WO2023087630A1 - 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法 - Google Patents

一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法 Download PDF

Info

Publication number
WO2023087630A1
WO2023087630A1 PCT/CN2022/090854 CN2022090854W WO2023087630A1 WO 2023087630 A1 WO2023087630 A1 WO 2023087630A1 CN 2022090854 W CN2022090854 W CN 2022090854W WO 2023087630 A1 WO2023087630 A1 WO 2023087630A1
Authority
WO
WIPO (PCT)
Prior art keywords
soil
index
sentinel
salinity
data
Prior art date
Application number
PCT/CN2022/090854
Other languages
English (en)
French (fr)
Inventor
史舟
王楠
薛杰
彭杰
Original Assignee
浙江大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 浙江大学 filed Critical 浙江大学
Publication of WO2023087630A1 publication Critical patent/WO2023087630A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N21/55—Specular reflectivity
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
    • G01N27/02—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance
    • G01N27/04—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance by investigating resistance
    • G01N27/041—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating impedance by investigating resistance of a solid body
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
    • G01N2021/1793—Remote sensing

Definitions

  • the invention belongs to the field of remote sensing inversion, and in particular relates to a method for estimating the soil salinity of straw residual farmland by using remote sensing to construct a characteristic index.
  • Soil salinization includes primary salinization and secondary salinization, especially farmland soil salinization seriously endangers soil health and crop production. Soil evapotranspiration in arid and semi-arid regions is much higher than precipitation, and salt accumulation on the soil surface forms a salt crust. However, unreasonable farmland management (such as flooding with salt) intensifies the secondary salinization process. Quantitative estimation of soil salinization is of great significance for the protection and utilization of cultivated soil resources. Remote sensing technology is widely used in the detection and evaluation of soil salinity because of its wide spatial coverage, high observation spatial resolution, and short time return period.
  • the purpose of the present invention is to solve the problems existing in the prior art, combined with ground survey data, to provide a method for estimating the soil salinity of straw residual farmland soil by using Sentinel-2 satellite data and Sentinel-1 radar data to construct characteristic index.
  • the present invention provides a method for estimating the salinity of straw residual farmland soil by using remote sensing index construction, the steps of which are as follows:
  • the spectral transformation forms include logarithmic transformation, reciprocal transformation and Differential transformation
  • the spectral data set contains the spectral collection of each grid in the image, and the spectral collection includes the original spectrum and a variety of transformed spectra obtained after the spectral transformation
  • the Sentinel-1 radar data is preprocessed, Obtain a radar data set containing radar backscatter coefficients for vertical emission-vertical reception (VV) and vertical emission-horizontal reception (VH);
  • the preprocessing of the multispectral images in the Sentinel-2 satellite data includes radiometric calibration and atmospheric correction, and the model selected for atmospheric correction is FLAASH.
  • the preprocessing of the Sentinel-1 radar data includes orbit correction, thermal noise removal, radiation correction, Lee filtering and Doppler terrain correction, and the water cloud model is selected to remove the influence of vegetation water content on the soil backscatter coefficient
  • the radar backscatter coefficients of vertical emission-vertical reception (VV) and vertical emission-horizontal reception (VH) are converted to obtain a radar data set in decibel (dB) format.
  • the salinity content of the soil surface layer corresponding to each sampling point is obtained by measuring the soil layer of 0-0.2 m on the soil surface by the electrical conductivity method.
  • the spatial resolution of the Sentinel-2 satellite data and the Sentinel-1 radar data is 10 meters.
  • the sensitive band combination is composed of optical remote sensing sensitive band and radar sensitive band, wherein the optical remote sensing sensitive band is the blue (Blue) band, green (Green) band, red (Red) band in the Sentinel-2 satellite data. band, near-infrared (NIR) band, short-wave infrared 1 band (SWIR 1), short-wave infrared 2 band (SWIR 2) and 3 vegetation red edge index (Red Edge 1, Red Edge 2, Red Edge 3) bands, the radar The sensitive bands are the radar backscatter coefficients of vertical emission-vertical reception (VV) and vertical emission-horizontal reception (VH) in Sentinel-1 radar data.
  • the optical remote sensing sensitive band is the blue (Blue) band, green (Green) band, red (Red) band in the Sentinel-2 satellite data. band, near-infrared (NIR) band, short-wave infrared 1 band (SWIR 1), short-wave infrared 2 band (SWIR 2) and 3 vegetation red edge index (Red Edge 1, Red Edge 2,
  • the differential transformation includes first-order differential transformation, second-order differential transformation and third-order differential transformation
  • the spectrum set of each grid in the image contains 6 spectra, which are original spectrum, logarithmic transformation spectrum, reciprocal transformation Spectrum, First Differential Transform Spectrum, Second Differential Transform Spectrum, and Third Differential Transform Spectrum.
  • the two-dimensional, three-dimensional and four-dimensional indices are obtained by combined calculations of bands of corresponding dimensions, and the form of combined calculations is one or more of difference calculation, ratio calculation and power calculation combined.
  • the regression model is a polynomial regression model, and its form is:
  • EC is the salinity content of the soil surface
  • a, b, and c are the regression coefficients, respectively.
  • a high sensitivity index has been screened, which is a new index enhanced straw residual salinity index (Enhanced Residues Soil Salinity Index, ERSSI) that can effectively indicate the soil salinity of straw residual farmland in arid areas
  • the high sensitivity index is a three-dimensional index calculated based on the second-order differential transformation spectrum, and the form of the three-dimensional index is Among them, ERSSI is the high sensitivity index, Green is the green (Green) band of Sentinel-2 satellite data, Blue is the blue (Blue) band of Sentinel-2 satellite data, SWIR 1 is the short-wave infrared 1 band (SWIR) of Sentinel-2 satellite data 1).
  • the present invention further simplifies the method of the first aspect, and proposes the technical scheme of the following second aspect.
  • the present invention provides a method for estimating the salinity of straw residual farmland soil by constructing an index using remote sensing, the steps of which are as follows:
  • ERSSI is the high sensitivity index
  • Green is the green (Green) band of Sentinel-2 satellite data
  • Blue is the blue (Blue) band of Sentinel-2 satellite data
  • SWIR 1 is the short-wave infrared 1 band (SWIR) of Sentinel-2 satellite data 1);
  • the present invention has the following beneficial effects:
  • the invention combines satellite data and ground survey data to provide a method for estimating soil salinity of straw residue farmland by using Sentinel-2 satellite data and Sentinel-1 radar data to construct characteristic index.
  • the method of the present invention estimates the soil salt content of straw residual farmland soil in arid areas, it makes up for the lack of characteristic index in the estimation of straw residual farmland soil salinity, and provides a new method for estimating the straw residual farmland soil salt content in arid areas , which is conducive to the formulation of large-scale farmland soil salinity control and improvement policies, and has certain theoretical and practical significance and application value.
  • the present invention obtained a new index, the Enhanced Residues Soil Salinity Index (ERSSI), which effectively indicates the soil salinity of straw residues in arid areas, and constructed the new index ERSSI based on the Sentinel-2 satellite remote sensing data and the new index ERSSI.
  • the method of polynomial linear regression inversion of soil salinity content in straw residue farmland in arid areas can finally obtain high spatial resolution and high-quality spatial variation results of soil salinity content with a spatial resolution of 10 meters.
  • Figure 1 is a scatter plot of the performance of the estimated soil salinity content in Q1 (modeling set) and Q2 (independent validation set) using the Enhanced Residues Soil Salinity Index (ERSSI) and polynomial regression;
  • EDSI Enhanced Residues Soil Salinity Index
  • Fig. 2 is a distribution map of the salt content of farmland soil with residues of straw in the arid area of southern Xinjiang in 2020 estimated by this embodiment.
  • a method for estimating the soil salinity of straw residue farmland by using remote sensing construction index is provided, and the method is used for estimating the soil salinity of the area to be measured whose land type is straw residue farmland in arid area .
  • the concrete steps of this method are as follows:
  • the Sentinel-2 satellite data is a kind of optical remote sensing image data
  • the Sentinel-1 radar data is a kind of radar data
  • the soil sampling data set is a kind of ground survey data.
  • the three sources are different and belong to the combination of multi-source data.
  • the three types of data, Sentinel-2 satellite data, Sentinel-1 radar data, and soil sampling data sets need to be collected synchronously at the time to be estimated. For example, if the time to be estimated is a certain period, then Sentinel-2 satellite data, Sentinel-1 radar data, and soil sampling datasets all need to be data collected synchronously during this period.
  • a certain time deviation is allowed in the estimated time of different data sources in the present invention.
  • Sentinel-2 satellite data and Sentinel-1 radar data can be downloaded through the corresponding satellite data publishers.
  • Sentinel-2 satellite data comes from Sentinel-2 satellite, which carries a multispectral imager (MSI) with an altitude of 786km and can cover 13 spectral bands.
  • Sentinel-1 radar data is derived from the Sentinel-1 satellite, consisting of two satellites carrying a C-band synthetic aperture radar that provides continuous imagery (day, night and all weather).
  • the spatial resolution of the Sentinel-2 satellite data and Sentinel-1 radar data used in this embodiment is 10 meters.
  • the soil sampling data set is obtained through ground investigation in the area to be measured. The ground survey should arrange sampling points in the area to be measured, and collect corresponding soil surface samples (0-0.2m soil surface) at each sampling point. layer), and then measure the soil salinity content of the soil surface samples, so that the soil sampling data set will eventually contain the soil surface salinity content corresponding to different sampling points in the area to be tested.
  • the salinity content of the soil surface layer corresponding to each sampling point in the soil sampling data set is obtained by measuring the soil surface layer sample by the electrical conductivity method.
