CN114446408B - Leaf nitrogen content estimation method based on radiation transmission model inversion - Google Patents
Leaf nitrogen content estimation method based on radiation transmission model inversion Download PDFInfo
- Publication number
- CN114446408B CN114446408B CN202210357777.8A CN202210357777A CN114446408B CN 114446408 B CN114446408 B CN 114446408B CN 202210357777 A CN202210357777 A CN 202210357777A CN 114446408 B CN114446408 B CN 114446408B
- Authority
- CN
- China
- Prior art keywords
- nitrogen content
- spectrum
- inversion
- blade
- leaf
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 title claims abstract description 200
- 229910052757 nitrogen Inorganic materials 0.000 title claims abstract description 100
- 230000005855 radiation Effects 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000005540 biological transmission Effects 0.000 title claims abstract description 24
- 238000001228 spectrum Methods 0.000 claims abstract description 77
- 230000003595 spectral effect Effects 0.000 claims abstract description 60
- 238000002310 reflectometry Methods 0.000 claims abstract description 12
- 238000002790 cross-validation Methods 0.000 claims abstract description 8
- 230000008569 process Effects 0.000 claims description 8
- 229930002875 chlorophyll Natural products 0.000 claims description 5
- 235000019804 chlorophyll Nutrition 0.000 claims description 5
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 claims description 5
- 238000002329 infrared spectrum Methods 0.000 claims description 5
- YGLMVCVJLXREAK-MTVMDMGHSA-N 1,1-dimethyl-3-[(1S,2R,6R,7S,8R)-8-tricyclo[5.2.1.02,6]decanyl]urea Chemical compound C([C@H]12)CC[C@@H]1[C@@H]1C[C@@H](NC(=O)N(C)C)[C@H]2C1 YGLMVCVJLXREAK-MTVMDMGHSA-N 0.000 claims description 3
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 claims description 3
- 229910052799 carbon Inorganic materials 0.000 claims description 3
- 235000021466 carotenoid Nutrition 0.000 claims description 3
- 150000001747 carotenoids Chemical class 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 150000001875 compounds Chemical class 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 238000004088 simulation Methods 0.000 claims description 2
- 230000012010 growth Effects 0.000 abstract description 5
- 238000010521 absorption reaction Methods 0.000 description 7
- 241000894007 species Species 0.000 description 7
- 241000196324 Embryophyta Species 0.000 description 6
- 238000010586 diagram Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- -1 nitrogen-containing compound Chemical class 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000209504 Poaceae Species 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 150000001722 carbon compounds Chemical class 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 235000008216 herbs Nutrition 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000029553 photosynthesis Effects 0.000 description 1
- 238000010672 photosynthesis Methods 0.000 description 1
- 230000035790 physiological processes and functions Effects 0.000 description 1
- 230000008635 plant growth Effects 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 230000005068 transpiration Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/20—Identification of molecular entities, parts thereof or of chemical compositions
-
- 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
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C10/00—Computational theoretical chemistry, i.e. ICT specially adapted for theoretical aspects of quantum chemistry, molecular mechanics, molecular dynamics or the like
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Crystallography & Structural Chemistry (AREA)
- Biochemistry (AREA)
- Pathology (AREA)
- Immunology (AREA)
- General Physics & Mathematics (AREA)
- Analytical Chemistry (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a leaf nitrogen content estimation method based on radiation transmission model inversion, which comprises the steps of acquiring leaf spectral reflectivity data by collecting plants in a wide geographical area, species and growth period, inverting a radiation transmission model through a full-spectrum wave band to obtain leaf structure parameters, selecting a semi-discrete optimal spectrum set for model inversion by using the leaf structure parameters as priori knowledge, and performing cross validation on an inversion strategy, thereby realizing high-precision and high-universality inversion estimation of leaf nitrogen content.
Description
Technical Field
The invention relates to the technical field of vegetation remote sensing, in particular to a leaf nitrogen content estimation method based on radiation transmission model inversion.
