Inversion method of total primary productivity and related equipment

Abstract

The invention discloses a total primary productivity inversion method and related equipment, and relates to the technical field of remote sensing data processing, and the method comprises the steps: obtaining land type purity data, first spectral data, second spectral data and first leaf area index data in a target region; screening the first spectral data and the first leaf area index data according to the land type purity data to obtain third spectral data and second leaf area index data; establishing a leaf area index inversion model according to the third spectral data and the second leaf area index data; substituting the second spectral data into a leaf area index inversion model to obtain target leaf area index data; and determining total primary productivity data of the target region according to the target leaf area index data. According to the method, the low-resolution data is screened through the land class purity data, and the inversion model is constructed to be applied to the high-resolution data, so that effective transition of cross-resolution data is realized, and the accuracy of total primary productivity inversion is improved.

Description

Technical Field

[0001] The present application relates to the technical field of remote sensing data processing, and more specifically, to an inversion method of total primary productivity and related equipment. Background Art

[0002] With the development of remote sensing technology and geographic information systems, the method of inverting gross primary productivity (GPP) based on remote sensing data has been widely used. However, in the current inversion method, the adaptation problem between low-resolution data and high-resolution data is still a technical problem that needs to be solved urgently; traditional inversion methods often directly use low-resolution data for modeling, but due to the problems of insufficient resolution and poor data quality of low-resolution data itself, the accuracy of the inversion results cannot meet the actual needs.

[0003] In the prior art, leaf area index inversion based on spectral data has been widely used in the estimation of total primary productivity. Usually, the inversion of leaf area index relies on high-resolution spectral data to obtain more accurate results, but the acquisition cost of high-resolution data is high. In practical applications, the acquisition cost of low-resolution data is low and the coverage is wide. However, the quality of low-resolution data is usually poor, and there is a large amount of noisy data, which poses a challenge to the accurate inversion of primary productivity, thereby affecting the reliability and accuracy of the inversion results. How to effectively apply the modeling results of low-resolution data to the inversion of high-resolution data to achieve an effective transition across resolutions is still an important technical bottleneck. That is, there is a technical problem of low accuracy in the inversion of high-resolution total primary productivity in the prior art. Summary of the invention

[0004] A series of simplified concepts are introduced in the summary of the invention, which will be further described in detail in the detailed description. The summary of the invention of this application does not mean to attempt to define the key features and essential technical features of the technical solution claimed for protection, nor does it mean to attempt to determine the scope of protection of the technical solution claimed for protection.

[0005] The inversion method and related equipment for total primary productivity provided in this application can filter low-resolution data through land purity data, improve its quality, build an inversion model and apply it to high-resolution data, thereby achieving an effective transition across resolution data and improving the accuracy of total primary productivity inversion.

[0006] In a first aspect, the present application provides a method for inverting total primary productivity, comprising: acquiring land type purity data, first spectral data, second spectral data and first leaf area index data in a target area, wherein the spatial unit area of ​​the first spectral data, the land type purity data and the first leaf area index data is a first area, the spatial unit area of ​​the second spectral data is a second area, and the first area is greater than the second area; based on the land type purity data, the first spectral data and the first leaf area index data are screened to obtain third spectral data and second leaf area index data; based on the third spectral data and the second leaf area index data, a leaf area index inversion model is established, wherein the leaf area index inversion model is used to deduce leaf area index data based on spectral data; substituting the second spectral data into the leaf area index inversion model to obtain target leaf area index data; and determining the total primary productivity data of the target area based on the target leaf area index data.

[0007] In some embodiments, the process of acquiring the land class purity data includes: acquiring first land class data and second land class data in the target area, wherein the spatial unit area of ​​the first land class data is the first area, and the spatial unit area of ​​the second land class data is the second area; determining the land class purity data based on the correspondence between the first land class data and the second land class data, wherein the spatial unit area of ​​the land class purity data is the first area.

[0008] In some embodiments, the method further includes: obtaining first normalized difference vegetation index data and second normalized difference vegetation index data in the target area, wherein the spatial unit area of ​​the first normalized difference vegetation index data is the first area, and the spatial unit area of ​​the second normalized difference vegetation index data is the second area; establishing a leaf area index inversion model based on the third spectral data and the second leaf area index data, including: dividing the first normalized difference vegetation index data into groups according to a preset numerical interval of the second leaf area index data, and removing the upper quarter data and the lower quarter data of each group in the first normalized difference vegetation index data. The method comprises the steps of: obtaining the third normalized difference vegetation index data from the lower quarter of the total area of ​​the plant and obtaining the third normalized difference vegetation index data; screening the third spectral data and the second leaf area index data according to the third normalized difference vegetation index data to obtain the fourth spectral data and the third leaf area index data; establishing the leaf area index inversion model according to the third normalized difference vegetation index data, the fourth spectral data and the third leaf area index data; substituting the second spectral data into the leaf area index inversion model to obtain the target leaf area index data, comprising: substituting the second normalized difference vegetation index data and the second spectral data into the leaf area index inversion model to obtain the target leaf area index data.

[0009] In some embodiments, the leaf area index inversion model is established based on the third normalized difference vegetation index data, the fourth spectral data and the third leaf area index data, including: obtaining leaf area index quality control data corresponding to the third leaf area index data; screening the third leaf area index data according to the leaf area index quality control data to obtain fourth leaf area index data; screening and obtaining fifth spectral data and fourth normalized difference vegetation index data corresponding to the fourth leaf area index data from the fourth spectral data and the third normalized difference vegetation index data; and establishing the leaf area index inversion model based on the fourth normalized difference vegetation index data, the fifth spectral data and the fourth leaf area index data.