  • the measurement process of the conductivity method is as follows: the soil sample is air-dried, ground and passed through a 2 mm sieve, and the soil leachate is obtained at a soil-water ratio of 1:5, and the supernatant is filtered to measure its conductivity EC 1:5 , and the conductivity of the soil sample is obtained.
  • Rate data as the representative value of the salinity content of the soil surface.
  • the electrical conductivity can be directly used to characterize the soil salinization level of the area to be tested.
  • the salt content of the soil surface corresponding to the sampling point in the soil sampling data set can be represented by the electrical conductivity EC 1:5 , or the absolute content can be represented by the soluble salt content c of the soil surface, which can be selected according to actual needs.
  • S2 For the fusion and processing of multi-source data, the Sentinel-2 satellite data and Sentinel-1 radar data obtained in S1 are preprocessed and calculated respectively, and the specific steps are as follows:
  • S21 Perform preprocessing and calculation on the multispectral image (MSI image) in the Sentinel-2 satellite data to obtain surface reflectance data in different bands, wherein the preprocessing includes radiometric calibration and atmospheric correction, and the model selected for atmospheric correction for FLAASH.
  • the Sentinel-1 radar data is also preprocessed, including orbit correction, thermal noise removal, radiation correction, Lee filtering and Doppler terrain correction, and the water cloud model is used to remove the influence of vegetation water content on the soil backscatter coefficient. Influence, convert the radar backscatter coefficients of vertical emission-vertical reception (VV) and vertical emission-horizontal reception (VH) to obtain a radar data set in decibel (dB) format, and obtain surface reflection characteristic data.
  • VV vertical emission-vertical reception
  • VH vertical emission-horizontal reception
  • S22 Perform spectral transformation on each band of the preprocessed Sentinel-2 satellite data to obtain a spectral data set after spectral transformation.
  • the spectral transformation in this step adopts a band-by-band calculation method, and the spectral transformation forms include three types: logarithmic transformation, reciprocal transformation and differential transformation, and differential transformation includes first-order differential transformation, second-order differential transformation and third-order differential transformation.
  • the logarithmic transformation, reciprocal transformation, first-order differential transformation, second-order differential transformation, and third-order differential transformation are performed on the Sentinel-2 image band by band to obtain the original spectral curve and the spectral curve transformed by five mathematical calculations.
  • the spectral set of each grid in the final image contains 6 spectra, which are the original spectrum, the logarithmic transformation spectrum, the reciprocal transformation spectrum, the first-order differential transformation spectrum, the second-order differential transformation spectrum and the third-order differential transformation spectrum.
  • the sensitive band combination of soil surface salinity refers to the set of all bands related to soil salinity, and which bands are related to soil salinity can be determined based on correlation analysis, and can also be determined through relevant references and historical research. Since the remote sensing data in the present invention includes Sentinel-2 satellite data and Sentinel-1 radar data, the combination of sensitive bands of soil surface salinity also needs to be composed of optical remote sensing sensitive bands and radar sensitive bands.
  • the present invention adopts an inductive statistical method to screen sensitive bands, and summarizes the existing salinity index, For the bands used by vegetation index and soil related index, blue, green, red, near-infrared, short-wave infrared, and vegetation red-edge index are selected as sensitive bands in optical remote sensing data, and a total of 9 optical sensitive bands are obtained; in radar data The radar backscatter coefficients VV and VH are screened out as sensitive bands, and a total of 2 radar sensitive bands are obtained.
  • 11 bands sensitive to soil salinity are obtained to form a sensitive band combination, wherein the optical remote sensing sensitive bands are the blue band (Blue, wavelength 458-523nm) and green band (Green, wavelength 458-523nm) in the Sentinel-2 satellite data.
  • the optical remote sensing sensitive bands are the blue band (Blue, wavelength 458-523nm) and green band (Green, wavelength 458-523nm) in the Sentinel-2 satellite data.
  • red band (Red, wavelength 650-680nm), near-infrared band (NIR, wavelength 785-900nm), short-wave infrared 1 band (SWIR 1, wavelength 1565-1655nm), short-wave infrared 2 band (SWIR 2, wavelength 2100-2280nm) and three vegetation red edge index bands (Red Edge, the wavelengths are 698-713nm, 733-748nm, 773-793nm), and the radar sensitive band is the vertical emission-vertical reception ( VV) and vertical transmit-horizontal receive (VH) radar backscatter coefficients.
  • VV vertical emission-vertical reception
  • VH vertical transmit-horizontal receive
  • the N-dimensional index refers to the index constructed with N sensitive bands.
  • the form of the index is not limited, and different types of calculation formulas can be used.
  • the formula form of combined calculation can be one of difference calculation, ratio calculation and power calculation.
  • the difference calculation form is (AB)
  • the ratio calculation form is (A/B)
  • the power calculation form is (A k ).
  • a and B are each a sensitive band, and k is a constant.
  • the calculation formula form of combined calculation can also be a combination of two or three of difference calculation, ratio calculation and power calculation, that is, to combine difference calculation, ratio calculation and power calculation to build a more complex calculation formula, such as (AB)/C, [(AB)/(A+B)] k , [(AB)/(A+C)] k and so on, C is also a sensitive band.
  • N-dimensional index the N sensitive bands used can be traversed and sampled from 11 sensitive bands. Therefore, in the multidimensional index set of the present invention, there are two distinguishing dimensions of exponential form and sensitive band. There are many variants of the index because of the different sensitive bands selected.
  • Each N-dimensional index can further derive a series of indices by selecting different N sensitive bands. These indices with different forms or different sensitive bands are included in the multi-dimensional index. collection and assign a unique ID as an index for subsequent filtering.
  • the index screening is carried out to find the index most relevant to the soil salinity content of straw residue farmland. Since each index in the multidimensional index set is calculated based on one or more bands, the band values of the bands need to be obtained from the spectral curve. In the aforementioned step S22, a spectral data set consisting of 6 spectra is obtained. Therefore, when screening the most sensitive index, each index in the above-mentioned multidimensional index set needs to be sampled from 6 spectra respectively.
  • the spectral set corresponding to each sampling point can be obtained from the spectral data set of the aforementioned S22.
  • Each spectrum calculates the value of each index in the aforementioned multidimensional index set; then each index under each spectrum is used as the index to be screened, and the soil surface salinity content of each soil sampling point in the above soil sampling data set is compared with Correlation analysis of different indexes to be screened. Different indices to be screened may have different sampled spectra, or may have different index IDs.
  • the soil surface salinity content of each soil sampling point in the soil sampling data set is used as the first data sequence, and a certain index calculated based on a certain spectrum of each soil sampling point in the soil sampling data set is used as For the second data series, calculate the correlation index of the two data series. Finally, the index to be screened with the highest correlation is selected as the highly sensitive index.
  • each sampling point in the soil sampling data set can be used as a regression sample, with the high-sensitivity index at the sampling point as the independent variable, and the salinity content of the soil surface at the sampling point as the dependent variable. Analyze and build regression models.
  • the regression model adopted is a polynomial regression model, and its form is:
  • EC is the salinity content of the soil surface
  • a, b, and c are the regression coefficients, respectively.
  • the value of the high sensitivity index corresponding to each grid in the area to be measured is obtained from the spectral data set and radar data set obtained in the aforementioned S2, and input into the constructed regression model to calculate each
  • a highly sensitive index is selected, which is a three-dimensional index calculated based on the second-order differential transformation spectrum, and the form of the three-dimensional index is Among them, ERSSI is the high sensitivity index, Green is the green (Green) band of Sentinel-2 satellite data, Blue is the blue (Blue) band of Sentinel-2 satellite data, SWIR 1 is the short-wave infrared 1 band (SWIR1) of Sentinel-2 satellite data ), the Green, Blue, and SWIR 1 band values all need to be sampled from the second-order differential transform spectrum.
  • the constructed multi-dimensional index sets are different, other highly sensitive indices may also be screened out, and there is no limit to this.
  • the present invention further provides a method for estimating the salinity of straw residual farmland soil by using remote sensing construction index, the steps are as follows:
  • the Sentinel-2 satellite data corresponding to the area to be measured at the time to be estimated is straw residue farmland in an arid area, that is, farmland located in an arid area with straw remaining on the surface.
  • the method of using the remote sensing construction index to estimate the soil salinity of straw residual farmland soil in steps S1 to S4 in the above embodiment is applied to a specific case in order to demonstrate specific technical effects.
  • the cotton planting fields (81°17'-81°22'E, 40°28'-40°31'N) in the Xinjiang Uygur Autonomous Region were selected as the area to be tested, and the soil samples obtained from the ground survey on November 1, 2020 were used.
  • the salinity content of the 0.2-meter soil obtained from the point data was used as the dependent variable
  • the single-scene Sentinel-2 satellite remote sensing image data on November 1, 2020 and the Sentinel-1 data on November 2, 2020 were used as independent variables, based on spectral transformation and Two-dimensional, three-dimensional and four-dimensional indexes were constructed to obtain the Enhanced Residues Soil Salinity Index (ERSSI), which is highly sensitive to the soil salinity of straw residues in arid areas.
  • the linear polynomial regression model was used to estimate the soil salt content distribution map of the straw residue farmland in the area to be tested.
  • the index and its estimation method are as follows:
  • Step 1) Data Acquisition: Taking straw residual farmland in arid areas as the estimation object, according to the time to be estimated, obtain the Sentinel-2 satellite data, Sentinel-1 radar data, and soil surface salt content data sets in the same period of the area to be measured; Sentinel- 2 Satellite data and Sentinel-1 radar data have a spatial resolution of 10 meters.
  • Sentinel-2 data is an L1 product released by the European Space Agency (ESA).
