A method for extracting and visualizing three-dimensional dynamic characteristics of vegetation evolution based on remote sensing data

Abstract

The application discloses a vegetation evolution three-dimensional dynamic feature extraction and visualization method based on remote sensing data, and belongs to the technical field of remote sensing information processing and ecological environment monitoring. The method comprises the following steps: step 1, basic data acquisition and processing; step 2, horizontal gradient vegetation coverage spatial centroid migration calculation; step 3, vertical gradient vegetation evolution feature identification; and step 4, vegetation evolution three-dimensional dynamic feature extraction and visualization. The method disclosed by the application is based on multi-dimensional space fusion analysis of remote sensing vegetation index and digital elevation model data, can identify the horizontal migration characteristics and vertical gradient evolution characteristics of vegetation coverage, comprehensively consider the horizontal and vertical vegetation dynamic change characteristics, construct a vegetation three-dimensional curved surface model, efficiently and accurately extract and visualize the spatial structure and dynamic change trend of large-scale vegetation, and provide high-precision spatial analysis technical support for ecological environment monitoring, surface vegetation dynamic research and ecological restoration evaluation.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing information processing and ecological environment monitoring technology, and in particular relates to a method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data. Background Technology

[0002] As a crucial component of terrestrial ecosystems, vegetation's spatial distribution patterns and changes directly reflect the quality and dynamic trends of a region's ecological environment. The Normalized Difference Vegetation Index (NDVI) is a vital indicator that quantitatively reflects vegetation cover using remote sensing data, and it is widely used in vegetation monitoring, ecological assessment, and climate change research. Accurately and comprehensively identifying the continuous changes in vegetation in three-dimensional space is of great significance for revealing ecosystem structural characteristics and improving the level of ecological environment monitoring.

[0003] Existing vegetation feature extraction methods often focus on analyzing time-series trends or planar spatial distribution characteristics, neglecting the vertical constraints of topographic relief on vegetation growth and failing to reveal the three-dimensional features of vegetation spatial structure as a whole. Especially in complex terrain areas, vegetation distribution is often affected by multiple factors such as elevation, aspect, and slope, making it difficult to comprehensively express the continuous characteristics of vegetation three-dimensional spatial structure using only two-dimensional spatial analysis.

[0004] Therefore, there is an urgent need for a comprehensive analysis method that combines horizontal and vertical spatial dimensions, which can extract the spatial distribution and change characteristics of vegetation more efficiently and comprehensively, and realize continuous, three-dimensional visualization identification of vegetation spatial characteristics over a large area, providing a new technical path for dynamic monitoring of ecological patterns and assessment of environmental changes. Summary of the Invention

[0005] The purpose of this invention is to provide a method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data, so as to solve the above-mentioned technical problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention discloses a method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data. The method includes the following steps:

[0008] Step 1: Basic Data Acquisition and Processing: Acquire basic data for the study area, including remote sensing vegetation index (MODIS-NDVI) data and digital elevation model (DEM) data; and perform smoothing and noise reduction processing on the acquired vegetation index data, including data quality control, cloud removal, and invalid value removal. Then, average the monthly NDVI values ​​for the year pixel by pixel to obtain the annual average NDVI value.

[0009] Step 2, Calculation of spatial centroid migration of horizontal gradient vegetation cover: This includes the following steps:

[0010] Step 21: Vegetation cover gradient classification: The continuous NDVI data is discretized by the method of vegetation cover multi-level gradient classification, so as to identify the differences in the evolution characteristics of different vegetation cover in the region; specifically, the annual NDVI raster data are classified into different vegetation cover levels according to the NDVI value range, and the spatial range of different vegetation cover levels is obtained by converting the classified raster into area features.

[0011] Step 22: Extraction of Centroid Coordinates for the Region: Treating raster pixels as spatial points, weights are assigned to each pixel based on its NDVI value to calculate the centroid coordinates for regions with different coverage levels. This represents the central location of the vegetation space and further analyzes its migration characteristics. The formula for calculating the centroid coordinates is as follows:

[0012] (1)

[0013] (2)

[0014] In the formula: , The centroid coordinates of areas with different vegetation cover levels are given in degrees. , ) represents the pixel coordinates; C i Let be the value of the i-th pixel;

[0015] Step 23, Vegetation Spatial Migration and Trend Analysis: Based on the centroid coordinates of each vegetation cover level area calculated year by year, further calculate the direction and distance of the centroid migration vector to identify the interannual variation trend of vegetation in the horizontal direction.