Background
Nitrogen is an important limiting factor for plant growth, can influence carbon change of a land ecosystem through a potential climate feedback effect, and is also an important input parameter of a process model of the ecosystem. Leaf nitrogen regulates a variety of plant physiological processes, mainly including photosynthesis, leaf respiration and transpiration, and is closely linked to canopy and plant level traits such as light energy utilization efficiency, woody growth and net primary productivity. Because of the important role of the nitrogen-containing compound in biodiversity and ecosystem functions, the nitrogen content of the leaves is also listed as one of key biodiversity indexes, and the nitrogen-containing compound is a research target which is focused in the remote sensing and ecological fields.
The leaf nitrogen content is only 0.2-6.4% of the dry weight of the leaf, and the proportion is very low, and the leaf nitrogen content estimated based on remote sensing mainly utilizes the absorption characteristics of protein and chlorophyll, such as spectral signals in visible light, near infrared and short wave infrared regions. As the spectral absorption wave band of the leaf nitrogen is mostly located in the short wave infrared region and is often covered by the absorption characteristics with stronger dry matters and more obvious moisture, the remote sensing inversion of the leaf nitrogen has high uncertainty.
In the past, a statistical regression method is mainly adopted for the research of inverting the nitrogen content of the blade by utilizing hyperspectral remote sensing, the method is used for predicting by establishing a regression equation of the nitrogen content of the blade and spectral reflectance data, the method is simple and easy to use, but the method is often inconsistent with the actual absorption characteristic of the nitrogen of the blade, so that the mechanism interpretation degree of a model is low; in addition, since the regression equation is obtained by data fitting, other independent data sets are difficult to use, and the generality of the model is low. Moreover, the radiation transmission model of the method for inverting the nitrogen content of the blade by hyperspectral remote sensing simulates the interaction of solar radiation and the blade based on the physical optics law, including the processes of reflection, scattering, absorption and the like, so as to simulate the spectral reflectivity of the blade, and the component parameters of the blade are inversely estimated by utilizing the reverse process of the model. The method is widely applied to components with obvious spectral absorption characteristics such as chlorophyll, moisture and the like, but the spectral absorption characteristics for leaf nitrogen are weak, and the combination of different leaf parameters can generate similar spectral reflectivity, so that the problem of 'ill-inversion' is caused, and the inversion accuracy is low.
In summary, a radiation transmission model inversion method capable of estimating the nitrogen content of the leaf blade with high precision is still lacking at present.
Disclosure of Invention
Aiming at the problems, the invention provides a leaf nitrogen content estimation method based on radiation transmission model inversion, which mainly solves the problems of low inversion precision and low model universality of the existing hyperspectral satellite remote sensing leaf nitrogen content inversion method.
In order to solve the technical problems, the technical scheme of the invention is as follows:
a leaf nitrogen content estimation method based on radiation transmission model inversion comprises the following steps:
collecting two groups of leaf samples of forest and grassland ecosystems located in different continents, and respectively obtaining spectral reflectance measured values and leaf nitrogen content measured values corresponding to the two groups of leaf samples, wherein the spectral reflectance measured values and the leaf nitrogen content measured values are respectively defined as a first data group and a second data group;
bringing the specular reflection spectral radiation into a leaf radiation transmission model PROSPECT to establish an inversion model;
dividing short-wave infrared spectrum bands in the inversion model into a plurality of spectrum subsets according to preset length intervals, taking the spectrum subsets of different bands as initial conditions, respectively substituting the spectral reflectivity measured values corresponding to the first data group and the second data group into two independent inversion models for preliminary inversion to obtain respective first blade nitrogen content estimated values, calculating a first root-mean-square error between the two first blade nitrogen content estimated values and the corresponding blade nitrogen content measured values, taking the first root-mean-square error with the highest precision as a screening condition, and selecting respective semi-discrete optimal spectrum sets from the spectrum subsets through backward iterative recursion;
and cross-substituting the two groups of semi-discrete optimal spectrum sets into the inversion model subjected to the preliminary inversion for cross validation to obtain respective second blade nitrogen content estimated values, calculating a second root mean square error between the two second blade nitrogen content estimated values and the corresponding blade nitrogen content measured values, and outputting the inversion model if the second root mean square error is within a preset value.