  • Sentinel-2 consists of two polar-orbiting satellites (Sentinel-2A and Sentinel-2B). The data of the two satellites are complementary.
  • the revisit period is 5 days, the orbital period is 100 minutes, and the orbital altitude is 786 kilometers.
  • the scan width is 290 kilometers, and the orbital inclination is 98.62°.
  • the multispectral scanning imaging data (MSI) provides a band with a spatial resolution of 60m, three bands of 10m and two bands of 30m in the visible light to the red edge region.
  • Sentinel-2B data can be downloaded for free from the US Geological Survey (USGS), The steps used Sentinel-2B data on November 1, 2020; Sentinel-1 satellite is an Earth observation satellite launched by the European Space Agency's Copernicus Program (GMES), consisting of two satellites (Sentinel-1A and Sentinel-1B ), carrying a C-band synthetic aperture radar, allowing all-sky imagery to be provided with a temporal resolution of 6 days.
  • GMES European Space Agency's Copernicus Program
  • This step uses the Sentinel-1A data from November 2, 2020, because the images from this period are close to the date of the field experiment, and the soil and canopy conditions are almost the same.
  • Sentinel-1A data were acquired in ascending direction and interferometric width (IW) mode with two dual polarizations and a spatial resolution of 10 m.
  • IW interferometric width
  • the data set of salinity content in the soil surface is obtained by sampling soil samples at different points in the 0-0.2m soil layer and transporting them back to the laboratory for analysis.
  • the obtained soil samples were measured by the conductivity method: the soil samples were air-dried, ground and sieved through a 2 mm sieve, and the soil leaching solution was obtained with a soil-water ratio of 1:5, and the supernatant was filtered to measure its conductivity, and the soil sample was obtained.
  • Conductivity data used to characterize the level of soil salinization in the area to be tested.
  • the soil sample point data is divided into two regions by using the geographical partition method, and divided into two sub-regions Q1 and Q2 according to the geographical location of the sample point distribution.
  • the conductivity data set of Q1 is used as a modeling set for the new index. Construction, correlation evaluation, and construction of polynomial regression equations, the conductivity data set of Q2 was used as an independent validation set to independently verify the validity of the exponential and polynomial regression equations.
  • Step 2) Data preprocessing: Based on the fusion and processing of multi-source data, the optical remote sensing image data, radar data and soil sample data obtained in step 1) are preprocessed and calculated.
  • Preprocess the Sentinel-2 remote sensing data obtained in step 1) use the Sentinel Application Platform (SNAP) module in the Sentinel Application Platform (SNAP) package to perform radiometric calibration and atmospheric correction on the Sentinel-2 data, and convert the MSI image into a surface reflectance format for output.
  • SNAP Sentinel Application Platform
  • the Sentinel-1 radar data obtained in step 1) are preprocessed: orbit correction, thermal noise removal, radiation correction, Lee filtering and Doppler terrain correction are performed in sequence, and the water cloud model is used to remove the influence of vegetation water content on soil backward Influence of scattering coefficient, the radar backscatter coefficients of vertical emission-vertical reception (VV) and vertical emission-horizontal reception (VH) are converted to obtain a radar data set in decibel (dB) format, and surface reflectance data are obtained.
  • VV vertical emission-vertical reception
  • VH vertical emission-horizontal reception
  • Each band of the preprocessed Sentinel-2 satellite data is spectrally transformed using the band-by-band calculation method, and logarithmic transformation, reciprocal transformation, first-order differential transformation, second-order differential transformation, and third-order differential transformation are performed on each band.
  • a total of 6 mathematics Calculate the spectral transformation to obtain the original spectral curve and the spectral curves transformed by five mathematical calculations.
  • the 6 spectra of each grid are used as the corresponding spectral set of the grid, and the spectral sets of all grids are used as the spectral data set.
  • Step 3) Construction of the new index Different bands in the sensitive band combination of soil surface salinity are traversed and combined into different two-dimensional, three-dimensional and four-dimensional indices to obtain a multi-dimensional index set.
  • the combination of sensitive bands sensitive to soil salinity is screened by the inductive statistical method, and the existing The bands used for the salinity index, vegetation index and soil related index are selected from the optical remote sensing data to obtain blue, green, red, near-infrared, short-wave infrared, and vegetation red-edge index as sensitive bands.
  • Sentinel-2 original spectrum and five The above bands are extracted from the spectrum transformed by mathematical calculations to obtain 9 optical sensitive bands; the radar backscattering coefficients VV and VH of the soil are screened from the radar data as the sensitive bands, and 2 radar bands are calculated from the Sentinel-1 data Sensitive band. A total of 11 bands sensitive to soil salinity were obtained.
  • the new index is constructed by inductive, adding and subtracting dimension methods.
  • B i , B j , B k , and B h are any four of the 11 bands of blue, green, red, near-infrared, short-wave infrared, vegetation red-edge index, and radar backscatter coefficients VV and VH, respectively.
  • Each index form needs to traverse the entire set of sensitive bands, and select different bands to form different indices and add them to the multidimensional index set for subsequent screening.
  • the spectral set corresponding to each sampling point is obtained from the spectral data set, based on each spectrum in the spectral set (respectively original spectrum, logarithmic transformation spectrum, reciprocal transformation spectrum, one First-order differential transformation spectrum, second-order differential transformation spectrum and third-order differential transformation spectrum) respectively calculate the value of each index in the multidimensional index set.
  • each index under each spectrum is used as the index to be screened, and the soil surface salt content of each soil sampling point in the soil sampling data set is correlated with different indexes to be screened, and at least one with the highest correlation is screened out.
  • the index to be screened is regarded as a highly sensitive index.
  • the correlation analysis is used to evaluate the correlation of different indices under different spectra (six groups in total, 27 index forms in each group) and the soil conductivity data set in Q1 (as shown in Table 2), and at the same time in Q2
  • the soil conductivity dataset was independently validated (shown in Table 3) and yielded the most correlated indices.
  • Table 3 The highest correlation (r) of 2D, 3D, and 4D indices constructed from six forms of spectra in Q2
  • the index with the highest correlation to the soil salinity of straw-covered farmland in arid areas is a three-dimensional index constructed after second-order derivative transformation, named Enhanced Residues Soil Salinity Index (Enhanced Residues Soil Salinity Index, ERSSI), the new index is calculated as:
  • ERSSI is the high sensitivity index
  • Green is the green (Green) band of Sentinel-2 satellite data
  • Blue is the blue (Blue) band of Sentinel-2 satellite data
  • SWIR 1 is the short-wave infrared 1 band (SWIR) of Sentinel-2 satellite data -1)
  • Green, Blue, and SWIR 1 band values all need to be sampled from the second-order differential transform spectrum.
  • Fig. 1 is the scatter diagram of the performance of the estimated value of soil salinity using ERSSI and polynomial regression in Q1 (modeling set) and Q2 (independent verification set), the surface salt data estimated by the present invention is compared with the simulated value of ground actual value. High degree of cooperation.
  • ERSSI combines the green, blue, and shortwave infrared bands in ratios.
  • the green band is sensitive to vegetation type and less sensitive to atmospheric effects, and the blue band is used to solve the problem of vegetation index decay caused by residual aerosol after initial atmospheric correction.
  • the combination of blue and green bands in ERSSI can effectively deal with interference. It can be seen from the evaluation results that the coverage type of the area to be tested is straw residue, and the salt intrusion state includes salt crust.
  • the conventional vegetation sensitive band is not suitable for straw residue farmland in arid areas.
  • SWIR 1 is used instead of NIR, which can effectively identify the cellular structure and soil water content of vegetation in water-deficient environments, and the combination of green and SWIR 1 bands can characterize iron content and soil moisture content and soil conditions , thus indirectly indicating salinization.
  • the soil salinization in the area to be measured mainly contains calcium ions and sulfate ions, both of which reflect strongly in the range of 1443-1745nm, and the SWIR 1 band is sensitive to these substances.
  • the use of SWIR 1 band provides a basis for monitoring soil salinization of straw residue farmland in Xinjiang. opportunity.
  • Step 4) Estimation of farmland soil salinity with straw residues in arid areas: According to the correlation evaluation results obtained in step 3), the Enhanced Residues Soil Salinity Index (Enhanced Residues Soil Salinity Index) calculated by the second derivative transformation of Sentinel-2 remote sensing images , ERSSI) as the independent variable, and the soil conductivity data set in Q1 obtained in step 2) as the dependent variable, construct a linear polynomial regression equation in R language:
  • EC is the soil surface salinity content (dS/m) of straw residue farmland in arid area.
  • the polynomial regression equation obtained in step 2) in Q1 has an accuracy R2 of 0.63; at the same time, the enhanced straw residual salt index (ERSSI) and polynomial regression equation obtained in step 2) Q2 (independent verification Set) for independent verification, the verification accuracy R 2 is 0.64;
  • the ERSSI index is calculated for each grid point in the area to be tested, and the constructed polynomial regression model is applied to the area to be tested, and the obtained The pixel-by-pixel soil salinity content of the area to be measured.
  • the cotton planting field in the Xinjiang Uygur Autonomous Region was selected as the area to be tested, and the distribution map of the soil salinity content in 2020 estimated by this implementation method is shown in Figure 2, with a spatial resolution of 10 meters. It should be noted that this soil salt content distribution map is a relative salt content distribution map characterized by electrical conductivity. The conversion formula can be converted.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Chemical & Material Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Pathology (AREA)
  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Electrochemistry (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Image Processing (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法。联合卫星数据和地面调查数据，提供了利用Sentinel-2卫星数据和Sentinel-1雷达数据构建特征指数用于估算秸秆残留农田土壤盐分的方法。通过对干旱区秸秆残留农田土壤盐分含量进行估算后，可得到大空间尺度的干旱区秸秆残留农田土壤盐分含量分布图，弥补了在秸秆残留农田土壤盐分估算中缺乏特征指数的短板，为干旱区秸秆残留农田土壤盐分含量的估算提供了新方法，有利于大区域尺度农田土壤盐分的治理改良政策的制定，具有一定的理论、实践意义和推广应用价值。