[0016] Step 3, Identification of Vertical Gradient Vegetation Evolution Features: This specifically includes the following steps:

[0017] Step 31, Elevation gradient division: The DEM data of the study area is divided into elevation partition thresholds using the K-means clustering algorithm. The topography of the study area is divided into several gradient levels. The spatial range of each gradient is extracted by raster reclassification and surface feature transformation.

[0018] Step 32: Calculation of NDVI mean for each elevation gradient: Based on the spatial range of each elevation gradient, perform regional statistics on the annual NDVI raster data to obtain the time series of NDVI mean values ​​for different elevation gradients.

[0019] Step 33: Vegetation Change Trend Analysis at Different Elevation Gradients: The slope trend analysis method is used to calculate the interannual variation rate of NDVI of vegetation at each elevation gradient, and a significance test is performed; further, the average elevation of each pixel within each elevation gradient is calculated, and the sensitivity characteristics of vegetation to elevation changes under different elevation conditions are identified based on its correspondence with the NDVI change rate; the sensitivity coefficient is defined as:

[0020] (6)

[0021] In the formula: Let be the vegetation change sensitivity coefficient for elevation gradient i; Let be the interannual rate of change of NDVI for this gradient; This represents the average elevation of the gradient; and These represent the average NDVI change rate and elevation for the entire region, respectively. A value greater than 0 indicates that the rate of change of vegetation cover with increasing elevation increases, representing a positively sensitive response; when... When <0, it indicates that the rate of change of vegetation cover with increasing elevation decreases, which is a negative sensitive response;

[0022] Step 4: Extraction and visualization of three-dimensional dynamic features of vegetation evolution: This includes the following steps:

[0023] Step 41: Horizontal and Vertical Feature Curve Fitting: In the horizontal direction, based on the time series of centroid coordinates of each vegetation cover level area, a continuous function model is used to fit the centroid points of different levels of areas to obtain a set of distribution feature curves reflecting the spatial distribution and migration trend of vegetation; in the vertical direction, based on the elevation average value of each vegetation cover gradient and the multi-year average value of NDVI, a polynomial function fitting method is used to establish the functional relationship between elevation and NDVI to obtain the distribution feature curves of vegetation in the vertical direction.

[0024] Step 42, 3D Surface Construction and Visualization: Spatial fusion and coordinate mapping of feature curves in the horizontal and vertical directions; the horizontal feature curves provide planar coordinate constraints, and the vertical curves provide elevation and vegetation cover information, using elevation as the spatial height dimension and NDVI as the vegetation cover attribute value; constructing a 3D surface spatial feature model of vegetation through surface interpolation or data fitting algorithms to obtain the continuous change expression of NDVI at different spatial locations, realizing the extraction of 3D dynamic features of vegetation evolution; based on the generated 3D surface model, using 3D visualization software tools to perform spatial rendering of the 3D features of vegetation, using NDVI values ​​as color mapping variables to generate stereoscopic visualization results.

[0025] Furthermore, the data quality control mentioned in step 1 refers to judging the quality of pixel NDVI according to the quality control file accompanying the MODIS data product, selecting and retaining high-quality pixels (Pixel Reliability=0), and removing low-quality data affected by clouds and atmosphere (Pixel Reliability=1~3); the data declouding process refers to using the Savitzky-Golay filtering method to perform declouding processing and time series reconstruction on the NDVI data time series to achieve data smoothing and noise removal; the invalid value removal refers to removing data with NDVI values ​​outside the valid range (0~1) to ensure that the analysis object is an area with effective vegetation cover, and using the maximum value synthesis method to obtain the monthly maximum value to further remove outliers and enhance the effective vegetation signal.