The invention has the beneficial effects that: the leaf spectral reflectivity data is acquired by collecting plants in a wide geographical area, species and growth period, the semi-discrete optimal spectrum set is inverted by adopting the priori knowledge and a selected model, and the inversion strategy is cross-validated, so that the high-precision and high-universality inversion estimation of the leaf nitrogen content is realized.
Drawings
FIG. 1 is a flow chart of a method for estimating nitrogen content of a blade based on inversion of a radiation transmission model according to an embodiment of the invention;
FIG. 2a is a schematic view of a first root mean square error relationship between a first estimated vane nitrogen content value and a corresponding measured vane nitrogen content value after a preliminary inversion of a first data set;
FIG. 2b is a schematic diagram of a first root mean square error relationship between a first estimated vane nitrogen content value and a corresponding measured vane nitrogen content value after a preliminary inversion of a second data set;
FIG. 3a is a schematic diagram of a relationship between inversion using a radiation transport model and a semi-discrete optimal spectrum set for a first data set;
FIG. 3b is a schematic diagram of a relationship between inversion using a radiation transport model and a semi-discrete optimal spectrum set for a second data set;
FIG. 4a is a graph showing the results of using a semi-discrete optimal set of spectra for the second data set for the first data set;
FIG. 4b is a graph showing the results of using the semi-discrete optimal set of spectra for the first data set for the second data set.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the following detailed description of the present invention is provided with reference to the accompanying drawings and detailed description. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some but not all of the relevant aspects of the present invention are shown in the drawings.
The embodiment provides a leaf nitrogen content estimation method based on radiation transmission model inversion, leaf spectral reflectivity data are obtained by collecting plants in a wide geographical area, species and growth period, a semi-discrete optimal spectrum set is inverted by adopting priori knowledge and a selected model, and cross validation of an inversion strategy is carried out, so that high-precision and high-universality inversion estimation of leaf nitrogen content is realized. The method is shown in figure 1 and comprises the following steps:
s1, collecting two groups of leaf samples of forest and grassland ecosystems located in different continents, respectively obtaining the spectral reflectance measured values and the leaf nitrogen content measured values corresponding to the two groups of leaf samples, and respectively defining the two groups of leaf samples as a first data group and a second data group.
In this embodiment, leaf samples are based on 6 field sites in south China and 18 field sites in east North America, mainly including subtropical, tropical, temperate forest and grassland ecosystems, and for the dominant species of arbor and shrub grasses, top male leaf branches are obtained and about 10 g of fresh leaves are collected. The plant leaf sample comprises different geographical regions, species, illumination conditions and growth stages and aims to cover the change interval of the measured spectral reflectance value and the measured leaf nitrogen content value. Finally, two data sets were obtained comprising 93 species 363 leaf samples (defined as the first data set) in south china and 384 species leaf samples (defined as the second data set) in east north america 161.
Specifically, the spectral reflectance measured value of the plant leaf sample is measured by using the ground object spectrometer and the leaf clamp, the measuring mode has a fixed incident light angle and a fixed radiation height angle, uncertainty in measurement can be greatly reduced, and the high-precision and high-consistency leaf spectral reflectance can be obtained. Wherein, broad-leaved tree species with larger leaf area can be directly measured, and narrow-leaved herbs can be measured by arranging a plurality of leaves. And drying and grinding the blade sample subjected to the spectral measurement, and then obtaining a measured value of the nitrogen content of the blade by using an element analyzer. The process of obtaining the measured reflectance value includes:
and S2, building an inversion model by introducing the specular reflection spectrum radiation into a blade radiation transmission model PROSPECT.