Description

一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法 技术领域
本发明属于遥感反演领域，具体涉及一种利用遥感构建特征指数用于估算秸秆残留农田土壤盐分的方法。
背景技术
土壤盐渍化包括原生盐渍化和次生盐渍化，尤其是农田土壤盐渍化严重危害了土壤健康和作物生产。干旱和半干旱地区土壤蒸散量远高于降水量，土壤表面盐分积累结成盐壳，然而，不合理的农田管理(如漫灌洗盐)加剧了二次盐渍化进程。土壤盐渍化的定量估算对于耕地土壤资源的保护和利用具有重要意义。遥感技术因其空间范围覆盖广、观测空间分辨率高、时间重返周期性短的特点，被广泛用于土壤盐分的检测与评估。
在利用遥感手段定量反演土壤盐分中，针对不同的土壤盐渍化类型，以往的研究构建并使用了多个盐分指数、植被指数、遥感波段等用于建模回归。然而，因为地理环境的不同，气候类型、地貌特征、盐渍化类型、土壤类型、植被类型、社会经济要素等本地化特征明显，在应用到不同地区时，不同指数的通用性表现出一定的差异。特别是对于存在秸秆残留的农田土壤而言，此类农田土壤由于表面存在秸秆覆盖等不利因素，因此其在遥感反演盐分含量时缺乏特征指数的问题。因此，针对干旱区秸秆残留农田土壤盐渍化的监测，需要构建并优选一系列指数用于指示农田土壤表层的盐分，从而实现对土壤盐分的高精度定量反演与可视化，为干旱区秸秆残留农田土壤盐分含量的估算提供了新方法
发明内容
本发明的目的在于解决现有技术中存在的问题，联合地面调查数据，提供一种利用Sentinel-2卫星数据和Sentinel-1雷达数据构建特征指数用于估算秸秆残留农田土壤盐分的方法。
为了实现上述发明目的，本发明采用的具体技术方案如下：
第一方面，本发明提供了一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其步骤如下：
S1、获取待测区域在待估算时间下对应的Sentinel-2卫星数据、Sentinel-1雷达数据、土壤采样数据集；所述待测区域的土地类型为干旱区秸秆残留农田；所述土壤采样数据集中包含待测区域内不同采样点对应的土壤表层盐分含量；
S2、对Sentinel-2卫星数据中的多光谱影像进行预处理，得到不同波段的表面反射率数据，再通过多种光谱变换得到光谱数据集；所述光谱变换形式包括对数变换、倒数变换和微分变换；所述光谱数据集中包含影像中每一个栅格的光谱集合，所述光谱集合包括原始光谱和经过所述光谱变换后得到的多种变换光谱；对Sentinel-1雷达数据进行预处理，得到包含垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数的雷达数据集；
S3、遍历土壤表层盐分的敏感波段组合中不同的波段，并将其组合成不同的二维、三维和四维指数，得到多维指数集合；再针对所述土壤采样数据集中的每个采样点，从所述光谱数据集中获取每个采样点对应的光谱集合，基于光谱集合中的每一条光谱分别计算所述多维指数集合中每一种指数的值；然后以每一条光谱下的每一种指数作为待筛选指数，将所述土壤采样数据集中各土壤采样点的土壤表层盐分含量分别与不同的待筛选指数进行相关性分析，筛选出相关性最高的至少一个待筛选指数作为高敏感指数；
S4、以所述土壤采样数据集中的各采样点为回归样本，以所述高敏感指数作为自变量，以土壤表层盐分含量作为因变量，建立回归模型；并从所述光谱数据集和所述雷达数据集中获取待测区域内每一个栅格对应的所述高敏感指数的值，根据所述回归模型计算得到每一个栅格处的土壤表层盐分含量，最终形成待测区域内秸秆残留农田土壤盐分含量的空间分布图。
作为优选，所述Sentinel-2卫星数据中的多光谱影像的预处理包括辐射定标和大气校正，且其中大气校正所选用的模型为FLAASH。
作为优选，所述Sentinel-1雷达数据的预处理包括轨道校正、热噪声去除、辐射校正、Lee滤波和多普勒地形校正，并选用水云模型去除植被含水量对土壤后向散射系数的影响后，将垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数转换并得到分贝(dB)格式的雷达数据集。
作为优选，所述土壤采样数据集中，每一个采样点对应的土壤表层盐分含量由土壤表面0～0.2m的土层通过电导率法测定得到。
作为优选，所述Sentinel-2卫星数据和Sentinel-1雷达数据的空间分辨率为10米。
作为优选，所述敏感波段组合由光学遥感敏感波段和雷达敏感波段组成，其中所述光学遥感敏感波段为Sentinel-2卫星数据中的蓝(Blue)波段、绿(Green)波段、红(Red)波段、近红外(NIR)波段、短波红外1波段(SWIR 1)、短波红外2波段(SWIR 2)和3个植被红边指数(Red Edge 1，Red Edge 2，Red Edge3)波段，所述雷达敏感波段为Sentinel-1雷达数据中垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数，。
作为优选，所述微分变换包括一阶微分变换、二阶微分变换和三阶微分变换，影像中每一个栅格的光谱集合均包含6条光谱，分别为原始光谱、对数变换光谱、倒数变换光谱、一阶微分变换光谱、二阶微分变换光谱和三阶微分变换光谱。
作为优选，所述多维指数集合中，二维、三维和四维指数均由相应维数的波段通过组合计算得到，组合计算的形式为差值计算、比值计算和幂计算中的一种或多种结合。
作为优选，所述回归模型为多项式回归模型，其形式为：
EC＝a×ERSSI 2+b×ERSSI+c
式中：EC为土壤表层盐分含量，a、b、c分别为回归系数。
需说明的是，上述第一方面中所提供的各技术方案，可以应用于任意的待测区域和时间，只要多维指数集合中所涵盖的指数可选范围足够广泛，一般都可以筛选得到高敏感指数并最终构建得到回归模型。
在本发明中，通过后续实施例，筛选到了一种高敏感指数，该指数是能有效指示干旱区秸秆残留农田土壤盐分的新指数增强型秸秆残留盐分指数(Enhanced Residues Soil Salinity Index，ERSSI)，该高敏感指数为基于二阶微分变换光谱计算的三维指数，三维指数形式为
Figure PCTCN2022090854-appb-000001
其中ERSSI为高敏感指数，Green为Sentinel-2卫星数据的绿(Green)波段，Blue为Sentinel-2卫星数据的蓝(Blue)波段，SWIR 1为Sentinel-2卫星数据的短波红外1波段(SWIR 1)。
基于这种高敏感指数，本发明进一步对第一方面的方法进行了简化，提出了 下述第二方面的技术方案。
第二方面，本发明提供了一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其步骤如下：
1)、获取待测区域在待估算时间下对应的Sentinel-2卫星数据；所述待测区域的土地类型为干旱区秸秆残留农田；
2)、对Sentinel-2卫星数据中的多光谱影像进行预处理，得到不同波段的表面反射率数据，再对每个栅格的光谱进行二阶微分变换得到二阶微分变换光谱；
3)、基于待测区域内每个栅格的二阶微分变换光谱计算对应的三维指数，三维指数形式为
Figure PCTCN2022090854-appb-000002
其中ERSSI为高敏感指数，Green为Sentinel-2卫星数据的绿(Green)波段，Blue为Sentinel-2卫星数据的蓝(Blue)波段，SWIR 1为Sentinel-2卫星数据的短波红外1波段(SWIR 1)；
4)、将待测区域内每个栅格的三维指数作为自变量，基于预先针对待测区域构建的回归模型估计每个栅格对应的土壤表层盐分含量，最终形成待测区域内秸秆残留农田土壤盐分含量的空间分布图。
本发明相对于现有技术而言，具有以下有益效果：
本发明联合卫星数据和地面调查数据，提供了一种利用Sentinel-2卫星数据和Sentinel-1雷达数据构建特征指数用于估算秸秆残留农田土壤盐分的方法。通过本发明的方法对干旱区秸秆残留农田土壤盐分含量进行估算后，弥补了在秸秆残留农田土壤盐分估算中缺乏特征指数的短板，为干旱区秸秆残留农田土壤盐分含量的估算提供了新方法，有利于大区域尺度农田土壤盐分的治理改良政策的制定，具有一定的理论、实践意义和推广应用价值。
另外，本发明筛选得到了一种有效指示干旱区秸秆残留农田土壤盐分的新指数增强型秸秆残留盐分指数(Enhanced Residues Soil Salinity Index，ERSSI)，以Sentinel-2卫星遥感数据和新指数ERSSI构建了多项式线性回归反演干旱区秸秆残留农田的土壤盐分含量的方法，最终可以得到空间分辨率为10米的高空间分辨率、高质量的土壤盐分含量空间变异结果。
附图说明
图1为使用增强型秸秆残留盐分指数(Enhanced Residues Soil Salinity Index，ERSSI)和多项式回归的土壤盐分含量估算值在Q1(建模集)和Q2(独立验证集) 中表现的散点图；
图2为本实施方式估算得到的2020年南疆干旱区秸秆残留农田土壤的盐分含量分布图。
具体实施方式
下面结合附图和具体实施方式对本发明做进一步阐述和说明。本发明中各个实施方式的技术特征在没有相互冲突的前提下，均可进行相应组合。
在本发明的一个较佳实施例中，提供了一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法，该方法用于对土地类型为干旱区秸秆残留农田的待测区域进行土壤盐分的估算。该方法的具体步骤如下：
S1、获取待测区域在待估算时间下对应的Sentinel-2卫星数据、Sentinel-1雷达数据、土壤采样数据集。
其中Sentinel-2卫星数据是一种光学遥感影像数据，Sentinel-1雷达数据是一种雷达数据，土壤采样数据集是一种地面调查数据，三者来源不同，属于多源数据组合。为了保证估算的准确性，Sentinel-2卫星数据、Sentinel-1雷达数据、土壤采样数据集这三类数据均需要是在待估算时间下同步采集的。例如，待估算时间是某一个时期，那么Sentinel-2卫星数据、Sentinel-1雷达数据、土壤采样数据集均需要是这一个时期同步采集的数据。但由于土壤盐分含量以及冠层信息在短时间内的相对稳定性，因此本发明中不同数据源的待估算时间允许存在一定的时间偏差。