[0026] Furthermore, the classification of annual NDVI raster data into different vegetation cover levels based on the NDVI value range in step 21 specifically involves classifying the annual NDVI raster data into five vegetation cover levels based on the NDVI value range: very low cover (0 < NDVI ≤ 0.1), low cover (0.1 < NDVI ≤ 0.3), medium cover (0.3 < NDVI ≤ 0.5), high cover (0.5 < NDVI ≤ 0.7), and very high cover (0.7 < NDVI ≤ 1.0).

[0027] Furthermore, step 23, which involves calculating the direction and distance of the centroid migration vector based on the centroid coordinates of each vegetation cover level area obtained annually, and identifying the interannual variation trend of vegetation in the horizontal direction, specifically includes the following process:

[0028] (1) Calculation of centroid migration rate and direction:

[0029] Based on the spatial offset relationship of the centroids of different vegetation cover levels in each year relative to the initial year, construct centroid migration vectors between consecutive years. And calculate the centroid migration rate and orientation angle:

[0030] (3)

[0031] (4)

[0032] (5)

[0033] in, Indicates the rate of interannual migration; This indicates the direction of migration, expressed in degrees (°), with due east as 0° and counterclockwise as positive. , This represents the centroid coordinates in year t, where t = 1 represents the initial centroid coordinates.

[0034] (2) Analysis of the centroid migration trend:

[0035] Construct time series of migration rate and orientation angle for each coverage level { }、{ The long-term slope of change is calculated using the trend test method. , And combine the Mann-Kendall significance test to determine significance: when When the vegetation cover is greater than 0, the vegetation cover expands to the northwest. If the value is greater than 0, the expansion trend intensifies. If the value is less than 0, the expansion trend weakens; When the vegetation cover is less than 0, the vegetation cover expands to the southeast. If the value is greater than 0, the expansion trend intensifies. If the value is less than 0, the expansion trend weakens.

[0036] Furthermore, the step 31, which involves dividing the topography of the study area into several gradient levels, specifically involves dividing the topography of the study area into five elevation levels: low, lower, medium, higher, and high.

[0037] The beneficial effects of this invention are as follows: The method described in this invention, based on the processing and multi-dimensional spatial fusion analysis of remote sensing vegetation index data and digital elevation model data, identifies the horizontal migration characteristics and vertical gradient evolution characteristics of vegetation cover. By integrating the dynamic change characteristics of horizontal and vertical vegetation, a three-dimensional surface model of the spatial characteristics of vegetation cover is constructed, achieving efficient and accurate extraction and visualization of the spatial structure and dynamic change trends of large-scale vegetation. Specifically, this is reflected in the following aspects:

[0038] (1) This invention improves the accuracy and reliability of basic data by denoising NDVI data; in the horizontal direction, based on the division of multi-level vegetation cover areas and extraction of spatial centroid coordinates, it quantitatively analyzes the spatial migration characteristics and change trends of different vegetation cover levels, and has the ability to efficiently extract the spatiotemporal distribution characteristics of vegetation in a large area; in the vertical direction, combined with DEM clustering and trend analysis methods, it identifies the differences and sensitivity response characteristics of vegetation changes under different elevation gradients, and reveals the systematic response mechanism of vegetation to topographic factors.

[0039] (2) This invention constructs a three-dimensional spatial distribution feature model of vegetation by spatial fusion of horizontal and vertical feature curves and surface modeling, realizing the continuous change expression of NDVI in planar position and elevation dimensions, breaking through the limitation of traditional two-dimensional analysis that is difficult to reflect the three-dimensional spatial distribution law;

[0040] (3) This invention can not only fully depict the dynamic changes of the horizontal space of regional vegetation, but also effectively identify the vertical evolution characteristics by combining topographic elevation information. It has technical advantages such as high analysis efficiency and quality, strong spatial feature expression and wide applicability in the identification of large-scale vegetation distribution characteristics. It can provide high-precision spatial analysis technical support for ecological environment monitoring, surface vegetation dynamic research and ecological restoration evaluation.

[0041] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the method flow described in Example 1;

[0043] Figure 2 This is a schematic diagram of the spatial centroid migration analysis results of horizontal vegetation cover in the Loess Plateau in Example 1.

[0044] Figure 3 This is a schematic diagram of the vertical variation characteristics of vegetation index in the Loess Plateau in Example 1.