The process of establishing the inversion model comprises the following steps:
s201, bringing the specular reflection spectral radiation into a blade radiation transmission model PROSPECT for forward modeling, and using the obtained spectral reflectance simulation value to establish a cost function;
the method for calculating the spectral reflectivity analog value comprises the following steps:
wherein, the lambda is the wavelength of the spectrum,
， 
，Nstruc, Cab, Cxc, Cw, Narea, Carea, R s respectively representing incident light angle, radiation height angle, leaf structure parameters and chlorophyll content (mug/cm)2) Content of carotenoid (μ g/cm)2) Water content (g/m)2) Leaf Nitrogen content (g/m)2) Carbon compound content (g/m)2) And specularly reflected spectral radiation.
In this embodiment, the leaf structure parameters, chlorophyll content, carotenoid content, moisture content, leaf nitrogen content, carbon-containing compound content and specular reflection spectrum radiation in formula (1) are evaluated according to the measured data set, and the leaf parameter N is calculatedstruc, Cab, Cxc, Cw, Narea, Carea, R s Is set to [1.5,30,5,100,2.26,10,0.2 ]]The minimum value is [1,0.5,0.5,10, 0.02,0.1, -0.2%]The maximum value is set to [3,100,20,440,6.77,200, 0.6%]。
S202, substituting the measured values of the spectral reflectivity into a cost function to carry out preliminary inversion to obtain estimated values of blade structure parameters and estimated values of specular reflection spectral radiation;
the cost function is:
wherein,θis a vector of the parameters of the model,
and
a spectral wavelength start value and an end value representing a cost function; BRFmes(λ) is the actual value of the spectral reflectance analog (i.e., atA measured bi-directional reflectance factor at wavelength λ); BRFsim(λ) is the spectral reflectance analog value (i.e., the simulated bi-directional reflectance factor at wavelength λ using the radiation transmission model parameters); obtaining estimated values of structural parameters of the blade through cost function operation
And specular reflection spectral radiance estimate
。
In the present embodiment, the spectral wavelength λ of formula (2) is set to 400-2500nm according to the full-band spectral reflectance.
S203, substituting the estimated value of the blade structure parameter and the estimated value of the specular reflection spectrum radiation as prior knowledge into the cost function again, determining the weight of the prior knowledge in inversion through normalization processing of the prior knowledge, wherein the weight is used for complementing the cost function.
The completed cost function is:
wherein,
is the maximum value of the structural parameter of the blade,
is the minimum value of the structural parameters of the blade,
is the maximum of the spectral radiance of the specular reflection,
is the spectral radiance minimum of specular reflection,
is the weight of the structural parameter of the blade,
is the weight of specularly reflected spectral radiation.
S3, dividing short-wave infrared spectrum wave bands in the inversion model into a plurality of spectrum subsets according to preset length intervals, taking the spectrum subsets of different wave bands as initial conditions, respectively substituting the spectrum reflectivity measured values corresponding to the first data group and the second data group into two independent inversion models for preliminary inversion to obtain respective first blade nitrogen content estimated values, calculating first root-mean-square errors between the two first blade nitrogen content estimated values and the corresponding blade nitrogen content measured values, taking the first root-mean-square error with the highest precision as a screening condition, and selecting respective semi-discrete optimal spectrum sets from the spectrum subsets through backward iterative recursion.
In the embodiment, the short-wave infrared spectrum band is set to 1400-2399nm and is divided into 20 spectrum subsets S = [ S ] according to the length interval of 50nm1,S2,…,S20]Wherein S is1Is 1400-1449nm, S21450-1499nm, …, S202350 and 2399 nm. Model iterative inversion with 20 different spectral subsets S = [ S ]1,S2,…,S20]As initial conditions.