Sentinel-2卫星数据、Sentinel-1雷达数据均可通过相应的卫星数据发布方处下载。Sentinel-2卫星数据来源于哨兵2号卫星，该卫星携带一枚多光谱成像仪(MSI)，高度为786km，可覆盖13个光谱波段。Sentinel-1雷达数据来源于哨兵1号卫星，由两颗卫星组成，载有C波段合成孔径雷达，可提供连续图像(白天、夜晚和各种天气)。本实施例中采用的Sentinel-2卫星数据和Sentinel-1雷达数据的空间分辨率为10米。而土壤采样数据集是通过对待测区域进行地面调查获得的，地面调查应当在待测区域内进行采样点的布设，并在每个采样点处采集相应的土壤表层样品(0～0.2m的土层)，然后对土壤表层样品测定其土壤盐分含量，由此土壤采样数据集中最终将包含待测区域内不同采样点对应的土壤表层盐分含量。
在本实施例中，土壤采样数据集中每一个采样点对应的土壤表层盐分含量由土壤表层样品通过电导率法测定得到。电导率法的测定过程如下：将土壤样品风干、研磨并过2毫米筛，以土水比1：5得到土壤浸出液，过滤取上清液测定其电导率EC 1:5，得到土壤样品的电导率数据，作为土壤表层盐分含量的表征值。
需说明的是，由于土壤样品的电导率与土壤表层盐分含量存在正相关关系，因此本实施例中可以直接用电导率来表征待测区域的土壤盐渍化水平。但是在其他实施例中，也可以根据两者的相关关系进行换算后得到土壤表层盐分含量的绝对值，土壤表层可溶性盐含量c的换公式为c(g/kg)＝0.0275*EC 1:5-0.0573，其中EC 1:5是指土壤的电导率，单位为mS/m。因此，土壤采样数据集中采样点对应的土壤表层盐分含量既可以用电导率EC 1:5代表相对含量，也可以用土壤表层可溶性盐含量c来代表绝对含量，具体可根据实际需要进行选择。
S2：对多源数据的融合和处理，将S1获取的Sentinel-2卫星数据、Sentinel-1雷达数据分别进行预处理和计算，具体依次进行如下步骤：
S21：对Sentinel-2卫星数据中的多光谱影像(MSI图像)进行预处理计算，得到不同波段的表面反射率数据，其中预处理包括辐射定标和大气校正，且其中大气校正所选用的模型为FLAASH。另外，也对Sentinel-1雷达数据进行预处理，预处理包括轨道校正、热噪声去除、辐射校正、Lee滤波和多普勒地形校正，应用水云模型去除植被含水量对土壤后向散射系数的影响，将垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数转换并得到分贝(dB)格式的雷达数据集，得到地表反射特征数据。
S22：对经过预处理的Sentinel-2卫星数据的各个波段进行光谱变换，得到光谱变换后的光谱数据集。本步骤中的光谱变换采用逐波段计算方法，而光谱变换形式包括对数变换、倒数变换和微分变换三类，微分变换又包含一阶微分变换、二阶微分变换和三阶微分变换。由此，对Sentinel-2影像逐个波段地进行对数变换、倒数变换、一阶微分变换、二阶微分变换和三阶微分变换，得到原始光谱曲线以及五种数学计算变换后的光谱曲线。因此，最终影像中每一个栅格的光谱集合均包含6条光谱，分别为原始光谱、对数变换光谱、倒数变换光谱、一阶微分变换光谱、二阶微分变换光谱和三阶微分变换光谱。
S3、遍历土壤表层盐分的敏感波段组合中不同的波段，并将其组合成不同的 二维、三维和四维指数，得到多维指数集合。
其中土壤表层盐分的敏感波段组合是指与土壤盐分相关的所有波段组成的集合，其中哪些波段与土壤盐分相关可以根据相关性分析确定，亦可通过相关的参考文献和历史研究进行确定。由于本发明中遥感数据包含了Sentinel-2卫星数据和Sentinel-1雷达数据，因此土壤表层盐分的敏感波段组合也需要由光学遥感敏感波段和雷达敏感波段组成。本发明在历史研究报道的基础上，采用归纳统计方法进行敏感波段筛选，根据影响土壤盐分监测精度的土壤、作物、水分因子，以及盐渍土的光谱反射特性，总结归纳现有的盐分指数、植被指数和土壤相关指数等使用的波段，在光学遥感数据中筛选得到蓝、绿、红、近红外、短波红外、植被红边指数为敏感波段，一共得到9个光学敏感波段；在雷达数据中筛选得到雷达后向散射系数VV、VH为敏感波段，一共得到2个雷达敏感波段。本实施例中，共得到11个对土壤盐分敏感的波段构成敏感波段组合，其中光学遥感敏感波段为Sentinel-2卫星数据中的蓝波段(Blue，波长458-523nm)、绿波段(Green，波长543-578nm)、红波段(Red，波长650-680nm)、近红外波段(NIR，波长785-900nm)、短波红外1波段(SWIR 1，波长1565-1655nm)、短波红外2波段(SWIR 2，波长2100-2280nm)和3个植被红边指数波段(Red Edge，波长分别为698-713nm、733-748nm、773-793nm)，而雷达敏感波段为Sentinel-1雷达数据中垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数。
当得到上述包含11个敏感波段的敏感波段组合后，即可将这些敏感波段组合成多维的指数。N维指数是指用N个敏感波段构建的指数，指数的形式不限，可以采用不同类型的计算式。在本实施例的多维指数集合中，指数的维数包括二维、三维和四维三种，分别由N＝3个、N＝4个、N＝5个敏感波段通过计算式组合计算得到。组合计算的计算式形式可以为差值计算、比值计算和幂计算中的一种，差值计算形式为(A-B)，比值计算形式为(A/B)，幂计算形式为(A k)，A和B各自为一个敏感波段，k为常数。当然，组合计算的计算式形式也可以是差值计算、比值计算和幂计算中两种或三种的结合，即将差值计算、比值计算和幂计算进行结合构建为更复杂的计算式，如(A-B)/C，[(A-B)/(A+B)] k，[(A-B)/(A+C)] k等等，C也为一个敏感波段。对于任意一种形式的N维指数，所用的N个敏感波段可从11个敏感波段中进行遍历采样，因此本发明的多维指数集 合中指数存在指数形式和敏感波段两个区分维度，相同形式的指数因为所选择的敏感波段不同还存在众多变种，每一个N维指数通过选择不同的N个敏感波段可以进一步衍生出一系列的指数，这些指数形式不同或者敏感波段不同的指数都被纳入多维指数集合中并分配唯一的ID，作为用于后续筛选的指数。
当得到多维指数集合后，再进行指数的筛选，从中找到与秸秆残留农田土壤盐分含量最相关的指数。由于多维指数集合中的每一个指数都是基于一个或多个波段进行计算的，而波段的波段值需要从光谱曲线上获取。在前述S22步骤中得到了6条光谱组成的光谱数据集，因此，在进行最敏感指数筛选时，上述多维指数集合中的每一个指数还需要分别从6条光谱上进行采样。具体进行筛选时，可针对前述S1的土壤采样数据集中的每个采样点，从前述S22的光谱数据集中获取每个采样点对应的光谱集合(一共包含6条光谱曲线)，基于光谱集合中的每一条光谱分别计算前述多维指数集合中每一种指数的值；然后以每一条光谱下的每一种指数作为待筛选指数，将上述土壤采样数据集中各土壤采样点的土壤表层盐分含量分别与不同的待筛选指数进行相关性分析。不同的待筛选指数之间可能是采样的光谱不同，也可能是指数的ID不同。每一次相关性分析时，以土壤采样数据集中各土壤采样点的土壤表层盐分含量作为第一条数据序列，以基于土壤采样数据集中各土壤采样点的某一条光谱计算得到的某一种指数作为第二条数据序列，计算两条数据序列的相关性指数。最终，筛选出相关性最高的一个待筛选指数作为高敏感指数。
S4、当确定高敏感指数后，即可以土壤采样数据集中的各采样点为回归样本，以采样点处的高敏感指数作为自变量，以采样点处的土壤表层盐分含量作为因变量，通过回归分析建立回归模型。本实施例中，所采用的回归模型为多项式回归模型，其形式为：
EC＝a×ERSSI 2+b×ERSSI+c
式中：EC为土壤表层盐分含量，a、b、c分别为回归系数。
当构建得到回归模型后，从前述S2得到的光谱数据集和雷达数据集中获取待测区域内每一个栅格对应的高敏感指数的值，将其输入构建得到的回归模型中进而计算得到每一个栅格处的土壤表层盐分含量，所有栅格计算完毕后，最终形成待测区域内秸秆残留农田土壤盐分含量的空间分布图。
在本发明的后续实施例中，筛选到了一个高敏感指数，即为基于二阶微分变换光谱计算的三维指数，三维指数形式为
Figure PCTCN2022090854-appb-000003
其中ERSSI为高敏感指数，Green为Sentinel-2卫星数据的绿(Green)波段，Blue为Sentinel-2卫星数据的蓝(Blue)波段，SWIR 1为Sentinel-2卫星数据的短波红外1波段(SWIR1)，Green、Blue和SWIR 1波段值都需要从二阶微分变换光谱上进行采样。当然，如果构建的多维指数集合不同，亦可能可以筛选到其他的高敏感指数，对此不做限定。
需说明的是，前述的三维指数ERSSI可以用于直接预测秸秆残留农田土壤盐分，对于干旱区秸秆残留农而言，仅依靠该指数即可实现农田土壤盐分的反演，无需依赖于其他的波段。因此，本发明中进一步提供了一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其步骤如下：
1)、获取待测区域在待估算时间下对应的Sentinel-2卫星数据；其中待测区域的土地类型为干旱区秸秆残留农田，即位于干旱区且地表残留有秸秆的农田。
2)、对Sentinel-2卫星数据中的多光谱影像进行预处理，得到不同波段的表面反射率数据，再对每个栅格的光谱进行二阶微分变换得到二阶微分变换光谱。多光谱影像的预处理可参见前述方法。
3)、基于待测区域内每个栅格的二阶微分变换光谱计算对应的三维指数ERSSI，计算公式如前所述。
4)、将待测区域内每个栅格的三维指数作为自变量，基于预先针对待测区域构建的回归模型估计每个栅格对应的土壤表层盐分含量，最终形成待测区域内秸秆残留农田土壤盐分含量的空间分布图。
为了进一步便于理解本发明的优点，下面将上述实施例中S1～S4步骤的利用遥感构建指数估算秸秆残留农田土壤盐分的方法应用于一个具体的案例中，以便于展示具体的技术效果。
实施例