[0045] Figure 4 This is a schematic diagram of the three-dimensional spatial feature extraction results of the vegetation index of the Loess Plateau in Example 1. Detailed Implementation

[0046] This invention discloses a method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data. The method includes the following steps:

[0047] Step 1: Basic Data Acquisition and Processing: Acquire basic data for the study area, including Remote Sensing Vegetation Index (MODIS-NDVI) data, Digital Elevation Model (DEM) data, etc.; and perform smoothing and noise reduction processing on the acquired NDVI data, including data quality control, cloud removal, and invalid value removal; then, average the monthly NDVI values ​​for the year pixel by pixel to obtain the annual average NDVI value.

[0048] Data quality control refers to assessing the NDVI quality of pixels based on the quality control documents accompanying MODIS data products, selecting and retaining high-quality pixels (Pixel Reliability=0, indicating high reliability), and removing low-quality data significantly affected by clouds and atmosphere (Pixel Reliability=1~3, indicating slight noise or atmospheric influence, snow / ice cover, and cloud interference, respectively). Cloud removal involves using the Savitzky-Golay (SG) filtering method to remove clouds and reconstruct the time series of NDVI data, achieving data smoothing and noise removal. Invalid value removal involves removing data with NDVI values ​​outside the valid range (0~1) to ensure the analyzed area represents effective vegetation cover. The Maximum Value Composite (MVC) method is then used to obtain the monthly maximum value to further remove outliers and enhance the effective vegetation signal.

[0049] Step 2, Calculation of spatial centroid migration of horizontal gradient vegetation cover: This includes the following steps:

[0050] Step 21: Vegetation Cover Gradient Delineation: To characterize the spatial differentiation features of vegetation in a large area in the horizontal direction, pixels within the region are classified and reclassified according to their NDVI values. By using a multi-level vegetation cover gradient delineation method, continuous NDVI data is discretized, thereby identifying the differences in vegetation cover evolution characteristics across regions and effectively improving the extraction efficiency of large-scale spatial vegetation feature information. For example, based on the NDVI value range, annual NDVI raster data is classified into five vegetation cover levels: very low cover (0 < NDVI ≤ 0.1), low cover (0.1 < NDVI ≤ 0.3), medium cover (0.3 < NDVI ≤ 0.5), high cover (0.5 < NDVI ≤ 0.7), and very high cover (0.7 < NDVI ≤ 1.0). By converting the classified raster into area features, the spatial range of different vegetation cover levels is obtained.

[0051] Gradient segmentation can adjust the grading threshold according to the spatial resolution of the research objective or remote sensing data, thereby maintaining the consistency of vegetation feature expression and analysis accuracy under different regional scales and topographic conditions.

[0052] Step 22: Extraction of Spatial Centroid Coordinates: Treating raster pixels as spatial centroids, weights are assigned to each pixel based on its NDVI value to calculate the spatial centroid coordinates of areas with different coverage levels. This represents the central location of vegetation space and facilitates further analysis of its migration characteristics. The resulting annual centroid sequence will serve as the foundational data for subsequent horizontal migration trend analysis and 3D spatial surface construction. The centroid coordinate calculation formula is as follows:

[0053] (1)

[0054] (2)

[0055] In the formula: , The centroid coordinates of areas with different vegetation cover levels are given in degrees (°). , ) represents the pixel coordinates; C i Let be the value of the i-th pixel.

[0056] Step 23, Vegetation Spatial Migration and Trend Analysis: Based on the centroid coordinates of each vegetation cover level area calculated year by year, the direction and distance of the centroid migration vector are further calculated to identify the interannual variation trend of vegetation in the horizontal direction. This specifically includes the following processes:

[0057] (1) Calculation of centroid migration rate and direction:

[0058] Based on the spatial offset relationship of the centroids of different vegetation cover levels in each year relative to the initial year, construct centroid migration vectors between consecutive years. And calculate the centroid migration rate and orientation angle:

[0059] (3)

[0060] (4)

[0061] (5)

[0062] In the formula: Indicates the rate of interannual migration; This indicates the direction of migration, expressed in degrees (°), with due east as 0° and counterclockwise as positive. , This represents the centroid coordinates in year t, where t = 1 represents the initial centroid coordinates.

[0063] (2) Analysis of the centroid migration trend:

[0064] Construct time series of migration rate and orientation angle for each coverage level { }、{ The long-term slope of change is calculated using trend testing methods (such as the Theil-Sen trend method). , And combine the Mann-Kendall (M-K) significance test to determine significance. When When the vegetation cover is greater than 0, the vegetation cover expands to the northwest. If the value is greater than 0, the expansion trend intensifies. If the value is less than 0, the expansion trend weakens; When the vegetation cover is less than 0, the vegetation cover expands to the southeast. If the value is greater than 0, the expansion trend intensifies. If the value is less than 0, the expansion trend weakens.

[0065] Step 3, Identification of Vertical Gradient Vegetation Evolution Features: This specifically includes the following steps:

[0066] Step 31: Elevation Gradient Delineation: The DEM data of the study area is divided into elevation zoning thresholds using the K-means clustering algorithm. The topography of the study area is divided into several representative gradient levels, such as five elevation grades (low, lower, medium, higher, and high). The spatial range of each gradient is extracted by raster reclassification and transformation of surface features to facilitate statistical analysis of vegetation changes within the same elevation interval.

[0067] Step 32: Calculation of NDVI mean for each elevation gradient: Based on the spatial range of each elevation gradient, perform regional statistics on the annual NDVI raster data to obtain the time series of NDVI mean values ​​for different elevation gradients.

[0068] Step 33: Vegetation Change Trend Analysis at Different Elevation Gradients: The slope trend analysis method is used to calculate the interannual variation rate of NDVI for vegetation at each elevation gradient, and significance tests are performed (t-test or P-test). Further, the average elevation of each pixel within each elevation gradient is calculated. Based on its correspondence with the NDVI change rate, the sensitivity characteristics of vegetation to elevation changes under different elevation conditions can be identified. The sensitivity coefficient is defined as:

[0069] (6)

[0070] In the formula: Let be the vegetation change sensitivity coefficient for elevation gradient i; Let be the interannual rate of change of NDVI for this gradient; This represents the average elevation of the gradient; and These represent the average NDVI change rate and elevation for the entire region, respectively. A value greater than 0 indicates that the rate of change of vegetation cover with increasing elevation increases, representing a positively sensitive response; when... When <0, it indicates that the rate of change of vegetation cover with increasing elevation decreases, which is a negative sensitive response.

[0071] Step 4: Extraction and visualization of three-dimensional dynamic features of vegetation evolution:

[0072] Step 41: Fitting Horizontal and Vertical Characteristic Curves: In the horizontal direction, based on the time series of centroid coordinates of each vegetation cover level region, a continuous function model (such as spline function or least squares regression) is used to fit the centroid points of different levels of regions to obtain a set of distribution characteristic curves reflecting the spatial distribution and migration trend of vegetation; in the vertical direction, based on the elevation average value of each vegetation cover gradient and the multi-year average value of NDVI, a polynomial function fitting method is used to establish the functional relationship between elevation and NDVI to obtain the distribution characteristic curves of vegetation in the vertical direction.

[0073] Step 42, 3D Surface Construction and Visualization: Spatial fusion and coordinate mapping of the horizontal and vertical feature curves. The horizontal feature curves provide planar coordinate constraints (x, y), and the vertical curves provide elevation and vegetation cover information (H, NDVI), with elevation H as the spatial height dimension (Z-axis) and NDVI as the vegetation cover attribute value. A 3D surface spatial feature model of vegetation is constructed using surface interpolation or data fitting algorithms (such as bicubic spline interpolation or inverse distance weighting), obtaining the continuous variation expression of NDVI at different spatial locations, thus realizing the extraction of the 3D dynamic features of vegetation evolution.

[0074] Based on the generated 3D surface model, the 3D visualization features of vegetation are further visualized using 3D visualization software tools (such as Origin or MATLAB). The NDVI value is used as a color mapping variable to generate a stereoscopic visualization result, thereby intuitively presenting a stereoscopic expression of the distribution relationship of vegetation in horizontal and vertical space, reflecting the spatial structural characteristics of vegetation and helping to identify its potential change trends.

[0075] Example 1

[0076] This embodiment is an application example of the above method.

[0077] This embodiment discloses a method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data, such as... Figure 1 As shown, the method includes the following steps:

[0078] Step 1: Basic Data Acquisition and Processing: This embodiment takes the Loess Plateau as the study area and acquires basic data of the study area, including MODIS-NDVI data, DEM data, etc.; and performs smoothing and noise reduction processing on the acquired NDVI data, including data quality control, data cloud removal, invalid value removal, etc. Then, the monthly NDVI values ​​of the processed year are averaged pixel by pixel to obtain the annual average NDVI value.

[0079] The NDVI data used in this embodiment comes from MODIS' MOD13Q1 v061 product, with a spatial resolution of 250m × 250m and a time series from 2001 to 2021. The original NDVI and its quality control data were stitched, converted in format, and reprojected using the MRT tool provided by MODIS. The processed data was then cropped to the study area, completing the quality control processing of the NDVI data.

[0080] The Savitzky-Golay filtering method (SG filtering method) was used to reconstruct and smooth the NDVI time series data. Based on the data characteristics of the study area and the debugging results, the filter window size was set to 5 and the degree of the fitted polynomial was set to 3 to remove noise caused by cloud cover and atmospheric influence.

[0081] Based on the range of pixel NDVI values, invalid data with NDVI≤0 are removed, and the maximum value synthesis method (MVC) is used to synthesize the 16-day resolution data into the monthly maximum value. Then, the average value of each month's data in the same year is calculated pixel by pixel to obtain the annual average NDVI value.

[0082] Step 2: Calculation of spatial centroid migration of horizontal gradient vegetation cover:

[0083] Step 21: Vegetation cover gradient classification: Using ArcGIS's Reclassify tool, the annual NDVI raster data are classified into five vegetation cover levels based on the NDVI value range: very low cover (0 < NDVI ≤ 0.1), low cover (0.1 < NDVI ≤ 0.3), medium cover (0.3 < NDVI ≤ 0.5), high cover (0.5 < NDVI ≤ 0.7), and very high cover (0.7 < NDVI ≤ 1.0). The classified raster is then converted into polygon features using the Raster to Polygon tool to obtain the spatial range of different vegetation cover levels.

[0084] Step 22: Extraction of Spatial Centroid Coordinates: The NDVI raster data for the study period is converted into point data. Combining the pixel values ​​and the corresponding spatial range of vegetation cover levels, the spatial centroid coordinates for each cover level are calculated. For example, taking every five years from 2001 to 2021 as an example, the horizontal vegetation cover area centroid migration characteristics are as follows: Figure 2 As shown.

[0085] Step 23, Vegetation Spatial Migration and Trend Analysis: Based on the centroid coordinates of each vegetation cover level area calculated year by year, further calculate the direction and distance of the centroid migration vector to identify the interannual variation trend of vegetation in the horizontal direction.

[0086] Step 3: Identification of vertical gradient vegetation evolution characteristics:

[0087] Step 31, Elevation gradient division: Based on the K-means clustering algorithm, the Loess Plateau DEM data is divided into five elevation levels (low, lower, medium, higher, and high), and the ArcGIS Reclassify tool is used to classify the raster data to obtain the spatial range of the corresponding elevation gradient.

[0088] Step 32: Calculate the mean NDVI of each elevation gradient: Using ArcGIS regional statistical analysis tools, calculate the mean NDVI of each year within the spatial range of different elevation gradients.

[0089] Step 33: Vegetation Change Trend Analysis at Different Elevation Gradients: The slope trend analysis method was used to calculate the interannual variation trend of NDVI for vegetation at each elevation gradient, and the significance was tested using the p-value. Interannual variation curves of NDVI at different gradients were plotted, as shown below. Figure 3 As shown.

[0090] Step 4: Extraction and visualization of three-dimensional dynamic features of vegetation evolution:

[0091] Step 41: Horizontal and Vertical Feature Curve Fitting: Based on the calculation results of the centroid coordinates of the horizontal gradient space, spline interpolation and least squares methods are used to fit curves of the centroid points of different vegetation cover levels to obtain the horizontal feature curves of vegetation evolution. Simultaneously, ArcGIS regional statistical analysis tools are used to calculate the spatial average elevation of different horizontal gradients and fit it to the NDVI value to obtain the vertical feature curves.

[0092] Step 42, 3D Surface Construction and Visualization: Combine the horizontal and vertical fitted curves, use multinomial regression to obtain the overall trend, and then combine this with a three-cubic weighted kernel (1-d) 3 ) 3 The local smoothing method refines the data through a grid interpolation algorithm (such as bicubic spline interpolation) to generate a three-dimensional surface model. The surface coordinates consist of horizontally fitted x and y values ​​and vertically fitted elevation values, achieving a three-dimensional representation of the spatial distribution of vegetation. Based on this, the three-dimensional visualization software Origin is used to create three-dimensional graphics depicting the distribution of different vegetation cover levels in the horizontal and vertical spaces of the study area. Figure 4 As shown, this allows for the identification of the horizontal and vertical three-dimensional spatial distribution characteristics of vegetation in the study area.

[0093] Finally, it should be noted that the above description is only used to illustrate the technical solution of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred arrangement, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data, characterized in that, The method includes the following steps: Step 1: Basic Data Acquisition and Processing: Acquire basic data for the study area, including remote sensing vegetation index (MODIS-NDVI) data and digital elevation model (DEM) data; and perform smoothing and noise reduction processing on the acquired vegetation index data, including data quality control, cloud removal, and invalid value removal. Then, average the monthly NDVI values ​​for the year pixel by pixel to obtain the annual average NDVI value. Step 2, Calculation of spatial centroid migration of horizontal gradient vegetation cover: This includes the following steps: Step 21: Vegetation cover gradient classification: The continuous NDVI data is discretized by the method of vegetation cover multi-level gradient classification, so as to identify the differences in the evolution characteristics of different vegetation cover in the region; specifically, the annual NDVI raster data are classified into different vegetation cover levels according to the NDVI value range, and the spatial range of different vegetation cover levels is obtained by converting the classified raster into area features. Step 22: Extraction of Centroid Coordinates for the Region: Treating raster pixels as spatial points, weights are assigned to each pixel based on its NDVI value to calculate the centroid coordinates for regions with different coverage levels. This represents the central location of the vegetation space and further analyzes its migration characteristics. The formula for calculating the centroid coordinates is as follows: （1） （2） In the formula: , The centroid coordinates of areas with different vegetation cover levels are given in degrees. , ) represents the pixel coordinates; C i Let be the value of the i-th pixel; Step 23, Vegetation Spatial Migration and Trend Analysis: Based on the centroid coordinates of each vegetation cover level area calculated year by year, further calculate the direction and distance of the centroid migration vector to identify the interannual variation trend of vegetation in the horizontal direction. Step 3, Identification of Vertical Gradient Vegetation Evolution Features: This specifically includes the following steps: Step 31, Elevation gradient division: The DEM data of the study area is divided into elevation partition thresholds using the K-means clustering algorithm. The topography of the study area is divided into several gradient levels. The spatial range of each gradient is extracted by raster reclassification and surface feature transformation. Step 32: Calculation of NDVI mean for each elevation gradient: Based on the spatial range of each elevation gradient, perform regional statistics on the annual NDVI raster data to obtain the time series of NDVI mean values ​​for different elevation gradients. Step 33: Vegetation Change Trend Analysis at Different Elevation Gradients: The slope trend analysis method is used to calculate the interannual variation rate of NDVI of vegetation at each elevation gradient, and a significance test is performed; further, the average elevation of each pixel within each elevation gradient is calculated, and the sensitivity characteristics of vegetation to elevation changes under different elevation conditions are identified based on its correspondence with the NDVI change rate; the sensitivity coefficient is defined as: （6） In the formula: Let be the vegetation change sensitivity coefficient for elevation gradient i; Let be the interannual rate of change of NDVI for this gradient; This represents the average elevation of the gradient; and These are the average values ​​of the NDVI change rate and elevation for the entire region, respectively; when A value greater than 0 indicates that the rate of change of vegetation cover with increasing elevation increases, representing a positively sensitive response; when... When <0, it indicates that the rate of change of vegetation cover with increasing elevation decreases, which is a negative sensitive response; Step 4: Extraction and visualization of three-dimensional dynamic features of vegetation evolution: This includes the following steps: Step 41: Horizontal and Vertical Feature Curve Fitting: In the horizontal direction, based on the time series of centroid coordinates of each vegetation cover level area, a continuous function model is used to fit the centroid points of different levels of areas to obtain a set of distribution feature curves reflecting the spatial distribution and migration trend of vegetation; in the vertical direction, based on the elevation average value of each vegetation cover gradient and the multi-year average value of NDVI, a polynomial function fitting method is used to establish the functional relationship between elevation and NDVI to obtain the distribution feature curves of vegetation in the vertical direction. Step 42, 3D Surface Construction and Visualization: Spatial fusion and coordinate mapping of feature curves in the horizontal and vertical directions; the horizontal feature curves provide planar coordinate constraints, and the vertical curves provide elevation and vegetation cover information, using elevation as the spatial height dimension and NDVI as the vegetation cover attribute value; constructing a 3D surface spatial feature model of vegetation through surface interpolation or data fitting algorithms to obtain the continuous change expression of NDVI at different spatial locations, realizing the extraction of 3D dynamic features of vegetation evolution; based on the generated 3D surface model, using 3D visualization software tools to perform spatial rendering of the 3D features of vegetation, using NDVI values ​​as color mapping variables to generate stereoscopic visualization results.

2. The method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data according to claim 1, characterized in that, The data quality control mentioned in step 1 refers to judging the quality of pixel NDVI according to the quality control document accompanying the MODIS data product, selecting and retaining high-quality pixels (Pixel Reliability=0), and removing low-quality data affected by clouds and atmosphere (Pixel Reliability=1~3); the data declouding process refers to using the Savitzky-Golay filtering method to perform declouding processing and time series reconstruction on the NDVI data time series to achieve data smoothing and noise removal; the invalid value removal refers to removing data with NDVI values ​​outside the valid range (0~1) to ensure that the analysis object is an area with effective vegetation cover, and using the maximum value synthesis method to obtain the monthly maximum value to further remove outliers and enhance the effective vegetation signal.

3. The method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data according to claim 1, characterized in that, The step 21, which describes classifying the annual NDVI raster data into different vegetation cover levels based on the NDVI value range, specifically classifies the annual NDVI raster data into five vegetation cover levels based on the NDVI value range: very low cover (0 < NDVI ≤ 0.1), low cover (0.1 < NDVI ≤ 0.3), medium cover (0.3 < NDVI ≤ 0.5), high cover (0.5 < NDVI ≤ 0.7), and very high cover (0.7 < NDVI ≤ 1.0).

4. The method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data according to claim 1, characterized in that, Step 23 involves calculating the centroid coordinates of each vegetation cover level area annually, then further calculating the direction and distance of the centroid migration vector to identify the specific interannual variation trend of vegetation in the horizontal direction. The process includes the following: (1) Calculation of centroid migration rate and direction: Based on the spatial offset relationship of the centroids of different vegetation cover levels in each year relative to the initial year, construct centroid migration vectors between consecutive years. And calculate the centroid migration rate and orientation angle: （3） （4） （5） in, Indicates the rate of interannual migration; This indicates the direction of migration, expressed in degrees (°), with due east as 0° and counterclockwise as positive. , This represents the centroid coordinates in year t, where t = 1 represents the initial centroid coordinates. (2) Analysis of the centroid migration trend: Construct time series of migration rate and orientation angle for each coverage level { }、{ The long-term slope of change is calculated using the trend test method. , And combine the Mann-Kendall significance test to determine significance: when When the vegetation cover is greater than 0, the vegetation cover expands to the northwest. If the value is greater than 0, the expansion trend intensifies. If the value is less than 0, the expansion trend weakens; When the vegetation cover is less than 0, the vegetation cover expands to the southeast. If the value is greater than 0, the expansion trend intensifies. If the value is less than 0, the expansion trend weakens.

5. The method for extracting and visualizing three-dimensional dynamic features of vegetation evolution based on remote sensing data according to claim 1, characterized in that, Step 31, which describes dividing the topography of the study area into several gradient levels, specifically involves dividing the topography of the study area into five elevation levels: low, lower, medium, higher, and high.

Images