Taking the first data set as an example, referring to formulas (2) and (3) recorded in steps S202 and S203, the single spectrum subsets are respectively subjected to preliminary inversion to obtain first blade nitrogen content estimation values for 20 spectrum subsets, the first blade nitrogen content estimation values obtained by the 20 spectrum subsets are respectively subjected to Root Mean Square Error (RMSE) calculation with the blade nitrogen content actual measurement values in the first data set, and the spectrum subset S with the highest accuracy of the first blade nitrogen content estimation values is searched and obtainedi(ii) a Secondly, the searched spectrum subset SiAdding the current optimal spectrum set SmaxAnd by combining the current optimal set S of spectramaxWith the remaining 19 spectral subsets S~iForming new 19 spectral subsets S~i,jWherein each spectral subset S~i,jSpectral subset S comprising two 50nm intervalsi(ii) a Again, these 19 different spectral subsets S are used~i,jThe inversion of the step S203 is carried out again, and the spectrum subset with the highest accuracy of the estimated value of the nitrogen content of the first blade at the current stage is obtained through searching; repeating the analogy until all the spectrum subsets with the length interval of 50nm are added into the optimal spectrum set S in sequencemax(ii) a Finally, calculating the Root Mean Square Error (RMSE) of the first blade nitrogen content estimated value obtained by the optimal spectrum set in each iteration process, and sequencing according to the RMSE to obtain the spectrum set S corresponding to the minimum value of the RMSEoptI.e. the final semi-discrete optimal set of spectra for the nitrogen content of the leaves, the order of each subset of spectra into the optimal set of spectra represents its importance for the model inversion.
Table 1 shows that the optimal set of the first data set is 2150-.
Fig. 2a,2b illustrate the relationship between the first root mean square error and the corresponding measured value of the nitrogen content of the blade for the first data set and the second data set, respectively. The X axis represents the sequence of the semi-discrete optimal spectrum subset obtained by the inversion of the backward iteration recursion selection model, and the Y axis represents the iteration sequence. The "+" sign represents the subset of spectra for which the minimum RMSE was obtained for the current iteration order; the asterisks indicate the subset of spectra that yielded the RMSE minimum over all iterations. The first rms error as a whole decreases with iteration order until a minimum is obtained, after which the first rms error increases with the addition of the spectral subset. The first data set requires six spectral subsets to form its semi-discrete optimal set of spectra, while the second data set requires four spectral subsets.
FIGS. 3a and 3b show that the root mean square error of the leaf nitrogen content estimation of the first data set is 0.65 g/m using radiation transmission model inversion and semi-discrete optimal spectrum aggregation2And the error of the second data set is 0.57 g/m2Slightly lower than the first data set. Meanwhile, the estimation errors of the leaf nitrogen contents of the two data sets are small, and the relative root mean square errors are within 20 percent.
And S4, the two groups of semi-discrete optimal spectrum sets are crossly substituted into the inversion model which is subjected to preliminary inversion for cross validation to obtain respective second blade nitrogen content estimated values, a second root mean square error between the two second blade nitrogen content estimated values and the corresponding blade nitrogen content measured values is calculated, and if the second root mean square error is within a preset value, the inversion model is output.
In the present embodiment, the semi-discrete optimal spectrum sets corresponding to the first data set and the second data set obtained in step S3 are cross-input to the inversion model that has undergone preliminary inversion for cross validation. Taking the first data group as an example, that is, the semi-discrete optimal spectrum set obtained by using the first data group is used in the inversion model of the second data group that is subjected to the preliminary inversion in step S3, the second blade nitrogen content estimated value is obtained by inversion, the second blade nitrogen content estimated value obtained by the first data group through the cross-inversion and the blade nitrogen content measured value measured by the second data group are used to calculate the second root mean square error, and the inversion accuracy of the inversion model and the universality of the inversion strategy are measured by the second root mean square error.
FIGS. 4a and 4b show that the root mean square error of the estimation of the leaf nitrogen content is 0.61 g/m by using the semi-discrete optimal spectrum set of the second data set for the first data set, which is obtained by the inversion cross validation of the radiation transmission model2And using the semi-discrete optimal spectral set of the first data set for the second data set with an estimation error of 0.59 g/m2. Therefore, the root mean square error of leaf nitrogen content estimation obtained by inversion strategy cross validation is similar to leaf nitrogen obtained by utilizing inversion semi-discrete optimal spectrum set of respective data setsThe content estimation root mean square error, namely the radiation transmission model inversion strategy has extremely high universality and generalizability.