[0010] In some embodiments, the method further includes: acquiring sixth spectral data and seventh spectral data in the target area, wherein the sixth spectral data includes a first blue band, a first green band, a first red band, a near infrared band, a first short-wave infrared 1 band, and a first short-wave infrared 2 band, and the seventh spectral data includes a second blue band, a second green band, a second red band, a narrow near infrared band, a second short-wave infrared 1 band, and a second short-wave infrared 2 band; according to the first blue band, the first green band, the first red band, the near infrared band, the first short-wave infrared 1 band and the first shortwave infrared 2 band, determining a first surface albedo; determining a second surface albedo according to the second blue band, the second green band, the second red band, the narrow near-infrared band, the second shortwave infrared 1 band and the second shortwave infrared 2 band; determining a third surface albedo according to the first surface albedo and the second surface albedo; determining the total primary productivity data of the target area according to the target leaf area index data, including: determining the total primary productivity data according to the third surface albedo and the target leaf area index data.

[0011] In some embodiments, the method also includes: obtaining meteorological reanalysis data in the target area, wherein the meteorological reanalysis data includes first minimum temperature data, first solar radiation data and first saturated air pressure difference data; resampling the first minimum temperature data, the first solar radiation data and the first saturated air pressure difference data to obtain second minimum temperature data, second solar radiation data and second saturated air pressure difference data with a spatial unit area of ​​the second area; determining the total primary productivity data based on the third surface albedo and the target leaf area index data includes: determining the total primary productivity data based on the second minimum temperature data, the second solar radiation data, the second saturated air pressure difference data, the third surface albedo and the target leaf area index data.

[0012] In some embodiments, the meteorological reanalysis data also includes precipitation data and wind speed data; the resampling of the first minimum temperature data, the first solar radiation data, and the first saturated air pressure difference data to obtain the second minimum temperature data, the second solar radiation data, and the second saturated air pressure difference data with a spatial unit area of ​​the second area includes: taking the first minimum temperature data, the first solar radiation data, and the first saturated air pressure difference data as dependent variables in turn, and screening the other two types of data in the meteorological reanalysis data through correlation analysis to establish a first multi-scale geographically weighted regression model; resampling the first regression parameter, the first regression constant, and the first error term of the first multi-scale geographically weighted regression model to obtain the second regression parameter, the second regression constant, and the second error term that match the second area; solving the dependent variable by using the second multi-scale geographically weighted regression model established by the second regression parameter, the second regression constant, and the second error term to obtain the second minimum temperature data, the second solar radiation data, and the second saturated air pressure difference data.

[0013] In the second aspect, the present application also provides a total primary productivity inversion device, including: a data acquisition unit, used to acquire land type purity data, first spectral data, second spectral data and first leaf area index data in a target area, wherein the spatial unit areas of the first spectral data, the land type purity data and the first leaf area index data are all first areas, the spatial unit area of ​​the second spectral data is second area, and the first area is larger than the second area; a data screening unit, used to screen the first spectral data and the first leaf area index data according to the land type purity data to obtain third spectral data and second leaf area index data; a model building unit, used to establish a leaf area index inversion model according to the third spectral data and the second leaf area index data; a calculation parameter acquisition unit, used to substitute the second spectral data into the leaf area index inversion model to obtain target leaf area index data; a productivity calculation unit, used to determine the total primary productivity data of the target area according to the target leaf area index data.

[0014] In a third aspect, the present application further provides an electronic device comprising: a memory and a processor, wherein the processor is used to implement the steps of the inversion method of total primary productivity described in the first aspect when executing a computer program stored in the memory.

[0015] In a fourth aspect, the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the inversion method of total primary productivity described in the first aspect.

[0016] In a fifth aspect, the present application also provides a computer program product, including a computer program or computer executable instructions, which, when executed by a processor, implements the inversion method of total primary productivity provided in an embodiment of the present application.

[0017] In summary, the present application screens the first spectral data and the first leaf area index data through land purity data, thereby improving the quality of low-resolution data whose spatial unit area is the first area, and eliminating low-quality data with low land purity, which can improve the reliability and representativeness of low-resolution data, making the correspondence between the remaining spectral data and leaf area index data more accurate, and providing a more reliable data basis for subsequent modeling; after establishing a leaf area index inversion model based on low-resolution data, this model can be applied to the inversion of high-resolution data whose spatial unit area is the second area; because high-resolution data can provide more details, the inversion result will be more accurate, realizing cross-resolution data utilization, and effectively transitioning from low-resolution data to high-resolution data, thereby improving the accuracy of leaf area index inversion. In summary, the inversion method of total primary productivity provided by the present application screens low-resolution data through land purity data, improves its quality, constructs an inversion model and applies it to high-resolution data, realizes the effective transition of cross-resolution data, and improves the accuracy of total primary productivity inversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present specification. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings: Figure 1 A schematic flow chart of a method for inverting total primary productivity provided in an embodiment of the present application; Figure 2 A schematic diagram of the composition structure of a total primary productivity inversion device provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] Terms in the specification, claims and drawings of this application, such as "first", "second", "third", "fourth", etc. (if any), are used to distinguish similar objects rather than to describe a specific order or sequence. Therefore, it is understood that these terms can be used interchangeably where appropriate, so that the described embodiments can be implemented in different orders unless otherwise required in the drawings or descriptions. In addition, the terms "is" and "has" and any variations thereof in this application are intended to cover all possible constituent elements on a non-exclusive basis. For example, a process, method, system, product or device comprising several steps or units is not necessarily limited to the steps or units that are explicitly listed, but may also include other steps or units that are not explicitly listed, or steps or units that are inherent to the process, method, product or device.

[0020] In this application, a "module" or "unit" refers to a computer program or a part of a computer program that has a specific function and works in conjunction with other related parts to achieve a predetermined goal. These modules or units can be implemented by software, hardware (such as processing circuits or memories), or a combination of the two. One or more processors or memories can implement one or more modules or units. At the same time, each module or unit can also be a part of a larger module or unit.

[0021] The technical solutions in the present application will be described in detail below in conjunction with the drawings in the embodiments. It should be noted that the described embodiments are only a part of the present application, not all embodiments. In the following description, the "some embodiments" mentioned are only subsets of all possible embodiments, which may be the same or different subsets, and different embodiments may be combined with each other without conflict.