选取新疆维吾尔自治区的棉花种植田块(81°17'-81°22'E，40°28'-40°31'N)作为待测区域，利用2020年11月1日地面调查获取的土壤样点数据得到的0.2米土壤盐分含量作为因变量，以2020年11月1日的单景Sentinel-2卫星遥感影像数据及2020年11月2日的Sentinel-1数据作为自变量，基于光谱变换和二维、 三维、四维指数的构建，得到与干旱区秸秆残留农田土壤盐分的高敏感型指数增强型秸秆残留盐分指数(Enhanced Residues Soil Salinity Index，ERSSI)。并以ERSSI为自变量，以土壤盐分含量为因变量，利用线性多项式回归模型估算，得到待测区域的秸秆残留农田土壤盐分含量分布图。该指数及估算方法具体如下：
步骤1)数据获取：以干旱区秸秆残留农田为估算对象，根据待估算的时间，获取待测区域同一时期的Sentinel-2卫星数据、Sentinel-1雷达数据、土壤表层盐分含量数据集；Sentinel-2卫星数据和Sentinel-1雷达数据的空间分辨率为10米。
其中，Sentinel-2数据是欧洲航天局(ESA)发布的L1级产品。Sentinel-2由两颗极轨卫星(Sentinel-2A和Sentinel-2B)组成，两颗卫星数据互补，重访周期为5天，轨道周期为100分钟，轨道高度为786公里。扫描宽度为290公里，轨道倾角为98.62°。多光谱扫描成像数据(MSI)在可见光至红色边缘区域提供一个空间分辨率为60m的波段、三个10m波段和两个30m波段，该数据可在美国地质调查局(USGS)下载免费下载，该步骤使用了2020年11月1日的Sentinel-2B数据；Sentinel-1卫星是欧洲航天局哥白尼计划(GMES)发射的一颗地球观测卫星，由两颗卫星(Sentinel-1A和Sentinel-1B)组成，携带C波段合成孔径雷达，允许以6天的时间分辨率提供全天图像。该步骤使用了2020年11月2日的Sentinel-1A数据，因为该时期的影像与野外实验日期接近，土壤和冠层条件几乎相同。Sentinel-1A数据上升方向和干涉宽(IW)模式下获得，具有两个双偏振，空间分辨率为10m。
土壤表层盐分含量数据集是通过对不同点位的0～0.2m土层进行土壤样品采样，并运回实验室进行分析得到的额。获取的土壤样品采用电导率法测定其电导率：将土壤样品风干、研磨并过2毫米筛，以土水比1：5得到土壤浸出液，过滤取上清液测定其电导率，得到土壤样品的电导率数据，用于表征待测区域的土壤盐渍化水平。采用地理分区法将土壤样点数据分为两个区域，根据样点分布的地理位置不同分为两个子区域Q1、Q2，其中，Q1的电导率数据集作为建模集，用于新指数的构建、相关性评估和多项式回归方程的构建，Q2的电导率数据集作为独立验证集，用于独立验证指数和多项式回归方程的有效性。
步骤2)数据预处理：基于多源数据的融合和处理，将步骤1)获取的光学遥感影像数据、雷达数据和土壤样品数据进行预处理和计算。
将步骤1)获取的Sentinel-2遥感数据进行预处理：利用Sentinel Application Platform(SNAP)包中的Sen2Cor模块对Sentinel-2数据进行辐射校准和大气校正，将MSI图像转换为表面反射率格式输出，得到分辨率为10米的Sentinel-2影像。另外，将步骤1)获取的Sentinel-1雷达数据进行预处理：依次进行轨道校正、热噪声去除、辐射校正、Lee滤波和多普勒地形校正，应用水云模型去除植被含水量对土壤后向散射系数的影响，将垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数转换并得到分贝(dB)格式的雷达数据集，得到表面反射率数据。
将经过预处理的Sentinel-2卫星数据的各个波段采用逐波段计算方法进行光谱变换，逐个波段进行对数变换、倒数变换、一阶微分变换、二阶微分变换和三阶微分变换一共6中数学计算光谱变换，得到原始光谱曲线以及五种数学计算变换后的光谱曲线，每一个栅格的6条光谱作为该栅格对应的光谱集合，所有栅格的光谱集合作为光谱数据集。
步骤3)新指数的构建：遍历土壤表层盐分的敏感波段组合中不同的波段，并将其组合成不同的二维、三维和四维指数，得到多维指数集合。
在本实施例中，的对土壤盐分敏感的敏感波段组合是采用归纳统计方法筛选的，根据影响土壤盐分监测精度的土壤、作物、水分因子，以及盐渍土的光谱反射特性，总结归纳现有的盐分指数、植被指数和土壤相关指数等使用的波段，在光学遥感数据中筛选得到蓝、绿、红、近红外、短波红外、植被红边指数为敏感波段，在Sentinel-2原始光谱及五种数学计算变换后的光谱中提取以上波段，得到9个光学敏感波段；在雷达数据中筛选得到土壤的雷达后向散射系数VV、VH为敏感波段，在Sentinel-1数据中计算得到2个雷达敏感波段。共得到11个对土壤盐分敏感的波段。采用归纳、增减维度方法进行新指数的构建。首先，对现有盐分指数的计算形式进行总结，在指数的构建形式上归纳得到差值、比值、幂计算三种计算形式，在此基础上，对三种计算形式中的变量改变维度、减少和增加维度，进行维度的改变，用于构建新指数，以11个敏感波段为变量，构建27个指数形式，包括9个二维指数、8个三维指数、10个四维指数，指数及其计算式如表1所示；
表1二维、三维、四维指数及其计算式
Figure PCTCN2022090854-appb-000004
其中，B i，B j，B k，B h分别为蓝、绿、红、近红外、短波红外、植被红边指数波段以及雷达后向散射系数VV、VH这11个波段中的任意四种。每一个指数形式均需要遍历整个敏感波段集合，选择不同的波段构成不同的指数加入多维指数集合中，用于后续的筛选。
最后，再针对Q1中的每个采样点，从光谱数据集中获取每个采样点对应的光谱集合，基于光谱集合中的每一条光谱(分别为原始光谱、对数变换光谱、倒数变换光谱、一阶微分变换光谱、二阶微分变换光谱和三阶微分变换光谱)分别计算多维指数集合中每一种指数的值。然后以每一条光谱下的每一种指数作为待筛选指数，将土壤采样数据集中各土壤采样点的土壤表层盐分含量分别与不同的待筛选指数进行相关性分析，筛选出相关性最高的至少一个待筛选指数作为高敏感指数。本实施例中，采用相关性分析评估不同光谱下不同指数(共六组，每组27种指数形式)与Q1中土壤电导率数据集的相关性(如表2所示)，同时在Q2中土壤电导率数据集进行独立验证(如表3所示)，得到相关性最高的指数。
表2由六种形式的光谱构建的二维、三维、四维指数在Q1中的最高相关性(r)
Figure PCTCN2022090854-appb-000005
表3由六种形式的光谱构建的二维、三维、四维指数在Q2中的最高相关性(r)
Figure PCTCN2022090854-appb-000006
在Q1和Q2的相关性分析中，得到对干旱区秸秆覆盖农田土壤盐分相关性最高的指数为二阶导数变换后构建的三维指数，命名为增强型秸秆残留盐分指数(Enhanced Residues Soil Salinity Index，ERSSI)，该新指数的计算公式为：
Figure PCTCN2022090854-appb-000007
其中ERSSI为高敏感指数，Green为Sentinel-2卫星数据的绿(Green)波段，Blue为Sentinel-2卫星数据的蓝(Blue)波段，SWIR 1为Sentinel-2卫星数据的短波红外1波段(SWIR-1)，Green、Blue和SWIR 1波段值都需要从二阶微分变换光谱上进行采样。
图1为使用ERSSI和多项式回归的土壤盐分含量估算值在Q1(建模集)和Q2(独立验证集)中表现的散点图，表面本发明估算得到的盐分数据相对于地面实测值的拟合程度较高。
ERSSI以比率的形式组合了绿、蓝和短波红外波段。绿波段对植被类型敏感，而对大气影响不太敏感，并使用蓝波段解决了初步大气校正后残留气溶胶引起的植被指数衰减问题。在处理土壤背景、大气和饱和噪声对盐分监测的影响方面，ERSSI中蓝绿波段组合可以有效应对干扰。从评价结果可以看出，待测区域的覆盖类型为秸秆残留，盐侵状态包括盐壳，常规植被敏感波段不适用于干旱区秸秆残留农田。为了解决植物残留的影响，使用SWIR 1代替NIR，可以有效识别缺水环境下植被的细胞结构和土壤含水量，并利用绿和SWIR 1波段的组合可以表征铁含量和土壤水分含量低和土壤状况，从而间接表明盐渍化。待测区域土壤盐渍化主要含有钙离子和硫酸根离子，两者在1443-1745nm范围内反射强烈，SWIR 1波段对这些物质敏感，使用SWIR 1波段为监测新疆秸秆残留农田土壤盐渍化提供了机会。
步骤4)干旱区秸秆残留的农田土壤盐分估算：根据步骤3)得到的相关性评价结果，以Sentinel-2遥感影像的二阶导数变换计算得到的增强型秸秆残留盐分指数(Enhanced Residues Soil Salinity Index，ERSSI)为自变量，以步骤2)得到的Q1中土壤电导率数据集为因变量，在R语言中构建线性多项式回归方程：
EC＝0.00072×ERSSI 2+0.072×ERSSI+0.92
其中，EC为干旱区秸秆残留农田的土壤表层盐分含量(dS/m)。该多项式回归方程在步骤2)得到的Q1(建模训练集)中精度R 2为0.63；同时，将增强型秸秆残留盐分指数(ERSSI)和多项式回归方程在步骤2)得到的Q2(独立验证集)中进行独立验证，验证精度R 2为0.64；
根据增强型秸秆残留盐分指数(ERSSI)的计算公式，基于Sentinel-2光学遥感影像，对待测区域的每一个栅格点进行ERSSI指数计算，将构建的多项式回归模型应用到待测区域中，得到待测区域的逐像元的土壤盐分含量。
选取新疆维吾尔自治区的棉花种植田块作为待测区域，以本实施方式估算得到的2020年土壤盐分含量分布图如图2所示，空间分辨率为10米。需说明的是，该土壤盐分含量分布图为以电导率作为表征的相对盐分含量分布图，如需转换为 绝对盐分含量的分布图，则将每个像元的值按照电导率与土壤盐分含量的转换公式进行转换即可。
以上所述的实施例只是本发明的一种较佳的方案，然其并非用以限制本发明。有关技术领域的普通技术人员，在不脱离本发明的精神和范围的情况下，还可以做出各种变化和变型。因此凡采取等同替换或等效变换的方式所获得的技术方案，均落在本发明的保护范围内。