It should be noted that the first and second data sets, the first blade nitrogen content estimated value and the second blade nitrogen content estimated value, and the "first" and "second" used in the first root mean square error and the second root mean square error are only used to indicate the attributes of the data. For example, the first data group represents the measured values of the spectral reflectance and the nitrogen content of the leaves measured by the leaf sample from China, and the second data group represents the measured values of the spectral reflectance and the nitrogen content of the leaves measured by the leaf sample from America; the first blade nitrogen content estimated value represents the blade nitrogen content estimated values respectively calculated by the first data group and the second data group in the preliminary inversion; the second blade nitrogen content estimated value represents the blade nitrogen content estimated values obtained by respectively calculating two groups of semi-discrete optimal spectrum sets in cross inversion according to a second data group and a first data group; the first root mean square error represents the root mean square error between a first blade nitrogen content estimated value obtained by the first data set through preliminary inversion and a blade nitrogen content measured value in the first data set, and also represents the root mean square error between the first blade nitrogen content estimated value obtained by the second data set through preliminary inversion and a blade nitrogen content measured value in the second data set; the second root mean square error represents the root mean square error between the second blade nitrogen content estimated value obtained by the first data set through cross inversion and the blade nitrogen content measured value in the second data set, and also represents the root mean square error between the first blade nitrogen content estimated value obtained by the second data set through cross inversion and the blade nitrogen content measured value in the first data set.
The above embodiments are only for illustrating the technical concept and features of the present invention, and the purpose thereof is to enable those skilled in the art to understand the contents of the present invention and implement the present invention accordingly, and not to limit the protection scope of the present invention accordingly. All equivalent changes or modifications made in accordance with the spirit of the present disclosure are intended to be covered by the scope of the present disclosure.
Claims (7)
1. A method for estimating the nitrogen content of a blade based on radiation transmission model inversion is characterized by comprising the following steps:
collecting two groups of leaf samples of forest and grassland ecosystems located in different continents, respectively obtaining a spectral reflectance measured value and a leaf nitrogen content measured value corresponding to the two groups of leaf samples, and respectively defining the two groups of leaf samples as a first data group and a second data group;
bringing the specular reflection spectral radiation into a leaf radiation transmission model PROSPECT to establish an inversion model;
dividing short-wave infrared spectrum bands in the inversion model into a plurality of spectrum subsets according to preset length intervals, taking the spectrum subsets of different bands as initial conditions, respectively substituting the spectral reflectivity measured values corresponding to the first data group and the second data group into two independent inversion models for preliminary inversion to obtain respective first blade nitrogen content estimated values, calculating a first root-mean-square error between the two first blade nitrogen content estimated values and the corresponding blade nitrogen content measured values, taking the first root-mean-square error with the highest precision as a screening condition, and selecting respective semi-discrete optimal spectrum sets from the spectrum subsets through backward iterative recursion;
and cross-substituting the two groups of semi-discrete optimal spectrum sets into the inversion model subjected to the preliminary inversion for cross validation to obtain respective second blade nitrogen content estimated values, calculating a second root mean square error between the two second blade nitrogen content estimated values and the corresponding blade nitrogen content measured values, and outputting the inversion model if the second root mean square error is within a preset value.
2. The method for estimating the nitrogen content of the blade based on the radiation transmission model inversion, according to claim 1, wherein the establishing process of the inversion model comprises the following steps:
bringing the specular reflection spectral radiation into the blade radiation transmission model PROSPECT for forward modeling, and using the obtained spectral reflectivity analog value for establishing a cost function;
substituting the measured values of the spectral reflectivity into the cost function to carry out preliminary inversion to obtain estimated values of blade structure parameters and estimated values of specular reflection spectral radiation;
and substituting the estimated values of the blade structural parameters and the estimated values of the specular reflection spectrum radiation as prior knowledge into the cost function again, processing the prior knowledge through normalization, and determining the weight occupied by the prior knowledge in inversion, wherein the weight is used for complementing the cost function.
3. The method for estimating the nitrogen content of the blade based on the radiation transmission model inversion according to claim 2, wherein the spectral reflectance simulation value is calculated by:
wherein, the lambda is the spectrum wavelength,
，
，Nstruc, Cab, Cxc, Cw, Narea, Carea, R s respectively representing incident ray angle, radiation height angle, leaf structure parameter, chlorophyll content, carotenoid content, moisture content, leaf nitrogen content, carbon-containing compound content and specular reflection spectrum radiation.
