tolerance threshold of the land use type of the k-th unit to the m-th environmental factor. This represents the sensitivity coefficient of the land use type of the k-th unit to the m-th environmental factor; This indicates an indicator function that takes the value 1 when the condition within the parentheses is true, and 0 otherwise.

[0056] Specifically, for the five land use types of cultivated land, forest land, water area, construction land and unused land in the urban study area of ​​the lower Yellow River, all spatial units corresponding to each type in the land use classification data were located. The specific values ​​of the dominant environmental factors such as topographic elevation, river network density and precipitation distribution in each spatial unit were extracted. The values ​​of the dominant environmental factors of all spatial units under the same land use type were integrated. The maximum and minimum values ​​were determined by comparing the values, and the numerical distribution range of each dominant environmental factor in different land use types was clarified.

[0057] Furthermore, a land use type and environmental factor interaction matrix, pre-defined based on the ecological characteristics of cities in the lower reaches of the Yellow River, was retrieved. This matrix sets sensitivity coefficients and tolerance thresholds for each dominant environmental factor, including topographic elevation, river network density, and precipitation distribution, for five land use types in the study area. The sensitivity coefficients are set according to the sensitivity of land use type to environmental factors, with water bodies showing a higher sensitivity coefficient to river network density than other types. The tolerance thresholds are set based on historical environmental monitoring data and ecological threshold experience in the study area. The precipitation tolerance threshold for cultivated land is taken as the critical precipitation value for crop growth. The extracted numerical distribution range is then input into the matrix for mapping processing.

[0058] Furthermore, each spatial unit in the study area was analyzed one by one, and the actual values ​​of each dominant environmental factor in the unit were extracted. The value of each dominant environmental factor was compared with the tolerance threshold corresponding to the land use type of the unit in the interaction matrix. If the value of a dominant environmental factor is greater than or equal to the tolerance threshold, the environmental factor is determined to be at a high stress level in the unit. If the value is less than the tolerance threshold, it is determined to be at a low stress level, thus clarifying the stress level of each dominant environmental factor in each spatial unit.

[0059] Furthermore, based on the stress level of each dominant environmental factor in each spatial unit, the corresponding sensitivity coefficient in the interaction matrix is ​​matched. Under high stress levels, the indicator function takes a value of 1, and the sensitivity coefficient is included in the calculation. Under low stress levels, the indicator function takes a value of 0, and the sensitivity coefficient is not included in the calculation.

[0060] Then, the products of the sensitivity coefficients of all the dominant environmental factors in the unit and the corresponding indicator function values ​​are added together, and the sum is the ecological response level of the unit. For example, if a certain water spatial unit has two dominant environmental factors, river network density and topographic elevation, and the river network density value reaches the tolerance threshold with a sensitivity coefficient of 0.9, while the topographic elevation does not reach the threshold with a sensitivity coefficient of 0.2, then the ecological response level of the unit is 0.9×1+0.2×0=0.9.

[0061] Furthermore, each spatial unit in the land use classification data is matched one by one with the calculated ecological response level, and the level is accurately filled into the corresponding spatial unit position in the land use classification data to ensure that the ecological response level of each spatial unit corresponds completely with its own land use type and dominant environmental factor value, without any mismatch.

[0062] Finally, we went through all the spatial units in the land use classification data of the urban study area in the lower reaches of the Yellow River one by one, collected the ecological response level that had been filled in for each unit, arranged them one by one according to the actual geographic location of each spatial unit in the study area, clarified the ecological response level corresponding to each spatial location in the study area, integrated the ecological response level information of all spatial units, and finally generated spatial distribution data of landscape vulnerability.

[0063] In summary, this scheme precisely couples the ecological characteristics of land use types with the spatial variation of dominant environmental factors, and achieves quantitative determination of ecological response levels based on the interaction matrix, effectively avoiding the bias of subjective qualitative assessment. At the same time, it accurately assigns response levels to each spatial unit, giving landscape vulnerability refined spatial distribution characteristics. It can clearly reflect the differences in ecological vulnerability of different land use types and different spatial locations in the study area. The generated spatial distribution data of landscape vulnerability is accurate and targeted, laying a high-quality data foundation for subsequent risk coupling and fusion.

[0064] In embodiments of the present invention, spatial distribution data of landscape disturbance degree and spatial distribution data of landscape vulnerability are fused to generate a spatial distribution map of landscape ecological risk, including: The grid-by-grid method retrieves the quantitative values ​​of disturbance intensity from the spatial distribution data of landscape disturbance intensity and the ecological response levels from the spatial distribution data of landscape vulnerability. The quantified value of the disturbance degree is normalized and aligned with the ecological response level to generate aligned disturbance degree contribution value and vulnerability degree contribution value. The ecological risk index of the grid is generated by multiplying and fusing the disturbance contribution value and the vulnerability contribution value. The ecological risk indices of all grid cells are collected and arranged according to their spatial location to generate a spatial distribution map of landscape ecological risk.