Claims (10)

  1. 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，步骤如下：
    S1、获取待测区域在待估算时间下对应的Sentinel-2卫星数据、Sentinel-1雷达数据、土壤采样数据集；所述待测区域的土地类型为干旱区秸秆残留农田；所述土壤采样数据集中包含待测区域内不同采样点对应的土壤表层盐分含量；
    S2、对Sentinel-2卫星数据中的多光谱影像进行预处理，得到不同波段的表面反射率数据，再通过多种光谱变换得到光谱数据集；所述光谱变换形式包括对数变换、倒数变换和微分变换；所述光谱数据集中包含影像中每一个栅格的光谱集合，所述光谱集合包括原始光谱和经过所述光谱变换后得到的多种变换光谱；对Sentinel-1雷达数据进行预处理，得到包含垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数的雷达数据集；
    S3、遍历土壤表层盐分的敏感波段组合中不同的波段，并将其组合成不同的二维、三维和四维指数，得到多维指数集合；再针对所述土壤采样数据集中的每个采样点，从所述光谱数据集中获取每个采样点对应的光谱集合，基于光谱集合中的每一条光谱分别计算所述多维指数集合中每一种指数的值；然后以每一条光谱下的每一种指数作为待筛选指数，将所述土壤采样数据集中各土壤采样点的土壤表层盐分含量分别与不同的待筛选指数进行相关性分析，筛选出相关性最高的至少一个待筛选指数作为高敏感指数；
    S4、以所述土壤采样数据集中的各采样点为回归样本，以所述高敏感指数作为自变量，以土壤表层盐分含量作为因变量，建立回归模型；并从所述光谱数据集和所述雷达数据集中获取待测区域内每一个栅格对应的所述高敏感指数的值，根据所述回归模型计算得到每一个栅格处的土壤表层盐分含量，最终形成待测区域内秸秆残留农田土壤盐分含量的空间分布图。
  2. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述Sentinel-2卫星数据中的多光谱影像的预处理包括辐射定标和大气校正，且其中大气校正所选用的模型为FLAASH。
  3. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述Sentinel-1雷达数据的预处理包括轨道校正、热噪声去除、 辐射校正、Lee滤波和多普勒地形校正，并选用水云模型去除植被含水量对土壤后向散射系数的影响后，将垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数转换并得到分贝(dB)格式的雷达数据集。
  4. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述土壤采样数据集中，每一个采样点对应的土壤表层盐分含量由土壤表面0～0.2m的土层通过电导率法测定得到。
  5. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述Sentinel-2卫星数据和Sentinel-1雷达数据的空间分辨率为10米。
  6. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述敏感波段组合由光学遥感敏感波段和雷达敏感波段组成，其中所述光学遥感敏感波段为Sentinel-2卫星数据中的蓝(Blue)波段、绿(Green)波段、红(Red)波段、近红外(NIR)波段、短波红外1波段(SWIR 1)、短波红外2波段(SWIR 2)和3个植被红边指数(Red Edge 1，Red Edge 2，Red Edge 3)波段，所述雷达敏感波段为Sentinel-1雷达数据中垂直发射-垂直接收(VV)和垂直发射-水平接收(VH)的雷达后向散射系数。
  7. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述微分变换包括一阶微分变换、二阶微分变换和三阶微分变换，影像中每一个栅格的光谱集合均包含6条光谱，分别为原始光谱、对数变换光谱、倒数变换光谱、一阶微分变换光谱、二阶微分变换光谱和三阶微分变换光谱。
  8. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述多维指数集合中，二维、三维和四维指数均由相应维数的波段通过组合计算得到，组合计算的形式为差值计算、比值计算和幂计算中的一种或多种结合。
  9. 如权利要求1所述的利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，所述回归模型为多项式回归模型，其形式为：
    EC＝a×ERSSI 2+b×ERSSI+c
    式中：EC为土壤表层盐分含量，a、b、c分别为回归系数。
  10. 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法，其特征在于，步骤如下：
    1)、获取待测区域在待估算时间下对应的Sentinel-2卫星数据；所述待测区域的土地类型为干旱区秸秆残留农田；
    2)、对Sentinel-2卫星数据中的多光谱影像进行预处理，得到不同波段的表面反射率数据，再对每个栅格的光谱进行二阶微分变换得到二阶微分变换光谱；
    3)、基于待测区域内每个栅格的二阶微分变换光谱计算对应的三维指数，三维指数形式为
    Figure PCTCN2022090854-appb-100001
    其中ERSSI为高敏感指数，Green为Sentinel-2卫星数据的绿(Green)波段，Blue为Sentinel-2卫星数据的蓝(Blue)波段，SWIR 1为Sentinel-2卫星数据的短波红外1波段(SWIR 1)；
    4)、将待测区域内每个栅格的三维指数作为自变量，基于预先针对待测区域构建的回归模型估计每个栅格对应的土壤表层盐分含量，最终形成待测区域内秸秆残留农田土壤盐分含量的空间分布图。
PCT/CN2022/090854 2021-11-22 2022-05-05 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法 WO2023087630A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111385163.2 2021-11-22
CN202111385163.2A CN114076738B (zh) 2021-11-22 2021-11-22 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法

Publications (1)

Publication Number Publication Date
WO2023087630A1 true WO2023087630A1 (zh) 2023-05-25

Family

ID=80284190

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/090854 WO2023087630A1 (zh) 2021-11-22 2022-05-05 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法

Country Status (2)

Country Link
CN (1) CN114076738B (zh)
WO (1) WO2023087630A1 (zh)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116645593A (zh) * 2023-07-20 2023-08-25 南京信息工程大学 监测海草床分布的遥感方法、装置
CN116663783A (zh) * 2023-07-28 2023-08-29 中国科学院东北地理与农业生态研究所 沼泽湿地生态系统碳储量统计分析系统
CN116702065A (zh) * 2023-05-30 2023-09-05 浙江时空智子大数据有限公司 基于影像数据黑臭水体生态治理污染监测方法及系统
CN116952849A (zh) * 2023-08-01 2023-10-27 中国科学院东北地理与农业生态研究所 Msi的表层土壤质地反演方法
CN117153291A (zh) * 2023-10-31 2023-12-01 水利部交通运输部国家能源局南京水利科学研究院 一种灌区稻田碳汇价值计算方法及系统
CN117540854A (zh) * 2023-11-09 2024-02-09 中国标准化研究院 一种基于云平台的多通道土壤信息监测系统
CN117630000A (zh) * 2024-01-26 2024-03-01 北京观微科技有限公司 土壤含盐量评估方法、装置、电子设备及存储介质
CN117789025A (zh) * 2023-12-26 2024-03-29 北京工业大学 一种基于物候信息的分作物秸秆露天焚烧排放估算方法
CN118015420A (zh) * 2024-03-28 2024-05-10 珠江水利委员会珠江水利科学研究院 基于植被盖度特征幂次函数的遥感影像融合方法、系统及介质
CN117789025B (zh) * 2023-12-26 2024-06-04 北京工业大学 一种基于物候信息的分作物秸秆露天焚烧排放估算方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114076738B (zh) * 2021-11-22 2023-04-18 浙江大学 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法
CN116662786A (zh) * 2023-07-13 2023-08-29 江西慧航工程咨询有限公司 一种用于地下建筑的数字化地质分析方法
CN116757099B (zh) * 2023-08-18 2024-03-19 中国科学院南京土壤研究所 一种土壤盐分反演及盐渍化风险评估方法、装置及设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016000088A1 (zh) * 2014-07-02 2016-01-07 中国林业科学研究院林业研究所 一种基于最佳指数-相关系数法的高光谱波段提取方法
CN108680509A (zh) * 2018-08-17 2018-10-19 山东农业大学 一种滨海盐渍区土壤盐分含量估算方法
CN112711989A (zh) * 2020-12-15 2021-04-27 中国农业大学 基于雷达遥感与光学遥感的玉米秸秆覆盖度估算方法
CN112949681A (zh) * 2021-01-27 2021-06-11 中国科学院东北地理与农业生态研究所 基于卫星光谱的盐渍土电导率估算方法
CN113221445A (zh) * 2021-04-21 2021-08-06 山东师范大学 利用遥感影像的联合特征对土壤盐分估测的方法及系统
CN114076738A (zh) * 2021-11-22 2022-02-22 浙江大学 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AUPR301401A0 (en) * 2001-02-09 2001-03-08 Commonwealth Scientific And Industrial Research Organisation Lidar system and method
CN102645350B (zh) * 2012-03-08 2013-11-13 浙江大学 基于高分卫星遥感数据的土壤采样方法
US9579700B2 (en) * 2014-05-30 2017-02-28 Iteris, Inc. Measurement and modeling of salinity contamination of soil and soil-water systems from oil and gas production activities
CN107992696B (zh) * 2017-12-13 2020-12-29 电子科技大学 一种复杂色散媒质中的改进的指数时间积分构造方法
CN108458978B (zh) * 2018-03-13 2021-03-19 山东农业大学 基于敏感波段和波段组合最优的树种多光谱遥感识别方法
CN108662991B (zh) * 2018-04-08 2020-02-07 浙江大学 基于遥感卫星数据的地块尺度冬小麦叶面积指数估算方法
CN108663330B (zh) * 2018-04-19 2019-08-16 中国自然资源航空物探遥感中心 一种基于叶片实测光谱的植被覆盖区土壤铜元素反演方法
CN108918820B (zh) * 2018-05-23 2020-08-14 中国水利水电科学研究院 获取耕地土壤盐渍化程度等级分布的方法和装置
CN108693333B (zh) * 2018-06-14 2021-05-28 中铁二院成都勘察设计研究院有限责任公司 一种确定粗颗粒硫酸钠盐渍土盐胀系数的方法
CN109884664B (zh) * 2019-01-14 2022-12-02 武汉大学 一种城市地上生物量光学微波协同反演方法及系统
CN109738380B (zh) * 2019-01-25 2022-09-30 西北农林科技大学 一种土壤盐渍化程度的高光谱遥感判断方法
CN112213287B (zh) * 2020-12-07 2021-04-06 速度时空信息科技股份有限公司 基于遥感卫星影像的沿海滩涂盐度反演的方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016000088A1 (zh) * 2014-07-02 2016-01-07 中国林业科学研究院林业研究所 一种基于最佳指数-相关系数法的高光谱波段提取方法
CN108680509A (zh) * 2018-08-17 2018-10-19 山东农业大学 一种滨海盐渍区土壤盐分含量估算方法
CN112711989A (zh) * 2020-12-15 2021-04-27 中国农业大学 基于雷达遥感与光学遥感的玉米秸秆覆盖度估算方法
CN112949681A (zh) * 2021-01-27 2021-06-11 中国科学院东北地理与农业生态研究所 基于卫星光谱的盐渍土电导率估算方法
CN113221445A (zh) * 2021-04-21 2021-08-06 山东师范大学 利用遥感影像的联合特征对土壤盐分估测的方法及系统
CN114076738A (zh) * 2021-11-22 2022-02-22 浙江大学 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
MA GUOLIN, DING JIANLI, HAN LIJNG, ZHANG ZIPENG, RAN SI: "Digital mapping of soil salinization based on Sentinel-1 and Sentinel-2 data combined with machine learning algorithms", REGIONAL SUSTAINABILITY, vol. 2, no. 2, 1 April 2021 (2021-04-01), pages 177 - 188, XP093068164, ISSN: 2666-660X, DOI: 10.1016/j.regsus.2021.06.001 *
TRIPATHI AKSHAR, TIWARI REET KAMAL: "A simplified subsurface soil salinity estimation using synergy ofSENTINEL-1 SAR and SENTINEL-2 multispectral satellite data,for early stages of wheat crop growth in Rupnagar, Punjab,India", LAND DEGRADATION & DEVELOPMENT, JOHN WILEY & SONS, INC., HOBOKEN, USA, vol. 32, no. 14, 30 August 2021 (2021-08-30), Hoboken, USA, pages 3905 - 3919, XP093068166, ISSN: 1085-3278, DOI: 10.1002/ldr.4009 *
WANG, ZHENG ET AL.: "Regional suitability prediction of soil salinization based on remote-sensing derivatives and optimal spectral index", SCIENCE OF THE TOTAL ENVIRONMENT, vol. 775, 12 February 2021 (2021-02-12), XP086538940, ISSN: 0048-9697, DOI: 10.1016/j.scitotenv.2021.145807 *
ZHANG QIANQIAN; LI LI; CHAO ZHANG; RUIZHI SUN: "Fusion of Sentinel-1 and Sentinel-2 Imagery for Land Salinity Mapping: A Case Study in Da'an, Jilin Province", 2021 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM IGARSS, 11 July 2021 (2021-07-11), pages 6560 - 6563, XP033986562, DOI: 10.1109/IGARSS47720.2021.9553204 *

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116702065A (zh) * 2023-05-30 2023-09-05 浙江时空智子大数据有限公司 基于影像数据黑臭水体生态治理污染监测方法及系统
CN116702065B (zh) * 2023-05-30 2024-04-16 浙江时空智子大数据有限公司 基于影像数据黑臭水体生态治理污染监测方法及系统
CN116645593B (zh) * 2023-07-20 2023-10-24 南京信息工程大学 监测海草床分布的遥感方法、装置
CN116645593A (zh) * 2023-07-20 2023-08-25 南京信息工程大学 监测海草床分布的遥感方法、装置
CN116663783A (zh) * 2023-07-28 2023-08-29 中国科学院东北地理与农业生态研究所 沼泽湿地生态系统碳储量统计分析系统
CN116663783B (zh) * 2023-07-28 2023-10-13 中国科学院东北地理与农业生态研究所 沼泽湿地生态系统碳储量统计分析系统
CN116952849A (zh) * 2023-08-01 2023-10-27 中国科学院东北地理与农业生态研究所 Msi的表层土壤质地反演方法
CN117153291A (zh) * 2023-10-31 2023-12-01 水利部交通运输部国家能源局南京水利科学研究院 一种灌区稻田碳汇价值计算方法及系统
CN117153291B (zh) * 2023-10-31 2024-01-02 水利部交通运输部国家能源局南京水利科学研究院 一种灌区稻田碳汇价值计算方法及系统
CN117540854A (zh) * 2023-11-09 2024-02-09 中国标准化研究院 一种基于云平台的多通道土壤信息监测系统
CN117540854B (zh) * 2023-11-09 2024-04-30 中国标准化研究院 一种基于云平台的多通道土壤信息监测系统
CN117789025A (zh) * 2023-12-26 2024-03-29 北京工业大学 一种基于物候信息的分作物秸秆露天焚烧排放估算方法
CN117789025B (zh) * 2023-12-26 2024-06-04 北京工业大学 一种基于物候信息的分作物秸秆露天焚烧排放估算方法
CN117630000A (zh) * 2024-01-26 2024-03-01 北京观微科技有限公司 土壤含盐量评估方法、装置、电子设备及存储介质
CN117630000B (zh) * 2024-01-26 2024-04-02 北京观微科技有限公司 土壤含盐量评估方法、装置、电子设备及存储介质
CN118015420A (zh) * 2024-03-28 2024-05-10 珠江水利委员会珠江水利科学研究院 基于植被盖度特征幂次函数的遥感影像融合方法、系统及介质

Also Published As

Publication number Publication date
CN114076738B (zh) 2023-04-18
CN114076738A (zh) 2022-02-22

Similar Documents

Publication Publication Date Title
WO2023087630A1 (zh) 一种利用遥感构建指数估算秸秆残留农田土壤盐分的方法
Bao et al. Surface soil moisture retrievals over partially vegetated areas from the synergy of Sentinel-1 and Landsat 8 data using a modified water-cloud model
Hosseini et al. Estimation of Leaf Area Index (LAI) in corn and soybeans using multi-polarization C-and L-band radar data
Liang et al. Multi-factor modeling of above-ground biomass in alpine grassland: A case study in the Three-River Headwaters Region, China
Fernandes et al. Landsat-5 TM and Landsat-7 ETM+ based accuracy assessment of leaf area index products for Canada derived from SPOT-4 VEGETATION data
CN101963664B (zh) 基于水陆地物分类信息的微波遥感混合像元分解方法
Chen et al. Vegetation water content estimation for corn and soybeans using spectral indices derived from MODIS near-and short-wave infrared bands
Ganguly et al. Generating global leaf area index from Landsat: Algorithm formulation and demonstration
CN110136194B (zh) 基于星载多光谱遥感数据的积雪覆盖度测算方法
Wang et al. Regional suitability prediction of soil salinization based on remote-sensing derivatives and optimal spectral index
CN113205475B (zh) 基于多源卫星遥感数据的森林高度反演方法
Sun et al. Retrieval of rapeseed leaf area index using the PROSAIL model with canopy coverage derived from UAV images as a correction parameter
CN102194127B (zh) 一种多频率sar数据农作物遥感分类方法
CN110109118B (zh) 一种森林冠层生物量的预测方法
CN111337434A (zh) 一种矿区复垦植被生物量估算方法及系统
US20220392215A1 (en) System and Method for Mapping Land Cover Types with Landsat, Sentinel-1, and Sentinel-2 Images
CN114062439B (zh) 一种利用时间序列遥感影像联合估算土壤剖面盐分的方法
Wu et al. Evaluation of winter wheat yield simulation based on assimilating LAI retrieved from networked optical and SAR remotely sensed images into the WOFOST model
Liu et al. Improved potato AGB estimates based on UAV RGB and hyperspectral images
CN111783288A (zh) 基于Landsat8对黄河三角洲土壤盐分的反演方法
Wang et al. Improving remote sensing of salinity on topsoil with crop residues using novel indices of optical and microwave bands
Zhang et al. Variation of intra-daily instantaneous FAPAR estimated from the geostationary Himawari-8 AHI data
Qi et al. Soil salinity inversion in coastal cotton growing areas: An integration method using satellite‐ground spectral fusion and satellite‐UAV collaboration
Dwivedi et al. Assessment of rice biomass production and yield using semi-physical approach and remotely sensed data
Guo et al. Estimating aboveground biomass of alpine grassland during the wilting period using in situ hyperspectral, Sentinel-2 and Sentinel-1 data

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22894175

Country of ref document: EP

Kind code of ref document: A1