4. The method of estimation of vane nitrogen content based on radiation transport model inversion of claim 3, wherein the cost function is:
wherein,θis a vector of the parameters of the model,
and
a spectral wavelength start value and an end value representing a cost function; BRFmes(λ) is the measured value of the spectral reflectance; BRFsim(λ) is a spectral reflectance analog value; obtaining the estimated value of the blade structure parameter through the cost function operation
And the specular reflection spectral radiance estimate
。
5. The method of estimating leaf nitrogen content based on radiation transport model inversion of claim 4, wherein the cost function that is completed is:
wherein,
is the maximum value of the structural parameter of the blade,
is the minimum value of the structural parameters of the blade,
is the maximum of the spectral radiance of the specular reflection,
is the spectral radiance minimum of specular reflection,
is the weight of the structural parameter of the blade,
is the weight of specularly reflected spectral radiation.
6. The method for estimating the leaf nitrogen content based on the radiation transmission model inversion as claimed in claim 1, wherein the short wave infrared spectrum band is set to 1400-2399nm and is divided into 20 spectrum subsets at the interval of 50 nm.
7. The method of claim 6, wherein the step of selecting a respective set of semi-discrete optimal spectra from the subset of spectra by backward iterative recursion comprises: dividing the first data group and the second data group into 20 spectrum subsets at length intervals of 50nm as initial conditions, respectively performing the preliminary inversion to obtain first blade nitrogen content estimated values of the 20 spectrum subsets, respectively performing root mean square error calculation on the first blade nitrogen content estimated values of the 20 spectrum subsets and blade nitrogen content measured values in corresponding data groups, searching to obtain a spectrum subset with the highest accuracy of the first blade nitrogen content estimated values, adding the searched spectrum subset into the current optimal spectrum set, forming new 19 spectrum subsets by combining the current optimal spectrum set and the remaining 19 spectrum subsets, wherein each spectrum subset comprises two spectrum subsets at intervals of 50nm, performing the preliminary inversion again by using the 19 different spectrum subsets, and searching to obtain a spectrum subset with the highest accuracy of the first blade nitrogen content estimated values at the current stage, and repeating iteration until all the spectrum subsets with the length of 50nm are sequentially added into the optimal spectrum set, calculating the root mean square error of the first blade nitrogen content estimated value obtained by the optimal spectrum set in each iteration process, and sequencing according to the root mean square error to obtain the spectrum set corresponding to the root mean square error minimum value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210357777.8A CN114446408B (en) | 2022-04-07 | 2022-04-07 | Leaf nitrogen content estimation method based on radiation transmission model inversion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210357777.8A CN114446408B (en) | 2022-04-07 | 2022-04-07 | Leaf nitrogen content estimation method based on radiation transmission model inversion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114446408A CN114446408A (en) | 2022-05-06 |
| CN114446408B true CN114446408B (en) | 2022-07-12 |
Family
ID=81359450
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210357777.8A Active CN114446408B (en) | 2022-04-07 | 2022-04-07 | Leaf nitrogen content estimation method based on radiation transmission model inversion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114446408B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107505271A (en) * | 2017-07-13 | 2017-12-22 | 北京农业信息技术研究中心 | Plant nitrogen evaluation method and system based on nitrogen fractions radiative transfer model |
| CN111929261A (en) * | 2020-09-16 | 2020-11-13 | 广州地理研究所 | Hyperspectral vegetation index-based leaf polyphenol content estimation method |