[0065] Specifically, using the ecological risk assessment grid of the urban study area in the lower reaches of the Yellow River as a unified spatial carrier, the quantitative value of the degree of landscape disturbance corresponding to each grid in the spatial distribution data of landscape disturbance is read sequentially according to the unique spatial number of the grid, and the ecological response level corresponding to the grid in the spatial distribution data of landscape vulnerability is read at the same time, to ensure accurate spatial matching of the two types of data in a single grid.

[0066] Furthermore, the maximum and minimum values ​​of the landscape disturbance degree quantification of all grids in the study area are taken, and the quantification value of a single grid is converted into a value between 0 and 1 to obtain the disturbance degree contribution value of that grid. At the same time, the maximum and minimum values ​​of the ecological response level of all grids are taken, and the level of a single grid is converted into a value between 0 and 1 to obtain the vulnerability contribution value of that grid. This completes the normalization and alignment processing of the two types of data.

[0067] Furthermore, the disturbance contribution value and vulnerability contribution value of each grid after normalization and alignment are multiplied together. Taking a grid with a disturbance contribution value of 0.6 and a vulnerability contribution value of 0.9 as an example, the result of multiplying the two is 0.54, which is the ecological risk index of that grid. The index calculation for all grids is completed in this way.

[0068] Finally, the ecological risk index of all grids in the study area was collected, and each grid was arranged according to its actual geographical location in the study area. The ecological risk index was accurately assigned to the corresponding spatial grid unit, and the spatial information of the ecological risk index of all grids was integrated to finally generate a spatial distribution map of landscape ecological risk.

[0069] In summary, this scheme achieves refined spatial coupling of the two core risk elements—landscape disturbance and ecological vulnerability—based on a grid. Normalization and alignment effectively avoid fusion bias caused by differences in data units. Product fusion can accurately quantify the comprehensive ecological risk level of a single grid, making the risk results more consistent with the actual risk characteristics of urban landscape ecology in the lower reaches of the Yellow River. The distribution map generated by spatial location can clearly show the spatial differentiation pattern of ecological risk in the study area. Quantitative verification based on historical events and objective data assessment of the reliability of the distribution map significantly improve the scientificity and credibility of the analysis results. This provides precise spatial data support and decision-making basis for the zoning and prevention of urban ecological risks and the formulation of landscape ecological protection plans in the lower reaches of the Yellow River.

[0070] In the several embodiments provided by this invention, it should be understood that the disclosed method can be implemented in other ways.

[0071] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0072] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, and technology that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data, characterized in that... The method includes: S1. Acquire high-resolution remote sensing images of the study area, interpret the ground features in the high-resolution remote sensing images, and obtain land use classification data; S2. Based on land use classification data, divide the ecological risk assessment grid, statistically analyze the landscape pattern index within each grid, and generate spatial distribution data of landscape disturbance. S3. Obtain multidimensional geographic environmental factor data of the study area, perform spatial correlation analysis between the multidimensional geographic environmental factor data and land use classification data, and screen out the dominant environmental factors; S4. Based on the spatial correlation between dominant environmental factors and land use classification data, generate spatial distribution data of landscape vulnerability; S5. The spatial distribution data of landscape disturbance degree and the spatial distribution data of landscape vulnerability are fused and processed to generate a spatial distribution map of landscape ecological risk.

2. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 1, characterized in that, The acquisition of high-resolution remote sensing images of the study area includes: Send image orders for the study area to the remote sensing satellite ground receiving station to obtain raw remote sensing image products; Geometric correction is performed on the original remote sensing image products to eliminate geometric distortion and generate spatially registered high-resolution remote sensing images.

3. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 2, characterized in that, The process of interpreting ground features from high-resolution remote sensing images to obtain land use classification data includes: Multi-scale segmentation is performed on the spatially registered high-resolution remote sensing image to generate an image object layer; Based on a pre-defined sample library of ground features, the image object layer is classified in an object-oriented manner, and land use type attributes are assigned to generate land use classification data.

4. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 1, characterized in that, The ecological risk assessment grid is divided based on land use classification data. Landscape pattern indices within each grid are statistically analyzed to generate spatial distribution data of landscape disturbance, including: Based on land use classification data, a rule grid covering the study area was created as an ecological risk assessment grid. Landscape fragmentation, landscape separation, and landscape dominance indices are extracted from each grid cell. The fragmentation, separation, and dominance indices are then fused to generate a quantitative value for the degree of landscape disturbance in each grid cell. Spatial autocorrelation test was performed on the quantitative values ​​of landscape disturbance to identify and mark anomalous grids; The median of the disturbance level quantization value of the abnormal grid is used to replace the disturbance level quantization value of the abnormal grid to generate the corrected landscape disturbance level quantization value. The corrected quantitative values ​​of landscape disturbance from all grid cells are collected to form spatial distribution data of landscape disturbance.

5. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 4, characterized in that, The process of performing spatial autocorrelation testing on the quantified values ​​of landscape disturbance to identify and mark anomalous grids includes: The global Moran index was used to measure the quantitative values ​​of landscape disturbance in all grid cells to obtain the global Moran index value. A significance test is performed on the global Moran index value to obtain the corresponding significance test value. The significance test value is compared with the preset significance level threshold. If the significance test value is less than the preset significance level threshold, it is determined that the quantitative value of landscape disturbance degree has significant spatial clustering characteristics. For landscape disturbances that are determined to have significant spatial clustering characteristics, the local Moran index of each grid is measured to obtain the local Moran index value and the corresponding local significance test value of each grid. Grids with negative local Moran index values ​​and local significance test values ​​less than a preset significance level threshold are selected and marked as spatially anomalous grids.

6. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 1, characterized in that, The process involves acquiring multidimensional geographic environmental factor data for the study area, performing spatial correlation analysis between this data and land use classification data, and identifying dominant environmental factors, including: Spatial overlay analysis was performed on multidimensional geographic environmental factor data and land use classification data to extract environmental factor values ​​corresponding to each land use type and generate environmental factor value sets. Geographically weighted principal component decomposition is performed on the environmental factor value set. The eigenvectors of the local covariance matrix are solved at each grid center point to generate the local factor loading coefficients at each spatial location. Spatial interpolation is performed on the local factor loading coefficients to generate a continuously distributed factor loading surface; On the factor loading surface, the loading coefficients of each environmental factor are compared pixel by pixel. The environmental factor with the largest loading coefficient at each pixel position is selected as the dominant environmental factor of that pixel, forming a set of dominant environmental factors for spatial variability.

7. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 6, characterized in that, The step of performing geographic weighted principal component decomposition on the environmental factor value set, solving for the eigenvectors of the local covariance matrix at each grid center point, and generating local factor loading coefficients at each spatial location includes: The environmental factor value set is standardized to generate a standardized environmental factor matrix. Construct a spatial weight matrix, and use the spatial weight matrix to weight the environmental factor values ​​for each grid center point to generate a locally weighted environmental factor matrix; Eigenvalue decomposition is performed on the locally weighted environmental factor matrix, and its principal component eigenvectors are extracted as the local factor loading coefficients of the grid center point.

8. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 1, characterized in that, The spatial distribution data of landscape vulnerability, generated based on the spatial correlation between dominant environmental factors and land use classification data, includes: Extract the numerical distribution range of dominant environmental factors across different land use types; The numerical distribution range is input into the preset ecological response coupling processor. The ecological response coupling processor has a built-in interaction matrix between land use type and environmental factors. The interaction matrix is ​​used to map the numerical distribution range and output the ecological response level of each land use type under different combinations of environmental factors. Fill in the ecological response level into the corresponding spatial unit in the land use classification data; The ecological response levels of all spatial units are aggregated to generate spatial distribution data of landscape vulnerability.

9. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 8, characterized in that, The ecological response coupling processor has a built-in interaction matrix between land use type and environmental factors. This interaction matrix maps the numerical distribution range and outputs the ecological response level of each land use type under different combinations of environmental factors, including: Read the pre-set interaction matrix between land use type and environmental factors. The interaction matrix records the sensitivity coefficient and tolerance threshold of each land use type to each dominant environmental factor. The dominant environmental factor values ​​of each spatial unit are compared with the tolerance threshold to determine the stress level of each environmental factor in that unit. Based on the sensitivity coefficients matched to the stress levels, the sensitivity coefficients of each environmental factor are fused to generate the ecological response level of the unit.

10. The method for analyzing the ecological risk of urban landscapes in the lower reaches of the Yellow River based on high-resolution data as described in claim 1, characterized in that, The process of fusing spatial distribution data of landscape disturbance degree with spatial distribution data of landscape vulnerability to generate a spatial distribution map of landscape ecological risk includes: The grid-by-grid method retrieves the quantitative values ​​of disturbance intensity from the spatial distribution data of landscape disturbance intensity and the ecological response levels from the spatial distribution data of landscape vulnerability. The quantified value of the disturbance degree is normalized and aligned with the ecological response level to generate aligned disturbance degree contribution value and vulnerability degree contribution value. The ecological risk index of the grid is generated by multiplying and fusing the disturbance contribution value and the vulnerability contribution value. The ecological risk indices of all grid cells are collected and arranged according to their spatial location to generate a spatial distribution map of landscape ecological risk.

Images