[0022] Figure 1 is a flow chart of a method for inverting total primary productivity provided in an embodiment of the present application. For example, see Figure 1 The inversion method of total primary productivity provided in the embodiment of the present application may include the following steps 101 to 105: Step 101, obtaining land purity data, first spectrum data, second spectrum data and first leaf area index data in a target area; The spatial unit area of ​​the first spectral data, the land purity data and the first leaf area index data is the first area, the spatial unit area of ​​the second spectral data is the second area, and the first area is greater than the second area; In some examples, the target area is the area where the total primary productivity inversion is performed. The land purity data is numerical data used to characterize the purity or quality of various types of land surfaces in the target area, reflecting the proportion of a specific land type in a specific area or pixel in the target area. The larger the proportion of the specific land type in the area, the higher the land purity in the area, and vice versa. For example, the specific land type can include forests, farmlands, and grasslands; remote sensing data classification technology can be used to determine the land type of each pixel and calculate the land purity data. For example, remote sensing data classification technology can include supervised classification, unsupervised classification, and machine learning. The first spectral data refers to the reflectance spectral data obtained by remote sensing equipment in a specific spectral band. The first spectral data has a lower spatial resolution and is used to cover a larger area. The specific spectral band can be a spectral band such as visible light and near infrared. For example, the surface reflectance spectral data obtained by remote sensing satellites such as Visible Infrared Imaging Radiometer Suite (VIIRS) and Moderate Resolution Imaging Spectroradiometer (MODIS) can cover a larger area, and the spatial resolution of each pixel can be 500 meters. The second spectral data is also remote sensing image data, but it has a higher spatial resolution. For example, remote sensing platforms such as Landsat8, Sentinel-2 and Satellite Pour l'Observation dela Terre (SPOT) series can be used to obtain remote sensing images, and the spatial resolution of each pixel can be 30 meters. Leaf Area Index (LAI) is the ratio of the total area of ​​plant leaves per unit ground area to the area, which is used to describe the density and health of plant communities; the first leaf area index data refers to the leaf area index data estimated based on lower resolution remote sensing data, which can be obtained through remote sensing inversion models or through the combination of field measurements and remote sensing data. The spatial unit area refers to the actual area of ​​the pixel or unit of remote sensing data on the ground. In remote sensing images, each pixel represents a fixed-size ground area; for example, the first area corresponding to the spatial unit area of ​​the first spectral data and the first leaf area index data is larger, such as a 500m×500m pixel, while the second area corresponding to the spatial unit area of ​​the second spectral data is smaller, such as a 30m×30m pixel.

[0023] By implementing step 101, the land purity data, the first spectral data, the second spectral data and the first leaf area index data are obtained, which can make full use of the characteristics of different data sources, provide more comprehensive ground information, and help to use cross-resolution data to perform inversion processing of total primary productivity.

[0024] Step 102, screening the first spectrum data and the first leaf area index data according to the land purity data to obtain third spectrum data and second leaf area index data; Exemplarily, the third spectral data is data filtered out from the first spectral data, and only the spectral data of the area with higher purity of land type will be retained. The second leaf area index data is data filtered out from the first leaf area index data, and only the spectral data of the area with higher purity of land type will be retained.

[0025] For example, areas with higher land purity can be screened out based on the land purity data, and the first spectral data and first leaf area index data corresponding to these areas are retained, and the relevant data of low land purity areas are eliminated to obtain third spectral data and second leaf area index data. The data finally obtained reflect higher quality, more representative and accurate surface information, and can be used for subsequent leaf area index inversion modeling.

[0026] By implementing step 102, data in areas with low land purity are removed, which can effectively improve the quality of the two low-resolution data, the first spectral data and the first leaf area index data, thereby ensuring that the correspondence between the spectral data and the leaf area index data is more accurate and that the data used for modeling is more reliable.

[0027] Step 103, establishing a leaf area index inversion model according to the third spectral data and the second leaf area index data; Among them, the leaf area index inversion model is used to deduce leaf area index data based on spectral data; In some examples, the leaf area index inversion model is a mathematical model that aims to use spectral data to predict or deduce the leaf area index of a target area; the leaf area index inversion model can be constructed by regression analysis, machine learning methods, or statistical methods, and trained using the third spectral data and the second leaf area index data obtained in step 102.

[0028] For example, a support vector machine (SVM), a random forest (RF), or a neural network may be used to train a model using the third spectral data, thereby deducing the second leaf area index data.

[0029] Through the implementation of step 103, a leaf area index inversion model is established based on the screened data, i.e., the third spectral data and the second leaf area index data, so that the process of deducing leaf area index data from spectral data can be realized, providing a model basis for subsequent high-precision inversion and improving the prediction accuracy of the model.

[0030] Step 104, substituting the second spectral data into the leaf area index inversion model to obtain target leaf area index data; In some examples, the target leaf area index data is obtained by substituting the second spectral data into a pre-established leaf area index inversion model and performing calculation deduction.

[0031] Exemplarily, the second spectral data can be used as input, and the leaf area index of the target area can be calculated by using the input spectral data through the inversion model, and the output target leaf area index data is the estimated leaf area index data of the target area, which is a grid representing the leaf area index value of each pixel in the area. Exemplarily, the second spectral data can include reflectance in visible light, infrared, near infrared and other bands; the inversion model can use a regression model, a machine learning model, etc.

[0032] By implementing step 104, the high-resolution data is deduced using the established inversion model to obtain target leaf area index data, which can more accurately reflect the actual situation of the target area.

[0033] Step 105, determining the total primary productivity data of the target area according to the target leaf area index data; Specifically, the total primary productivity of the target area refers to the total amount of light energy converted into chemical energy by plants in the target area through photosynthesis within a specific period of time. The total primary productivity data of the target area can be effectively derived by combining the target leaf area index data with the light radiation transfer model.

[0034] Through the implementation of step 105, the total primary productivity data of the target area of ​​the earth is further calculated based on the target leaf area index data. Through the refined modeling and data screening of the previous steps, the accuracy of the total primary productivity inversion prediction of step 105 is effectively improved.

[0035] In summary, the embodiment of the present application screens the first spectral data and the first leaf area index data through the land purity data, thereby improving the quality of the low-resolution data whose spatial unit area is the first area, and eliminating the low-quality data with low land purity, which can improve the reliability and representativeness of the low-resolution data, so that the correspondence between the remaining spectral data and the leaf area index data is more accurate, providing a more reliable data basis for subsequent modeling; after establishing a leaf area index inversion model based on low-resolution data, this model can be applied to the inversion of high-resolution data whose spatial unit area is the second area; because high-resolution data can provide more details, the inversion result will be more accurate, realizing the cross-resolution data utilization, and the effective transition from low-resolution data to high-resolution data, thereby improving the accuracy of leaf area index inversion. In summary, the inversion method of total primary productivity provided in the embodiment of the present application screens low-resolution data through land purity data, improves its quality, constructs an inversion model and applies it to high-resolution data, realizes the effective transition of cross-resolution data, and improves the accuracy of total primary productivity inversion.

[0036] In some embodiments, the process of acquiring the aforementioned land class purity data may include: acquiring first land class data and second land class data in the target area, wherein the spatial unit area of ​​the first land class data is a first area, and the spatial unit area of ​​the second land class data is a second area; determining the land class purity data based on the corresponding relationship between the first land class data and the second land class data, wherein the spatial unit area of ​​the land class purity data is the first area.

[0037] In some examples, the first land class data is a land use / land cover dataset of the target area, and its spatial resolution is the "first area" size; for example, the MODIS land cover product (MCD12Q1.061) dataset, which is provided by the MODIS satellite of the National Aeronautics and Space Administration (NASA), has a spatial resolution of 500 meters and provides land use / land cover types on a global scale. The second land class data is another land use / land cover dataset, and its spatial resolution is the "second area" size; for example, the global land cover (GlobeLand) dataset, which provides land use data with a resolution of 30 meters, and can provide higher spatial details than 500-meter resolution data. The land class purity can be calculated by comparing the MCD12Q1.061 with a resolution of 500 meters and the GlobeLand data with a resolution of 30 meters, two datasets with different resolutions. The land class purity reflects the matching degree of data with different resolutions at the same location; if the high-resolution (30-meter) data is very consistent in the land class type within a large area of ​​low-resolution (500-meter) cells, the land class purity is high.

[0038] For example, the land class data in the MCD12Q1.061 dataset include water bodies, grasslands, shrublands, broadleaf croplands, savannas, evergreen broadleaf forests, deciduous broadleaf forests, evergreen needleleaf forests, deciduous needleleaf forests, non-vegetated lands and urban and built-up lands; the land class data in the GlobeLand dataset include cultivated land, forest, grass land, shrub land, wetland, water bodies, tundra, artificial surface The "Water Body" category in MCD12Q1.061 corresponds to the "Water Body" category in GlobeLand30; "Grassland" corresponds to "Grassland"; "Shrub" corresponds to "Shrubland"; "Broadleaved Farmland" corresponds to "Cultivated Land"; "Savanna" corresponds to "Forest Land"; "Evergreen Broadleaf Forest", "Deciduous Broadleaf Forest", "Evergreen Coniferous Forest", "Deciduous Coniferous Forest" all correspond to the "Forest Land" category; "Bare Land" corresponds to "Bare Land"; "Urban and Built-up Land" corresponds to "Artificial Surfaces".

[0039] Through the implementation of the above embodiment, the first land class data and the second land class data (such as GlobeLand30, 30-meter resolution data) in the target area are obtained, and data fusion is performed through the correspondence between the two, which can overcome the differences between data of different resolutions, and can balance between larger scales (500 meters) and smaller scales (30 meters), and eliminate low-quality and low-representative areas; illustratively, the first land class data can be 500-meter resolution data such as MCD12Q1.061), and the second land class data can be 30-meter resolution data such as GlobeLand30. Specifically, the first land class data may have greater spatial heterogeneity due to its low resolution, while the second land class data provides finer details and can supplement the deficiencies of the first land class data; and then by accurately determining the land class purity data, the subsequent model establishment and prediction analysis are more reliable.

[0040] In some embodiments, the aforementioned method may further include: acquiring first normalized difference vegetation index data and second normalized difference vegetation index data in the aforementioned target area, wherein the spatial unit area of ​​the first normalized difference vegetation index data is the first area, and the spatial unit area of ​​the second normalized difference vegetation index data is the second area; step 103 may include: grouping the first normalized difference vegetation index data according to a preset numerical interval of the second leaf area index data, and removing the upper quarter data and the lower quarter data of each group in the first normalized difference vegetation index data to obtain third normalized difference vegetation index data; according to the third normalized difference vegetation index data, screening the third spectral data and the second leaf area index data to obtain fourth spectral data and third leaf area index data; establishing a leaf area index inversion model according to the third normalized difference vegetation index data, the fourth spectral data and the third leaf area index data; step 104 may include: substituting the second normalized difference vegetation index data and the second spectral data into the leaf area index inversion model to obtain target leaf area index data.

[0041] In some examples, the first normalized difference vegetation index data is normalized difference vegetation index data calculated by remote sensing images. The second normalized difference vegetation index data is normalized difference vegetation index data calculated by high-resolution remote sensing data. The leaf area index has a strong positive correlation with the normalized difference vegetation index. Within a specific leaf area index value range, the normalized difference vegetation index value tends to fall within a relatively stable normal range; therefore, dividing the second leaf area index data into preset numerical intervals can screen out more representative normalized difference vegetation index data; for example, the preset numerical interval can be an interval of 0.2m² / m² (such as 0m² / m²-0.2m² / m², 0.2m² / m²-0.4m² / m², etc.). In the implementation process, the first normalized difference vegetation index data can be divided into several groups according to the preset numerical interval of the second leaf area index data, and the data in each group can be sorted. In each group, the upper quarter and the lower quarter of the first normalized difference vegetation index data are eliminated to remove potential extreme values ​​or abnormal data, and obtain the third normalized difference vegetation index data; this processing can effectively reduce the noise in the data and enhance the robustness and stability of the subsequent model. The fourth spectral data and the third leaf area index data are selected based on the third normalized difference vegetation index data, and the third spectral data and the second leaf area index data are screened to select the band data and index data that are consistent with the area / pixel where the third normalized difference vegetation index data is located.

[0042] Through the implementation of the above embodiment, the first normalized difference vegetation index data is divided into groups, and abnormal samples in the upper quarter and the lower quarter are eliminated, which can effectively remove noise and extreme values ​​in the data, making subsequent model training more stable and the inversion results more accurate.

[0043] In some embodiments, the aforementioned establishment of a leaf area index inversion model based on the third normalized difference vegetation index data, the fourth spectral data and the third leaf area index data may include: obtaining leaf area index quality control data corresponding to the third leaf area index data; screening the third leaf area index data according to the leaf area index quality control data to obtain fourth leaf area index data; screening and obtaining fifth spectral data and fourth normalized difference vegetation index data corresponding to the fourth leaf area index data from the fourth spectral data and the third normalized difference vegetation index data; and establishing a leaf area index inversion model based on the fourth normalized difference vegetation index data, the fifth spectral data and the fourth leaf area index data.

[0044] In some examples, leaf area index quality control data is an indicator used to evaluate the quality of leaf area index data. Through these quality control data, leaf area index data points with poor quality or unreliable can be identified and eliminated to ensure the accuracy of subsequent data analysis. Leaf area index quality control data can come from the quality mark added by the remote sensing platform when acquiring leaf area index data; for example, MODIS's leaf area index data product (MOD15A2H) provides leaf area index quality control data information. The fourth leaf area index data is a more accurate leaf area index data obtained after quality control screening. The fifth spectral data and the fourth normalized difference vegetation index data are data sets screened from the spectral data corresponding to the fourth leaf area index data, and are consistent with the area / pixel where the fourth leaf area index data is located.

[0045] For example, in the leaf area index data of MODIS, the leaf area index quality control data is used to mark which pixels have reliable leaf area index values. The fourth leaf area index data can be obtained by deleting the leaf area index quality control data outside the range of 0 to 8 and the leaf area index quality control data that is an odd number.

[0046] Through the implementation of the above embodiments, by performing quality control screening on the leaf area index data and removing low-quality or unreliable data points, the quality of training data for establishing the leaf area index inversion model can be significantly improved, thereby improving the accuracy of the inversion results.

[0047] In some embodiments, the aforementioned method may further include: acquiring sixth spectral data and seventh spectral data in the aforementioned target area, wherein the sixth spectral data may include a first blue band, a first green band, a first red band, a near infrared band, a first shortwave infrared 1 band, and a first shortwave infrared 2 band, and the aforementioned seventh spectral data may include a second blue band, a second green band, a second red band, a narrow near infrared band, a second shortwave infrared 1 band, and a second shortwave infrared 2 band; determining a first surface albedo according to the aforementioned first blue band, first green band, first red band, near infrared band, first shortwave infrared 1 band, and first shortwave infrared 2 band; determining a second surface albedo according to the second blue band, second green band, second red band, narrow near infrared band, second shortwave infrared 1 band, and second shortwave infrared 2 band; determining a third surface albedo according to the first surface albedo and the second surface albedo; the aforementioned step 105 may include: determining total primary productivity data according to the third surface albedo and target leaf area index data.

[0048] In some examples, the first blue band, the first green band, the first red band, the near infrared band, the first shortwave infrared band 1 and the first shortwave infrared band 2 are derived from a certain type of remote sensing satellite image data, which can be obtained through remote sensing images; the second blue band, the second green band, the second red band, the narrow near infrared band, the second shortwave infrared band 1 and the second shortwave infrared band 2 are derived from remote sensing satellites or different sensor devices different from those used to obtain the sixth spectral data. Albedo is an indicator of reflectivity, which indicates the ratio of solar radiation reflected by the surface to incident radiation; the first surface albedo is calculated based on the sixth spectral data; the second surface albedo is calculated based on the seventh spectral data; the third surface albedo is a comprehensive albedo obtained by combining the first surface albedo and the second surface albedo, using weighted average or other data fusion methods; illustratively, the sixth spectral data can be the data of each band in the Landsat image, and the seventh spectral data can be the data of each band in the Sentinel-2 image. The total primary productivity data can be effectively derived using radiative transfer models and energy balance models by combining the third surface albedo and target leaf area index data.

[0049] Exemplarily, the first surface albedo can be calculated by the following formula: In the formula, is the first surface albedo, represents the first blue band, represents the first green band, represents the first red band, represents the near-infrared band, Indicates the first short-wave infrared 1 band, Indicates the first shortwave infrared 2 band.

[0050] The second surface albedo can be calculated by the following formula: In the formula, is the second surface albedo, represents the second blue band, represents the second green band, represents the second red band, represents a narrow near-infrared band, Indicates the second short-wave infrared 1 band, Indicates the second shortwave infrared 2 band.

[0051] Through the implementation of the above-mentioned embodiment, the use of the sixth spectral data and the seventh spectral data can effectively compensate for the surface features that may be ignored by single-source data. By calculating the first surface albedo and the second surface albedo and combining the two to obtain the third surface albedo, the surface reflection characteristics of the target area can be more accurately evaluated, and the total primary productivity data of the target area can be more accurately inverted.

[0052] In some embodiments, the aforementioned method may further include: obtaining meteorological reanalysis data within the aforementioned target area, wherein the meteorological reanalysis data may include first minimum temperature data, first solar radiation data and first saturated air pressure difference data; resampling the aforementioned first minimum temperature data, first solar radiation data and first saturated air pressure difference data to obtain second minimum temperature data, second solar radiation data and second saturated air pressure difference data with a spatial unit area of ​​a second area; determining total primary productivity data based on the third surface albedo and target leaf area index data, which may include: determining total primary productivity data based on the second minimum temperature data, second solar radiation data, second saturated air pressure difference data, third surface albedo and target leaf area index data.

[0053] In some examples, meteorological reanalysis data is historical meteorological conditions obtained by reanalyzing the combination of numerical weather forecast models and historical meteorological observation data, which usually includes multiple meteorological elements such as temperature, radiation, humidity, and air pressure, and has high temporal continuity and spatial coverage; the meteorological reanalysis data of the embodiment of the present application can often be extracted from data released by meteorological agencies such as the National Oceanic and Atmospheric Administration (NOAA) of the United States and the European Centre for Medium-Range Weather Forecasts (ECMWF). The first minimum temperature data, the first solar radiation data, and the first saturated air pressure difference data are specific meteorological elements from the meteorological reanalysis data; the second minimum temperature data, the second solar radiation data, and the second saturated air pressure difference data are data after resampling the first minimum temperature data, the first solar radiation data, and the first saturated air pressure difference data, and the resampling process can be achieved by spatial interpolation or regridding operations; for example, meteorological data with a resolution of 1 km can be resampled to data with a resolution of 500 m or 30 m, and processed using Geographic Information System (GIS) software or programming tools. The total primary productivity data of the target area can be comprehensively derived by combining the second minimum temperature, second solar radiation, second saturated air pressure difference, third surface albedo and target leaf area index data, using multivariate radiation transfer model and energy balance model, combining the effects of temperature, moisture and radiation.

[0054] Through the implementation of the above-mentioned embodiment, by resampling and fusing multiple data sources such as remote sensing data and meteorological data, lower resolution meteorological data can be converted into a format compatible with high-resolution remote sensing data, thereby combining meteorological data with factors such as surface albedo and leaf area index to more accurately estimate the total primary productivity data of the target area.

[0055] In some embodiments, the aforementioned meteorological reanalysis data may also include precipitation data and wind speed data; the aforementioned resampling of the first minimum temperature data, the first solar radiation data and the first saturated air pressure difference data to obtain the second minimum temperature data, the second solar radiation data and the second saturated air pressure difference data with a spatial unit area of ​​the second area may include: taking the first minimum temperature data, the first solar radiation data and the first saturated air pressure difference data as dependent variables in turn, and screening the other two types of data in the meteorological reanalysis data through correlation analysis to establish a first multi-scale geographically weighted regression model. In the process of establishing the first multi-scale geographically weighted regression model, the normalized difference vegetation index data and the digital elevation data of the target area may also be used as independent variables; resampling the first regression parameter, the first regression constant and the first error term of the first multi-scale geographically weighted regression model to obtain the second regression parameter, the second regression constant and the second error term matching the second area; solving the dependent variable by using the second multi-scale geographically weighted regression model established by using the second regression parameter, the second regression constant and the second error term to obtain the second minimum temperature data, the second solar radiation data and the second saturated air pressure difference data.

[0056] In some examples, a Geographically Weighted Regression (GWR) model is a spatial statistical model that can handle spatial heterogeneity between independent variables and dependent variables in spatial data, that is, the regression relationship in different regions may be different; Multi-Scale Geographically Weighted Regression (MGWR) is an extension of GWR, which performs regression analysis at different scales to better capture the multi-level characteristics of spatial data; the first multi-scale GWR model refers to an initial regression model constructed based on the first minimum temperature data, the first solar radiation data, and the first saturated air pressure difference data; the first regression parameter, the first regression constant, and the first error term are obtained when constructing the first multi-scale GWR model; the second multi-scale GWR model is a new regression model obtained by resampling the first regression parameter, the first regression constant, and the first error term, and is used to process new spatial data; different from the first multi-scale GWR model, the second multi-scale GWR model will perform analysis based on new spatial units to obtain more refined spatial predictions.

[0057] Exemplarily, firstly, a correlation analysis is performed on the meteorological data to find the independent variable data with the most similar changing trend to the first minimum temperature data; for example, solar radiation and precipitation may have a strong correlation with the minimum temperature; then, the minimum temperature data is used as the dependent variable, and the solar radiation and precipitation are used as independent variables to establish a first multi-scale geographically weighted regression model, and then the first regression parameter, the first regression constant, and the first error term of the first multi-scale geographically weighted regression model are resampled to obtain a second multi-scale geographically weighted regression model, and then more refined second minimum temperature data are obtained; finally, the above steps are repeated to perform similar regression analysis on the solar radiation data and the saturated pressure difference data, respectively, as the dependent variables, to obtain the second solar radiation data and the second saturated pressure difference data.

[0058] Through the implementation of the above-mentioned embodiment, the most relevant independent variables are screened by correlation analysis, which can ensure that the information used by the regression model is more relevant and improve the prediction accuracy. The use of a multi-scale geographically weighted regression model can reveal the relationship between the dependent variable and the independent variable in different geographical regions and at different scales, and can effectively integrate a variety of meteorological data to provide accurate basic data for further inversion of total primary productivity data.

[0059] Furthermore, as an implementation of the aforementioned method embodiment, the present application also provides a total primary productivity inversion device for implementing the aforementioned method embodiment. The device embodiment corresponds to the aforementioned method embodiment. For ease of reading, the present total primary productivity inversion device embodiment will no longer repeat the details of the aforementioned method embodiment one by one, but it should be clear that the device in the embodiment of the present application can correspond to and implement all the contents of the aforementioned method embodiment. Figure 2As shown, the total primary productivity inversion device 20 includes: a data acquisition unit 201, a data screening unit 202, a model building unit 203, a calculation parameter acquisition unit 204 and a productivity calculation unit 205, wherein the data acquisition unit 201 is used to acquire land purity data, first spectral data, second spectral data and first leaf area index data in the target area, wherein the spatial unit area of ​​the first spectral data, the land purity data and the first leaf area index data is the first area, the spatial unit area of ​​the second spectral data is the second area, and the first area is greater than the second area; the data screening unit 202 is used to screen the first spectral data and the first leaf area index data according to the land purity data to obtain the third spectral data and the second leaf area index data; the model building unit 203 is used to establish a leaf area index inversion model according to the third spectral data and the second leaf area index data; the calculation parameter acquisition unit 204 is used to substitute the second spectral data into the leaf area index inversion model to obtain the target leaf area index data; the productivity calculation unit 205 is used to determine the total primary productivity data of the target area according to the target leaf area index data.

[0060] In some embodiments, the data acquisition unit 201 is also used to acquire first land class data and second land class data in the target area, wherein the spatial unit area of ​​the first land class data is a first area, and the spatial unit area of ​​the second land class data is a second area; based on the correspondence between the first land class data and the second land class data, the land class purity data is determined, wherein the spatial unit area of ​​the land class purity data is the first area.

[0061] In some embodiments, the data acquisition unit 201 is further used to acquire first normalized difference vegetation index data and second normalized difference vegetation index data in the target area, wherein the spatial unit area of ​​the first normalized difference vegetation index data is the first area, and the spatial unit area of ​​the second normalized difference vegetation index data is the second area; the model establishment unit 203 is further used to group the first normalized difference vegetation index data according to a preset numerical interval of the second leaf area index data, and remove the upper quarter data and the lower quarter data of each group in the first normalized difference vegetation index data to obtain third normalized difference vegetation index data; according to the third normalized difference vegetation index data, the third spectral data and the second leaf area index data are screened to obtain fourth spectral data and third leaf area index data; according to the third normalized difference vegetation index data, the fourth spectral data and the third leaf area index data, a leaf area index inversion model is established; the calculation parameter acquisition unit 204 is further used to substitute the second normalized difference vegetation index data and the second spectral data into the leaf area index inversion model to obtain target leaf area index data.

[0062] In some embodiments, the model building unit 203 is also used to obtain leaf area index quality control data corresponding to the third leaf area index data; according to the leaf area index quality control data, the third leaf area index data is screened to obtain fourth leaf area index data; from the fourth spectral data and the third normalized difference vegetation index data, fifth spectral data and fourth normalized difference vegetation index data corresponding to the fourth leaf area index data are screened; according to the fourth normalized difference vegetation index data, the fifth spectral data and the fourth leaf area index data, a leaf area index inversion model is established.

[0063] In some embodiments, the data acquisition unit 201 is further used to acquire sixth spectral data and seventh spectral data in the target area, wherein the sixth spectral data includes a first blue band, a first green band, a first red band, a near infrared band, a first shortwave infrared 1 band and a first shortwave infrared 2 band, and the seventh spectral data includes a second blue band, a second green band, a second red band, a narrow near infrared band, a second shortwave infrared 1 band and a second shortwave infrared 2 band; a first surface albedo is determined based on the first blue band, the first green band, the first red band, the near infrared band, the first shortwave infrared 1 band and the first shortwave infrared 2 band; a second surface albedo is determined based on the second blue band, the second green band, the second red band, the narrow near infrared band, the second shortwave infrared 1 band and the second shortwave infrared 2 band; a third surface albedo is determined based on the first surface albedo and the second surface albedo; the productivity calculation unit 205 is also used to determine total primary productivity data based on the third surface albedo and target leaf area index data.

[0064] In some embodiments, the data acquisition unit 201 is also used to acquire meteorological reanalysis data in the target area, wherein the meteorological reanalysis data includes first minimum temperature data, first solar radiation data and first saturated air pressure difference data; the first minimum temperature data, the first solar radiation data and the first saturated air pressure difference data are resampled to obtain second minimum temperature data, second solar radiation data and second saturated air pressure difference data with a spatial unit area of ​​a second area; the productivity calculation unit 205 is also used to determine the total primary productivity data based on the second minimum temperature data, the second solar radiation data, the second saturated air pressure difference data, the third surface albedo and the target leaf area index data.

[0065] In some embodiments, the meteorological reanalysis data also includes precipitation data and wind speed data; the data acquisition unit 201 is also used to use the first minimum temperature data, the first solar radiation data and the first saturated air pressure difference data as dependent variables in turn, and to establish a first multi-scale geographically weighted regression model by screening the other two types of data in the meteorological reanalysis data through correlation analysis; resample the first regression parameter, the first regression constant and the first error term of the first multi-scale geographically weighted regression model to obtain the second regression parameter, the second regression constant and the second error term that match the second area; solve the dependent variables by using the second multi-scale geographically weighted regression model established by the second regression parameter, the second regression constant and the second error term to obtain the second minimum temperature data, the second solar radiation data and the second saturated air pressure difference data.

[0066] The present application also provides a computer-readable storage medium, which stores computer-executable instructions or computer programs. When the computer-executable instructions or computer programs are executed by a processor, the processor will be caused to execute any step of the inversion method of total primary productivity provided in the present application.

[0067] In some embodiments, the computer-readable storage medium may be a random access memory (RAM), a read-only memory (ROM), a flash memory, a magnetic surface memory, an optical disk, or a compact disc read-only memory (CD-ROM); or it may be various devices including one or any combination of the above memories.

[0068] In some embodiments, computer executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0069] In some embodiments, computer-executable instructions may, but need not, correspond to a file in a file system, may be stored as part of a file storing other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions).

[0070] In some embodiments, computer executable instructions may be deployed to be executed on one electronic device, or on multiple electronic devices located at one site, or on multiple electronic devices distributed at multiple sites and interconnected by a communication network.

[0071] like Figure 3 As shown, the present application also provides an electronic device 30, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor, and when the processor 320 executes the computer program 311, any step of the above-mentioned total primary productivity inversion method is implemented.

[0072] The present application also provides a computer program product, which includes a computer program or a computer executable instruction, and the computer program or the computer executable instruction is stored in a computer-readable storage medium. The processor of the electronic device reads the computer program or the computer executable instruction from the computer-readable storage medium, and the processor executes the computer program or the computer executable instruction, so that the electronic device performs any step of the inversion method of the total primary productivity described above in the present application.

[0073] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for inversion of total primary productivity, characterized in that: include: Acquire land purity data, first spectral data, second spectral data and first leaf area index data in a target area, wherein the spatial unit areas of the first spectral data, the land purity data and the first leaf area index data are all first areas, the spatial unit area of ​​the second spectral data is second area, and the first area is larger than the second area; According to the land type purity data, the first spectrum data and the first leaf area index data are screened to obtain third spectrum data and second leaf area index data; Establishing a leaf area index inversion model according to the third spectral data and the second leaf area index data, wherein the leaf area index inversion model is used to deduce the leaf area index data based on the spectral data; Substituting the second spectral data into the leaf area index inversion model to obtain target leaf area index data; The total primary productivity data of the target area is determined based on the target leaf area index data.

2. The inversion method of total primary productivity according to claim 1, characterized in that: The process of obtaining the land purity data includes: Acquire first land category data and second land category data in the target area, wherein the spatial unit area of ​​the first land category data is the first area, and the spatial unit area of ​​the second land category data is the second area; The land class purity data is determined according to the corresponding relationship between the first land class data and the second land class data, wherein the spatial unit area of ​​the land class purity data is the first area.

3. The inversion method of total primary productivity according to claim 1, characterized in that: The method further comprises: Acquire first normalized difference vegetation index data and second normalized difference vegetation index data in the target area, wherein the spatial unit area of ​​the first normalized difference vegetation index data is the first area, and the spatial unit area of ​​the second normalized difference vegetation index data is the second area; The step of establishing a leaf area index inversion model according to the third spectral data and the second leaf area index data comprises: The first normalized difference vegetation index data are divided into groups according to the preset value interval of the second leaf area index data, and the upper quarter data and the lower quarter data of each group in the first normalized difference vegetation index data are eliminated to obtain third normalized difference vegetation index data; According to the third normalized difference vegetation index data, the third spectral data and the second leaf area index data are screened to obtain fourth spectral data and third leaf area index data; Establishing the leaf area index inversion model according to the third normalized difference vegetation index data, the fourth spectral data and the third leaf area index data; Substituting the second spectral data into the leaf area index inversion model to obtain target leaf area index data includes: Substituting the second normalized difference vegetation index data and the second spectral data into the leaf area index inversion model, the target leaf area index data is obtained.

4. The inversion method of total primary productivity according to claim 3, characterized in that: The step of establishing the leaf area index inversion model according to the third normalized difference vegetation index data, the fourth spectral data and the third leaf area index data comprises: Acquire leaf area index quality control data corresponding to the third leaf area index data; According to the leaf area index quality control data, the third leaf area index data is screened to obtain fourth leaf area index data; Filtering the fourth spectral data and the third normalized difference vegetation index data to obtain fifth spectral data and fourth normalized difference vegetation index data corresponding to the fourth leaf area index data; The leaf area index inversion model is established according to the fourth normalized difference vegetation index data, the fifth spectral data and the fourth leaf area index data.

5. The inversion method of total primary productivity according to claim 1, characterized in that: The method further comprises: Acquire sixth spectral data and seventh spectral data within the target area, wherein the sixth spectral data includes a first blue band, a first green band, a first red band, a near infrared band, a first short-wave infrared 1 band, and a first short-wave infrared 2 band, and the seventh spectral data includes a second blue band, a second green band, a second red band, a narrow near infrared band, a second short-wave infrared 1 band, and a second short-wave infrared 2 band; Determine a first surface albedo according to the first blue band, the first green band, the first red band, the near infrared band, the first shortwave infrared 1 band, and the first shortwave infrared 2 band; determining a second surface albedo according to the second blue band, the second green band, the second red band, the narrow near infrared band, the second shortwave infrared 1 band, and the second shortwave infrared 2 band; Determining a third surface albedo according to the first surface albedo and the second surface albedo; Determining the total primary productivity data of the target area according to the target leaf area index data comprises: The total primary productivity data is determined based on the third surface albedo and the target leaf area index data.

6. The inversion method of total primary productivity according to claim 5, characterized in that: The method further comprises: Acquiring meteorological reanalysis data in the target area, wherein the meteorological reanalysis data includes first minimum temperature data, first solar radiation data, and first saturated air pressure difference data; Resampling the first minimum temperature data, the first solar radiation data, and the first saturated air pressure difference data to obtain second minimum temperature data, second solar radiation data, and second saturated air pressure difference data whose spatial unit area is the second area; Determining the total primary productivity data according to the third surface albedo and the target leaf area index data comprises: The total primary productivity data is determined according to the second minimum temperature data, the second solar radiation data, the second saturated air pressure difference data, the third surface albedo and the target leaf area index data.

7. The inversion method of total primary productivity according to claim 6, characterized in that: The meteorological reanalysis data also includes precipitation data and wind speed data; The resampling of the first minimum temperature data, the first solar radiation data, and the first saturated air pressure difference data to obtain second minimum temperature data, second solar radiation data, and second saturated air pressure difference data having a spatial unit area of ​​the second area includes: The first minimum temperature data, the first solar radiation data and the first saturated air pressure difference data are used as dependent variables in sequence, and the other two data in the meteorological reanalysis data are selected through correlation analysis to establish a first multi-scale geographically weighted regression model; Resampling the first regression parameter, the first regression constant and the first error term of the first multi-scale geographically weighted regression model to obtain a second regression parameter, a second regression constant and a second error term matching the second area; By using the second multi-scale geographically weighted regression model established by using the second regression parameter, the second regression constant and the second error term, the dependent variable is solved to obtain the second minimum temperature data, the second solar radiation data and the second saturated air pressure difference data.

8. A total primary productivity inversion device, characterized in that: include: A data acquisition unit, used for acquiring land purity data, first spectral data, second spectral data and first leaf area index data in a target area, wherein the spatial unit areas of the first spectral data, the land purity data and the first leaf area index data are all first areas, the spatial unit area of ​​the second spectral data is a second area, and the first area is larger than the second area; A data screening unit, configured to screen the first spectrum data and the first leaf area index data according to the land type purity data to obtain third spectrum data and second leaf area index data; A model building unit, used for building a leaf area index inversion model according to the third spectral data and the second leaf area index data; A calculation parameter acquisition unit, used for substituting the second spectral data into the leaf area index inversion model to obtain target leaf area index data; The productivity calculation unit is used to determine the total primary productivity data of the target area according to the target leaf area index data.

9. An electronic device, comprising: A memory and a processor, characterized in that the processor is used to implement the steps of the inversion method of total primary productivity as described in any one of claims 1 to 7 when executing the computer program stored in the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the inversion method of total primary productivity according to any one of claims 1 to 7 are implemented.

Images