| CN113552096A (en) * | 2021-07-20 | 2021-10-26 | 中国热带农业科学院湛江实验站 | Spectrum-based pineapple leaf nitrogen content estimation method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110455722A (en) * | 2019-08-20 | 2019-11-15 | 中国热带农业科学院橡胶研究所 | Rubber tree blade phosphorus content EO-1 hyperion inversion method and system |
| CN113076692B (en) * | 2021-03-29 | 2021-09-28 | 中国农业科学院农业资源与农业区划研究所 | Method for inverting nitrogen content of leaf |
-
2022
- 2022-04-07 CN CN202210357777.8A patent/CN114446408B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107505271A (en) * | 2017-07-13 | 2017-12-22 | 北京农业信息技术研究中心 | Plant nitrogen evaluation method and system based on nitrogen fractions radiative transfer model |
| CN111929261A (en) * | 2020-09-16 | 2020-11-13 | 广州地理研究所 | Hyperspectral vegetation index-based leaf polyphenol content estimation method |
| CN113552096A (en) * | 2021-07-20 | 2021-10-26 | 中国热带农业科学院湛江实验站 | Spectrum-based pineapple leaf nitrogen content estimation method |
Non-Patent Citations (1)
| Title |
|---|
| 不同树种叶片养分含量提取的高光谱方法及精度评价;孙中宇等;《热带地理》;20200331;第40卷(第2期);第175-183页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114446408A (en) | 2022-05-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Serbin et al. | From the Arctic to the tropics: multibiome prediction of leaf mass per area using leaf reflectance | |
| Jin et al. | Deep neural network algorithm for estimating maize biomass based on simulated Sentinel 2A vegetation indices and leaf area index | |
| CN109115725B (en) | Corn canopy LAI and chlorophyll content joint inversion method and equipment | |
| Stroppiana et al. | Plant nitrogen concentration in paddy rice from field canopy hyperspectral radiometry | |
| CN111855589A (en) | Remote sensing inversion model and method for rice leaf nitrogen accumulation | |
| CN107271382A (en) | A kind of different growing rape leaf SPAD value remote sensing estimation methods | |
| Lunagaria et al. | Evaluation of PROSAIL inversion for retrieval of chlorophyll, leaf dry matter, leaf angle, and leaf area index of wheat using spectrodirectional measurements | |
| Whitehead et al. | Photosynthesis and reflectance indices for rainforest species in ecosystems undergoing progression and retrogression along a soil fertility chronosequence in New Zealand | |
| Zhang et al. | Monitoring daily variation of leaf layer photosynthesis in rice using UAV-based multi-spectral imagery and a light response curve model | |
| Sonobe et al. | Estimating leaf carotenoid contents of shade-grown tea using hyperspectral indices and PROSPECT–D inversion | |
| CN113252583B (en) | Method for calculating alpine hay coverage based on hay vegetation index | |
| Mengistu et al. | Sun-induced fluorescence and near-infrared reflectance of vegetation track the seasonal dynamics of gross primary production over Africa | |
| CN110455726A (en) | A kind of method of real-time Forecasting Soil Moisture and total nitrogen content | |
| CN115170341A (en) | Forestry resource carbon sink accurate measurement and calculation method for correcting small samples by utilizing laser radar | |
| Qian et al. | Exploring the potential of leaf reflectance spectra for retrieving the leaf maximum carboxylation rate | |
| CN118072178B (en) | Corn yield estimation method and system based on classified percentage data assimilation | |
| CN110569605A (en) | Non-glutinous rice leaf nitrogen content inversion model method based on NSGA2-ELM | |
| Yuan et al. | Research on rice leaf area index estimation based on fusion of texture and spectral information | |
| Du et al. | Evaluation of spectral scale effects in estimation of vegetation leaf area index using spectral indices methods | |
| CN112257531B (en) | Remote sensing monitoring method for forest land change based on diversity feature combination | |
| Ollila | The roles of greenhouse gases in global warming | |
| CN114446408B (en) | Leaf nitrogen content estimation method based on radiation transmission model inversion | |
| Jin et al. | Combing both simulated and field-measured data to develop robust hyperspectral indices for tracing canopy transpiration in drought-tolerant plant | |
| Kong et al. | An integrated field and hyperspectral remote sensing method for the estimation of pigments content of Stipa Purpurea in Shenzha, Tibet | |
| CN111027523A (en) | Satellite remote sensing monitoring method for carotenoid content in cotton canopy |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |