Improved algorithm for relative contribution of human activity to long time series vegetation index

Abstract

The application provides a long time series vegetation index relative contribution rate improvement algorithm of human activities, improves the resolution accuracy of long time series NDVI by comprehensively utilizing multi-source remote sensing and space-time data reconstruction modeling, and realizes the function of accurately reflecting the change process of long time series NDVI value of human activities. The application removes the common trend effect of climate factors and human activities by using the detrend method, comprehensively utilizes the advantages of various influencing factors of climate (air temperature, precipitation, solar radiation), soil, topography and land use type, and overcomes the disadvantages of analyzing the influence of human activities on vegetation index by a single climate factor. The application improves the accuracy of the long time series vegetation index relative contribution rate algorithm of human activities, provides a new algorithm for the influence of human activities on vegetation coverage, and provides a theoretical reference for the ecological protection of vegetation in the ecological environment.

Description

Technical Field

[0001] This invention belongs to the field of vegetation remote sensing technology, specifically involving an improved algorithm for the relative contribution rate of human activities to long-term vegetation indices. Background Technology

[0002] Vegetation plays a vital regulatory role in global ecosystems by influencing energy exchange between the Earth's surface and the atmosphere. Currently, monitoring and attributing large-scale, long-term vegetation changes has become an important aspect of global change research. Using satellite remote sensing to monitor the Normalized Difference Vegetation Index (NDVI) is a widely used method for monitoring vegetation dynamics.

[0003] Current research on the impact of human activities on long-term vegetation indices often employs linear regression with temperature and precipitation as independent variables and NDVI as the dependent variable. The simulated NDVI value is calculated using the regression coefficients determined by the regression model and the original temperature and precipitation data. However, vegetation cover is influenced by more than just climatic factors like temperature and precipitation. Environmental factors such as solar radiation, soil, topography, and land use also comprehensively affect the NDVI result. This is especially true when studying vegetation changes in high-altitude, cold mountainous regions with strong spatiotemporal heterogeneity. Linear regression based solely on climatic variables like temperature and precipitation often fails to accurately reflect the true impact of human activities on long-term NDVI values. Furthermore, when climatic factors and human activities exhibit a co-trend, using the conventional regression residual trend method can confuse the two, underestimating the contribution of human activities and overestimating the role of climatic factors. Therefore, it is essential to first eliminate the co-trend effect between climatic factors and human activities. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide an improved algorithm for the relative contribution rate of human activities to long-term vegetation index, which can accurately reflect the change process of human activities on long-term NDVI values.

[0005] The technical solution adopted by this invention to solve the above-mentioned technical problems is: an improved algorithm for the relative contribution rate of human activities to long-term vegetation indices, comprising the following steps:

[0006] S1: Collect long-term NDVI remote sensing images of the target area as the data source and preprocess them. Use data fusion algorithms to reconstruct the remote sensing image data to obtain the remote sensing observation NDVI values, and fuse them into a long-term NDVI dataset with the accuracy required by the preset requirements.

[0007] S2: Collect climate data and other environmental factors for the target area, and perform detrending processing on the climate data and NDVI data;

[0008] S3: Using multiple environmental factors, including climate data and other environmental elements, as independent variables, and the fused high-precision long-term NDVI data as dependent variables, we use a linear regression model and the least squares method to perform trend analysis to obtain the time-series predicted NDVI value.

[0009] S4: Subtract the predicted NDVI value of the time series calculated in step S3 from the remote sensing observation NDVI value obtained in step S1 to obtain the NDVI residual value, which is used to represent the impact of human activities on the long-term NDVI.

[0010] S5: Calculate the NDVI residual trend rate based on the residual values ​​obtained in step S4, and calculate the relative contribution rate H of human activities to the long-term series NDVI using a combination of scenario simulation and evaluation.

[0011] According to the above scheme, in step S1, the long-term remote sensing images include SPOT NDVI remote sensing images and GIMMS NDVI remote sensing images; the SPOT NDVI remote sensing images are SPOT-4 product data with a spatial resolution of 1 km and a temporal resolution of 10 days, covering the period from 1999 to 2013; the GIMMS NDVI remote sensing images are GIMMS NDVI3g product data with a spatial resolution of 8 km and a temporal resolution of 15 days, covering the period from 1982 to 2015.

[0012] Furthermore, step S1 further includes the following step:

[0013] The NDVI data was extracted and clipped using a mask, and the annual maximum NDVI value of the study area was calculated by the maximum value synthesis method to form the NDVI time series dataset of the study area, which was then projected onto the WGS84 coordinate system.

[0014] Furthermore, in step S1, the preprocessing includes data stitching, projection transformation, and maximum value synthesis processing of the remote sensing images; the temporal resolution of the preprocessed SPOT NDVI remote sensing images and GIMMS NDVI remote sensing images is both monthly, while the spatial resolution and temporal coverage remain unchanged.

[0015] In the overlapping remote sensing images, the pixel values ​​of the 1 km spatial resolution SPOT NDVI remote sensing images from 1999 to 2013 were used as the prediction domain, and the pixel values ​​of the 8 km spatial resolution GIMMS NDVI remote sensing images from 1982 to 2015 were used as the response domain.

[0016] By utilizing the spatial distribution characteristics of pixel values ​​in the prediction domain and the response domain, an empirical orthogonal telecorrelation model between pixel values ​​in the prediction domain and the response domain is established to perform spatiotemporal fusion of SPOT NDVI remote sensing images and GIMMS NDVI remote sensing images.

[0017] Spatial spline interpolation was performed on the pixel values ​​of GIMMS NDVI images with a spatial resolution of 8 km from 1982 to 2002 to obtain NDVI pixel values ​​with a spatial resolution of 1 km from 1982 to 2015, which were then projected onto the WGS84 coordinate system.

[0018] According to the above scheme, in step S2, the climate data includes temperature, precipitation and solar radiation; other environmental factors include soil moisture, topography and land use type data.

[0019] The time coverage of remote sensing images of climate data and other environmental elements is from 1982 to 2015. The preprocessing of climate data includes weather station screening, missing data imputation and spatial data interpolation. It also includes summing the daily average temperatures of weather stations throughout the year and dividing by the length of the year to obtain the annual average temperature, and summing the cumulative precipitation within 24 hours of each day throughout the year to obtain the total annual precipitation.

[0020] The solar radiation dataset is obtained by calculating the annual average solar radiation of meteorological stations by performing an arithmetic average of monthly average solar radiation data.

[0021] Soil moisture, topography, and land use type data were extracted and clipped using a mask and then reprojected onto the WGS84 coordinate system.

[0022] Furthermore, in step S2, the detrending method is to perform univariate linear regression simulations on the factors of temperature, precipitation and solar radiation for each year, and then add the mean of the corresponding variables to the difference; the NDVI, temperature, precipitation and total solar radiation after detrending are denoted as D(NDVI), D(T), D(P) and D(S) respectively.

[0023] Furthermore, in step S3, the specific steps are as follows:

[0024] The independent variables include temperature (D(T)), precipitation (D(P)), total solar radiation (D(S)), soil moisture (SM), topography (DEM), and land use type (LU). A multiple linear regression model is established to calculate the NDVI predicted value for each parameter in the model, including annual average temperature, precipitation, total solar radiation, soil moisture, topography, and land use type. CC This is used to predict the NDVI value of vegetation; the specific calculation formula is as follows:

[0025] NDVI CC=β0+β1D(T)+β2D(P)+β3D(S)+β4SM+β5DEM+β6LU;

[0026] Where β0 is a constant term, β1 is the regression coefficient between NDVI and air temperature D(T), β2 is the regression coefficient between NDVI and precipitation P, β3 is the regression coefficient between NDVI and total solar radiation S, β4 is the regression coefficient between NDVI and soil moisture SM, β5 is the regression coefficient between NDVI and topography DEM, β6 is the regression coefficient between NDVI and land use type LU, and β0 is the regression constant term. The undetermined coefficients are calculated using the least squares method.

[0027] Furthermore, in step S4, the specific steps are as follows:

[0028] Calculate NDVI observations NDVI obs NDVI Predicted Value CC The difference is used to obtain the NDVI residual value. HA This is used to represent the impact of human activities on long-term NDVI series; the specific calculation formula is as follows:

[0029] NDVI HA =NDVI obs -NDVI CC ;

[0030] If residual NDVI HA >0 indicates that human activities have a promoting effect on vegetation growth;

[0031] If residual NDVI HA <0 indicates that human activities are detrimental to vegetation growth;

[0032] If residual NDVI HA =0 indicates that human activities have a weak impact on vegetation change.

[0033] Furthermore, in step S5, the specific steps are as follows:

[0034] For NDVI respectively HA and NDVI CC Linear regression was performed between the time series and the year to obtain the NDVI change trends β(NDVIres) and β(NDVIpre) under the influence of human activities and climate change alone, as shown in the following formulas:

[0035]

[0036] In the formula, β represents the trend of residual NDVI, i and j are time series, and NDVI i NDVI jThese represent the NDVI values ​​at time i and j, respectively.

[0037] If β>0, it indicates that human activities or changes in climate and other environmental factors have a promoting effect on vegetation growth.

[0038] If β < 0, it indicates that human activities or changes in climate and other environmental factors have an inhibitory effect on vegetation growth;

[0039] If β = 0, it indicates that human activities have a weak impact on vegetation change.

[0040] Furthermore, the following steps are also included: The contribution rate H of human activities to long-term NDVI is obtained using a combination of residual analysis and scenario simulation assessment.

[0041] Scenario 1: When S(NDVI) obs ) rise, NDVI CC >0, NDVI HA >0, the contribution rate of climate, soil, topography and land use type change is Human activity contribution rate This indicates that changes in climate and other environmental factors, as well as human activities, have led to an increase in vegetation.

[0042] Scenario 2: When S(NDVI) obs ) rise, NDVI CC >0, NDVI HA <0, the contribution rate of climate, soil, topography and land use type changes is 100%, and the contribution rate of human activities is 0, indicating that changes in climate and other environmental factors have caused vegetation increase;

[0043] Scenario 3: When S(NDVI) obs ) rise, NDVI CC <0, NDVI HA >0, the contribution rate of climate, soil, topography and land use type change is 0, and the contribution rate of human activities is 100%, indicating that human activities have caused vegetation increase;

[0044] Scenario 4: When S(NDVI) obs ) decrease, NDVI CC <0, NDVI HA <0, the contribution rate of climate, soil, topography and land use type change is Human activity contribution rate This indicates that changes in climate and other environmental factors, as well as human activities, have caused vegetation degradation.

[0045] Scenario 5: When S(NDVI) obs ) decrease, NDVI CC <0, NDVI HA>0, the contribution rate of climate, soil, topography and land use type change is 100%, and the contribution rate of human activities is 0, indicating that changes in climate and other environmental factors have caused vegetation degradation;

[0046] Scenario 6: When S(NDVI) obs ) decrease, NDVI CC >0, NDVI HA <0, the contribution rate of climate, soil, topography and land use type change is 0, and the contribution rate of human activities is 100%, indicating that human activities cause vegetation degradation.

[0047] The beneficial effects of this invention are as follows:

[0048] 1. The improved algorithm for the relative contribution rate of human activities to long-term vegetation index in this invention improves the resolution accuracy of long-term NDVI by comprehensively utilizing multi-source remote sensing and spatiotemporal data for reconstruction modeling, and realizes the function of accurately reflecting the change process of human activities on long-term NDVI values.

[0049] 2. This invention uses a detrending method to remove the co-trend effect of climate factors and human activities, and integrates the advantages of various influencing factors such as climate (temperature, precipitation, solar radiation), soil, topography, and land use type, overcoming the drawbacks of analyzing the impact of human activities on vegetation index through a single climate factor.

[0050] 3. This invention improves the accuracy of the algorithm for the relative contribution rate of human activities to long-term vegetation index, provides a new algorithm for the impact of human activities on vegetation cover, and provides a theoretical reference for the ecological protection of vegetation in the ecological environment. Attached Figure Description

[0051] Figure 1 This is a flowchart of an embodiment of the present invention.

[0052] Figure 2 These are NDVI spatial distribution diagrams before and after accuracy improvement on the XX plateau, and a difference diagram before and after accuracy improvement, according to embodiments of the present invention.

[0053] Figure 3 This is a comparison chart of residual analysis trends before and after the improvement of the XX plateau in this invention.

[0054] Figure 4 This is a graph showing the relative contribution rate of human activities to NDVI before the improvement of the XX Plateau in this invention.

[0055] Figure 5 This is a graph showing the relative contribution rate of human activities to NDVI after the improvement of the XX Plateau in this invention. Detailed Implementation

[0056] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0057] See Figures 2 to 5 This study selected the XX Plateau as the research area, reconstructed the NDVI vegetation index at a spatial resolution of 1 km in the XX Plateau region, and analyzed the impact of human activities on the NDVI vegetation index in this region. (See also...) Figure 1 The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices in this invention includes the following steps:

[0058] S1: Using long-term SPOT NDVI and GIMMS NDVI remote sensing images as data sources, the remote sensing image data are reconstructed using data fusion algorithms to obtain remote sensing observation NDVI values, which are then fused into a high-precision long-term NDVI dataset.

[0059] The SPOT remote sensing imagery is SPOT-4 product data with a spatial resolution of 1 km and a temporal resolution of 10 days, covering the period from 1999 to 2013; the GIMMS remote sensing imagery is GIMMS NDVI3g product data with a spatial resolution of 8 km and a temporal resolution of 15 days, covering the period from 1982 to 2015.

[0060] NDVI data of the target area were collected and preprocessed. The remote package in R was used to reconstruct the data and merge it into a high-precision long-term NDVI dataset of 1km. The NDVI data was cropped using mask extraction in ArcGIS, and the annual maximum NDVI value of the study area was calculated by the maximum value synthesis method to form the NDVI time series dataset of the study area, which was then projected onto the WGS84 coordinate system.

[0061] The preprocessing includes remote sensing image data stitching, projection conversion, maximum value synthesis, etc. The temporal resolution of the preprocessed SPOT and GIMMS remote sensing images is monthly, while the spatial resolution and temporal coverage remain unchanged. In the overlapping remote sensing images, the pixel values ​​of the 1 km spatial resolution SPOT NDVI image from 1999 to 2013 were used as the prediction domain, and the pixel values ​​of the 8 km spatial resolution GIMMS NDVI image from 1982 to 2015 were located in the response domain. Then, using the spatial distribution characteristics of the pixel values ​​in the prediction domain and the response domain, an empirical orthogonal teleconnection model was established between the pixel values ​​in the prediction domain and the response domain to perform spatiotemporal fusion of the SPOT NDVI and GIMMS NDVI remote sensing images. Finally, spatial spline interpolation was performed on the pixel values ​​of the 8 km spatial resolution GIMMS NDVI image from 1982 to 2002 to obtain the NDVI pixel values ​​with a spatial resolution of 1 km from 1982 to 2015, which were then projected onto the WGS84 coordinate system.

[0062] S2: Collect climate data and other environmental factors for the target area, such as climate data (temperature, precipitation, solar radiation), soil moisture, topography, and land use type data, and perform detrending processing on the climate data and NDVI.

[0063] The time coverage of remote sensing images of climate data and other environmental elements from meteorological stations is from 1982 to 2015. Preprocessing of climate data includes meteorological station screening, missing data imputation, and spatial data interpolation. The annual average temperature is obtained by summing the daily average temperature of meteorological stations throughout the year and dividing by the year. The annual total precipitation is obtained by summing the cumulative precipitation over 24 hours each day throughout the year. The solar radiation dataset is obtained by calculating the annual average solar radiation of meteorological stations by arithmetic averaging the monthly average solar radiation. Soil moisture, topography, and land use data are cropped using mask extraction in ArcGIS and reprojected onto the WGS84 coordinate system.

[0064] The detrending method involves performing univariate linear regression simulations on temperature, precipitation, and solar radiation factor for each year, and then adding the mean of the variable to the difference. The detrended NDVI, temperature, precipitation, and total solar radiation are denoted as D(NDVI), D(T), D(P), and D(S).

[0065] S3: Using multiple environmental factors, including climate data and other environmental elements, as independent variables, and the fused high-precision long-term NDVI data as dependent variables, the time-series predicted NDVI values ​​are obtained by using a linear regression model and the least squares method for trend analysis.

[0066] Using high-precision NDVI as the dependent variable and temperature D(T), precipitation D(P), total solar radiation D(S), soil moisture (SM), topography (DEM), and land use type (LU) as independent variables, a multiple linear regression model was established. The predicted NDVI value was obtained by calculating the parameters in the model. CC This is used to predict the NDVI value of vegetation; the specific calculation formula is as follows:

[0067] NDVI CC =β0+β1D(T)+β2D(P)+β3D(S)+β4SM+β5DM+β6LU

[0068] In the formula, NDVI CC The NDVI values ​​are predicted for factors such as annual average temperature, precipitation, solar radiation, soil moisture, topography, and land use type. β0 is a constant term, β1 is the regression coefficient between NDVI and temperature T, β2 is the regression coefficient between NDVI and precipitation P, β3 is the regression coefficient between NDVI and solar radiation S, β4 is the regression coefficient between NDVI and soil moisture SM, β5 is the regression coefficient between NDVI and topography DEM, and β6 is the regression coefficient between NDVI and land use LU. β0 is the regression constant term, and the undetermined coefficients are calculated using the least squares method.

[0069] S4: The residual value is obtained by subtracting the time series predicted NDVI value calculated in step S3 from the remote sensing observation NDVI value obtained in step S1. This residual value is used to represent the impact of human activities on vegetation NDVI.

[0070] Calculate NDVI observations NDVI obs With NDVI CC The difference between them is the NDVI residual (NDVI) HA The residual is used to represent the impact of human activities on vegetation NDVI. If the residual is greater than 0, it indicates that human activities promote vegetation growth; if the residual is less than 0, it indicates that human activities are detrimental to vegetation growth; if the residual is equal to 0, it indicates that human activities have a weak impact on vegetation change. The specific calculation formula is as follows:

[0071] NDVI HA =NDVI obs -NDVI CC

[0072] NDVI HA The actual value of vegetation cover (NDVI) obs Compared with the predicted value NDVI CC The residual between; when NDVI HA A value greater than 0 indicates that human activities have a positive impact on the vegetation index NDVI; when NDVI > 0, it indicates that human activities have a positive impact on the vegetation index NDVI. HAWhen NDVI < 0, it indicates that human activities have a negative impact; when NDVI < 0, it indicates that human activities have a negative impact. HA When the value is 0, it indicates that the impact of human activities is weak.

[0073] S5: Calculate the residual trend rate of NDVI based on the residual values ​​obtained in step S4, and calculate the relative contribution rate H of human activities to the long-term vegetation index using a combination of scenario simulation and evaluation.

[0074] The NDVI calculated based on the formula in step S4 HA and NDVI CC NDVI HA and NDVI CC Linear regression was performed between the time series and the year to obtain the NDVI change trends β(NDVIres) and β(NDVIpre) under the influence of human activities and climate change alone, as shown in the following formulas:

[0075]

[0076] In the formula, β represents the trend of residual NDVI, i and j are time series, and NDVI i NDVI j Let β and β represent the NDVI values ​​at time i and j, respectively. If β > 0, the residual NDVI trend is positive, indicating that human activities or climate change have a promoting effect on vegetation growth; if β < 0, the trend is negative, indicating that human activities or climate change have an inhibitory effect on vegetation growth; if β = 0, it indicates that the residual NDVI sequence has no trend.

[0077] The contribution rate H of human activities to vegetation NDVI was obtained by combining residual analysis and scenario simulation assessment. The relative contribution rate of human activities to vegetation indices was calculated as follows:

[0078] Table 1. Calculation Methods for the Contribution Rate of Environmental Factors such as Climate and Human Activities to Vegetation NDVI Change

[0079]

[0080]

[0081] Referring to Table 2, the correlation between vegetation NDVI values ​​and climate factors in areas such as the southern part of the XX Plateau is poor. The relative contribution rate H identified before the improvement was between 10% and 50%, while the H identified by the improved algorithm was greater than 50%. This indicates that the improved residual trend method effectively avoids the biases present in the residual trend method and effectively identifies the role of human activities.

[0082] Table 2. Percentage of relative contribution rates of human activities to vegetation index in the XX Plateau from 1982 to 2015 before and after improvement.

[0083] Before improvement H 17.43％ 28.86％ 53.72％ Improved H 16.37％ 10.2％ 74.43％

[0084] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0085] The above embodiments are only used to illustrate the design concept and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, all equivalent changes or modifications made based on the principles and design ideas disclosed in the present invention are within the protection scope of the present invention.

Claims

1. An improved algorithm for the relative contribution rate of human activities to long-term vegetation indices, characterized by: Includes the following steps: S1: Collect long-term NDVI remote sensing images of the target area as the data source and preprocess them. Use data fusion algorithms to reconstruct the remote sensing image data to obtain the remote sensing observation NDVI values, and fuse them into a long-term NDVI dataset with the accuracy required by the preset requirements. S2: Collect climate data and other environmental factors for the target area, and perform detrending processing on the climate data and NDVI data; S3: Using multiple environmental factors, including climate data and other environmental elements, as independent variables, and the fused high-precision long-term NDVI data as dependent variables, we use a linear regression model and the least squares method to perform trend analysis to obtain the time-series predicted NDVI value. The specific steps are as follows: The independent variables include temperature (D(T)), precipitation (D(P)), total solar radiation (D(S)), soil moisture (SM), topography (DEM), and land use type (LU). A multiple linear regression model is established to calculate the NDVI predicted value for each parameter in the model, including annual average temperature, precipitation, total solar radiation, soil moisture, topography, and land use type. CC This is used to predict the NDVI value of vegetation; the specific calculation formula is as follows: ； in β 0 is a constant term. β 1 is the regression coefficient between NDVI and temperature D(T). β 2 is the regression coefficient between NDVI and precipitation P. β 3 is the regression coefficient between NDVI and total solar radiation S. β 4 is the regression coefficient between NDVI and soil moisture SM. β 5 is the regression coefficient between NDVI and topographic DEM. β 6 is the regression coefficient between NDVI and land use type LU. β 0 represents the regression constant term, and the undetermined coefficients are calculated using the least squares method; S4: Subtract the predicted NDVI value of the time series calculated in step S3 from the remote sensing observation NDVI value obtained in step S1 to obtain the NDVI residual value, which is used to represent the impact of human activities on the long-term NDVI. S5: Calculate the NDVI residual trend rate based on the residual values ​​obtained in step S4, and calculate the relative contribution rate H of human activities to the long-term series NDVI using a combination of scenario simulation and evaluation.

2. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 1, characterized in that: In step S1, the long-term remote sensing images include SPOT NDVI remote sensing images and GIMMS NDVI remote sensing images; the SPOT NDVI remote sensing images are SPOT-4 product data with a spatial resolution of 1 km and a temporal resolution of 10 days, covering the period from 1999 to 2013; the GIMMS NDVI remote sensing images are GIMMS NDVI3g product data with a spatial resolution of 8 km and a temporal resolution of 15 days, covering the period from 1982 to 2015.

3. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 2, characterized in that: Step S1 further includes the following step: The NDVI data was extracted and clipped using a mask, and the annual maximum NDVI value of the study area was calculated by the maximum value synthesis method to form the NDVI time series dataset of the study area, which was then projected onto the WGS84 coordinate system.

4. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 3, characterized in that: In step S1, the preprocessing includes data stitching, projection transformation, and maximum value synthesis of the remote sensing images; the temporal resolution of the preprocessed SPOT NDVI remote sensing images and GIMMS NDVI remote sensing images is both monthly, while the spatial resolution and temporal coverage remain unchanged. In the overlapping remote sensing images, the pixel values ​​of the 1 km spatial resolution SPOT NDVI remote sensing images from 1999 to 2013 were used as the prediction domain, and the pixel values ​​of the 8 km spatial resolution GIMMS NDVI remote sensing images from 1982 to 2015 were used as the response domain. By utilizing the spatial distribution characteristics of pixel values ​​in the prediction domain and the response domain, an empirical orthogonal telecorrelation model between pixel values ​​in the prediction domain and the response domain is established to perform spatiotemporal fusion of SPOT NDVI remote sensing images and GIMMS NDVI remote sensing images. Spatial spline interpolation was performed on the pixel values ​​of GIMMS NDVI images with a spatial resolution of 8 km from 1982 to 2002 to obtain NDVI pixel values ​​with a spatial resolution of 1 km from 1982 to 2015, which were then projected onto the WGS84 coordinate system.

5. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 1, characterized in that: In step S2, the climate data includes temperature, precipitation, and solar radiation; other environmental factors include soil moisture, topography, and land use type data. The time coverage of remote sensing images of climate data and other environmental elements is from 1982 to 2015. The preprocessing of climate data includes weather station screening, missing data imputation and spatial data interpolation. It also includes summing the daily average temperatures of weather stations throughout the year and dividing by the length of the year to obtain the annual average temperature, and summing the cumulative precipitation within 24 hours of each day throughout the year to obtain the total annual precipitation. The solar radiation dataset is obtained by calculating the annual average solar radiation of meteorological stations by performing an arithmetic average of monthly average solar radiation data. Soil moisture, topography, and land use type data were extracted and clipped using a mask and then reprojected onto the WGS84 coordinate system.

6. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 5, characterized in that: In step S2, the detrending method is to perform univariate linear regression simulations on the factors of temperature, precipitation and solar radiation for each year, and then add the mean of the corresponding variables to the difference. The detrended NDVI, temperature, precipitation, and total solar radiation are denoted as D(NDVI), D(T), D(P), and D(S), respectively.

7. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 1, characterized in that: The specific steps in step S4 are as follows: Calculate NDVI observations NDVI obs NDVI Predicted Value CC The difference is used to obtain the NDVI residual value. HA This is used to represent the impact of human activities on long-term NDVI series; the specific calculation formula is as follows: NDVI HA =NDVI obs -NDVI CC ; If residual NDVI HA >0 indicates that human activities have a promoting effect on vegetation growth; If residual NDVI HA <0 indicates that human activities are detrimental to vegetation growth; If residual NDVI HA =0 indicates that human activities have a weak impact on vegetation change.

8. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 7, characterized in that: The specific steps in step S5 are as follows: For NDVI respectively HA and NDVI CC Linear regression was performed between the series and the year to obtain the trend of NDVI changes under the influence of human activities and climate change alone. β (NDVIres) and β (NDVIpre), the formula is as follows: ； In the formula, β The trend of residual NDVI change. i , j It is a time series. NDVI i , NDVI j They represent the first i , j NDVI value over time; if β >0 indicates that human activities or changes in climate and other environmental factors have a promoting effect on vegetation growth. if β <0 indicates that human activities or changes in climate and other environmental factors have an inhibitory effect on vegetation growth; if β =0 indicates that human activities have a weak impact on vegetation change.

9. The improved algorithm for the relative contribution rate of human activities to long-term vegetation indices according to claim 8, characterized in that: It also includes the following steps: The contribution rate H of human activities to long-term NDVI was obtained by combining residual analysis and scenario simulation assessment. Scenario 1: When S(NDVI) obs ) rise, NDVI CC >0, NDVI HA >0, the contribution rate of climate, soil, topography and land use type change is Human activity contribution rate This indicates that changes in climate and other environmental factors, as well as human activities, have led to an increase in vegetation. Scenario 2: When S(NDVI) obs ) rise, NDVI CC >0, NDVI HA <0, the contribution rate of climate, soil, topography and land use type changes is 100%, and the contribution rate of human activities is 0, indicating that changes in climate and other environmental factors have caused vegetation increase; Scenario 3: When S(NDVI) obs ) rise, NDVI CC <0, NDVI HA >0, the contribution rate of climate, soil, topography and land use type change is 0, and the contribution rate of human activities is 100%, indicating that human activities have caused vegetation increase; Scenario 4: When S(NDVI) obs ) decrease, NDVI CC <0, NDVI HA <0, the contribution rate of climate, soil, topography and land use type change is Human activity contribution rate This indicates that changes in climate and other environmental factors, as well as human activities, have caused vegetation degradation. Scenario 5: When S(NDVI) obs ) decrease, NDVI CC <0, NDVI HA >0, the contribution rate of climate, soil, topography and land use type change is 100%, and the contribution rate of human activities is 0, indicating that changes in climate and other environmental factors have caused vegetation degradation; Scenario 6: When S(NDVI) obs ) decrease, NDVI CC >0, NDVI HA <0, the contribution rate of climate, soil, topography and land use type change is 0, and the contribution rate of human activities is 100%, indicating that human activities cause vegetation degradation.

Images