CN110160967A - A kind of total nitrogen content evaluation method of crop canopies blade - Google Patents
A kind of total nitrogen content evaluation method of crop canopies blade Download PDFInfo
- Publication number
- CN110160967A CN110160967A CN201910305370.9A CN201910305370A CN110160967A CN 110160967 A CN110160967 A CN 110160967A CN 201910305370 A CN201910305370 A CN 201910305370A CN 110160967 A CN110160967 A CN 110160967A
- Authority
- CN
- China
- Prior art keywords
- reflectivity
- nitrogen content
- wave band
- total nitrogen
- canopy
- 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.)
- Pending
Links
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
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
-
- 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
- G01N2021/1797—Remote sensing in landscape, e.g. crops
Abstract
The invention belongs to crop biochemical component spectrum technical field of nondestructive testing, disclose a kind of total nitrogen content evaluation method of crop canopies blade: first, algorithm SPA, sensitivity spectrum characteristic data set, position feature data set and vegetation index characteristic data set are converted using successive projection；Then, the small the smallest sensitivity spectrum variables set of synteny is filtered out from sensitivity spectrum characteristic data set, position feature data set and vegetation index characteristic concentration；Finally, sensitivity spectrum variables set is carried out Partial Least Squares Regression PLS modeling, the appraising model of the total nitrogen content LNC of crop canopies blade is obtained.The model can accurately estimate crop canopies leaf total nitrogen content, be not much different with measured value.
Description
Technical field
The invention belongs to crop biochemical component spectrum technical field of nondestructive testing, in particular to a kind of crop canopies blade
Total nitrogen content evaluation method.
Background technique
Nitrogen is photosynthesis of plant and diagnosis crop nutrition condition necessary element, it and crop growth, length
Gesture and quality is closely related and the emphasis of agronomy Experts ' Attention.Normally apply under nitrogen, plant amount of nitrogen sucking, fertilizer utilization efficiency,
Crop yield and quality increase with the increase of amount of nitrogen.If but nitrogenous fertilizer crosses Sheng, will cause assessing index, rice produces
The problems such as amount and quality decline, it also will cause land pollution when serious.Meanwhile the variation of crop leaf Different Nitrogen Concentration will affect plant
The photosynthesis of object blade, so that certain variation occurs for leaf spectra.When crop leaf is compromised stress, leaf-nitrogen
Concentration variation can spectrally generate response.Therefore, plant nitrogen situation can be monitored by the variation of leaf spectra.
High-spectral data spectral resolution with higher, can detect the minor change of plant leaf blade biochemical component,
Detection has the characteristics that lossless, quick, has carried out quick, lossless monitoring to plant leaf blade biochemical component using EO-1 hyperion means
Important content as vegetation growth of plant status evaluation.However, the high-spectral data with hundreds or even thousands wave band, is compared
Although can provide more careful richer spectral information in multispectral, there is also high between mass data redundancy, neighbouring wave band
The problems such as spending correlation.
Currently, the spectral variables extracted based on hyperspectral information can be generally divided into three classes: (1) reflected waveband feature: high
Spectrum has more accurate wave band, can be good at the feature for reacting vegetation.Bai Limin et al. is mentioned using successive projection algorithm
Take Wheat Leavess total nitrogen content canopy spectra 8 (1985nm, 2474nm, 1751nm, 1916nm, 2507nm, 1955nm,
2465nm, 344nm) sensitive band, wheat during jointing stage blade content appraising model is established, estimation precision is high, and stability is good；Zhang Xiao
Flower bud et al. based on successive projection algorithm extract 12 characteristic wavelengths of Brassica campestris L seedling-flower-silique time of infertility blade (467nm,
557nm, 665nm, 686nm, 706nm, 752nm, 874nm, 879nm, 886nm, 900nm, 978nm, 995nm), inverting blade
Nitrogen content estimation models, realize the visualization of nitrogen level；(2) sensitive position feature: Fu Yuanyuan et al. is to canopy of winter wheat
Wave band depth analysis is carried out within the scope of spectrum 550nm-750nm, then to improve crop field winter wheat in conjunction with Partial Least Squares Regression
The estimation precision of biomass；Wang Ping et al. extracts reflection to EO-1 hyperion visible light wave range (400nm~800nm) and absorbs position etc.
Parameter, pick out ginseng higher to Chlorophyll from Maize correlation and relevant with corn pollution stress chlorophyll minor change
Number, establishes BP neural network model, gradually enhances and extract the small change of Chlorophyll from Maize under farmland pollution stress state
Change information；Du Huaqiang et al. analyzes 9 position features such as red side area, red paddy absorption depth in masson pine spectrum, establishes
The Folium Pini chlorophyll content prediction model of Partial Least Squares Regression based on gaussian kernel function transformation, mean square error
It is 0.0088, average absolute percentage error 0.7617%；(3) vegetation index feature: vegetation index is close red by visible light-
The combination of the linearity and non-linearity of wave section, can quantitative response vegetation in any case upgrowth situation.Chen et al. is utilized
Bimodal canopy nitrogen index (DCNI) is compared with some existing vegetation indexs, determines that the DCNI of wheat and corn is in research
Optimal spectrum index, to assess the efficiency of crop nitrogen processing；Shu Meiyan et al. improves crop by constructing novel spectral index
The precision of leaf area index EO-1 hyperion inverting；Tan Changwei et al. comprehensive analysis 10 common spectral vegetation indexes and summer corn
The correlation and predictability of LAI.
Research shows that: current existing research uses EO-1 hyperion characteristics of variables, although monitoring and evaluation plant growth shape can be used for
Condition parameter, but most researchs only use the spectral signature variable of single type, and different type spectral variables are all from difference
Angle provides the useful information of plant growth parameter, therefore, if different types of spectral variables can be comprehensively utilized to monitor and make
Object target component, which is one, may improve the approach of monitoring accuracy, and how to comprehensively utilize polymorphic type EO-1 hyperion variable and further dig
The abundant information contained in pick EO-1 hyperion, the precision of crop nitrogen nutrition spectrum estimation is improved with this, is rarely reported.Another party
Face, Partial Least Squares Regression PLS method are the extensions of multiple linear regression model, because it can reduce being total between data variable
Linear problem and be widely used.
Summary of the invention
It is an object of the present invention to for existing using spectrum existing for EO-1 hyperion variable monitoring and evaluation crop nutrition aspect
The excessively single problem of characteristic variable, the application provide a kind of total nitrogen content evaluation method of crop canopies blade.
To achieve the goals above, the technical solution that the application uses are as follows:
A kind of total nitrogen content evaluation method of crop canopies blade, comprising steps of
S1, algorithm SPA is converted using successive projection, from 400nm~1350nm, 1400nm~1790nm, 1990nm~
It is selected in the canopy leaves spectral region of 2350nm 6 the smallest with the total nitrogen content LNC1 synteny of known crop canopies blade
Sensitive band constitutes sensitivity spectrum characteristic data set；
6 sensitive bands include for 4 canopy leaves reflectivity Ref wave bands and 2 first derivative FD wave bands；
4 canopy leaves reflectivity Ref wave bands are respectively 412nm, 724nm, 1084nm, 1343nm；
2 first derivative FD wave bands are 658nm and 937nm；
The total nitrogen content LNC1 of the known crop canopies blade detects to obtain by azotometer；
S2, algorithm SPA is converted using successive projection, in the position wave band used when being studied from crop EO-1 hyperion nitrogen, choosing
The smallest two positions wave band SDb and Dr of synteny is taken, position feature data set is constituted；
The SDb is the sum of the canopy leaves spectral reflectivity first derivative FD of wave band at 490nm~530nm；
The Dr is the canopy leaves spectral reflectivity first derivative FD maximum value of wave band at 680nm~760nm；
S3, SPA algorithm is converted using successive projection, it is two the smallest that synteny is chosen from typical vegetation index VIs
Vegetation index, i.e. normalized differential vegetation index NDVIg-b #With difference vegetation index DVI II#, constitute vegetation index characteristic data set；
The NDVIg-b #=(R573-R440)/(R573+R440)；
Wherein, R573The canopy leaves spectral reflectivity for being wavelength at 573nm；
R440The canopy leaves spectral reflectivity for being wavelength at 440nm；
The DVI II#=RNIR-RR；
Wherein, RNIRFor the mean value of the canopy leaves spectral reflectivity of wave band at 841nm~876nm；
RRFor the canopy leaves spectral reflectivity of wave band at 680nm~760nm；
S4, SPA algorithm is converted using successive projection, from sensitivity spectrum characteristic data set, position feature data set and vegetation
The small the smallest sensitivity spectrum variables set of synteny is filtered out in index characteristic data set；
Wherein, the sensitivity spectrum variables set includes 2 canopy leaves reflectivity Ref wave bands, 1 first derivative FD wave band
With normalized differential vegetation index NDVIg-b #；
2 canopy leaves reflectivity Ref wave bands are respectively 724nm and 1343nm；
1 order derivative FD wave band is 658nm；
Sensitivity spectrum variables set is subjected to Partial Least Squares Regression PLS modeling again, the full nitrogen for obtaining crop canopies blade contains
Measure the appraising model of LNC: Y=3.6344-5.7351X1-4.4275X2+1236.6X3+2.0886X4；
Wherein, X1: wave band is the canopy leaves reflectivity Ref at 1343nm；
X2: normalized differential vegetation index NDVIg-b #；
X3: wave band is the first derivative values of reflectivity at 658nm；
X4: wave band is the canopy leaves reflectivity Ref at 724nm.
Further, the 400nm~1350nm, 1400nm~1790nm, the canopy leaves light of 1990nm~2350nm
Spectral limit is chosen obtain by the following method:
Reflectivity of the crop canopies blade in 350nm~2500nm wave-length coverage is measured, obtaining abscissa is wavelength, is indulged
Coordinate is the spectrogram of reflectivity；Removing steam influences, and remaining wave band is 400nm~1350nm, 1400nm in spectrogram
The canopy leaves spectrum of~1790nm and 1990nm~2350nm.
Further, the total nitrogen content LNC1 of the known crop canopies blade is obtained by the following method:
Crop canopies blade after progress spectral reflectivity measurement is put into 100 DEG C~110 DEG C of baking oven and dries 25min
~35min is dried to weight at then 75 DEG C~85 DEG C, obtains leaf samples；The leaf samples for weighing drying, then by sample leaf
Piece crushes, then measures leaf samples nitrogen content with azotometer, obtains LNC 1.
Further, the formula of the first derivative FD is as follows:
Wherein, Rλ(j)It is the reflectivity of wave band j, Rλ(j+1)It is the reflectivity of wave band j+1, λ (j+1)-λ (j) is wave band j and j
Wavelength difference between+1；
The first derivative of first derivative FD reflectivity of wavelength midpoint i between wave band j and j+1.
Further, the crop canopies blade is maize canopy blade or wheat canopy blade.
Compared with prior art, the beneficial effects of the present invention are:
The application converts SPA algorithm using successive projection, from sensitivity spectrum characteristic data set, position feature data set and plant
The small the smallest sensitivity spectrum variables set of synteny is filtered out in index characteristic data set；Sensitivity spectrum variables set is carried out again inclined
Least square regression PLS modeling, obtains the appraising model of the total nitrogen content LNC of crop canopies blade.Because establishing crop canopies
EO-1 hyperion sensitive band, the spectral position that canopy leaves have been taken into account during the appraising model of the total nitrogen content LNC of blade are special
It seeks peace typical vegetation index these three types spectral variables, compared to existing using single spectral variables monitoring and evaluation crop canopies
Leaf total nitrogen content, stability is more preferable, more can accurately predict the total nitrogen content LNC of the crop canopies blade in sample to be tested.
Detailed description of the invention
Fig. 1 is the correlation spectrogram of training set reflectance spectrum (Ref) and first derivative spectrum (FD) and LNC 1 of the present invention；
Fig. 2 is the practical total nitrogen content LNC' of maize canopy blade of the present invention and the prediction total nitrogen content of maize canopy blade
The precision evaluation figure of LNC.
Specific embodiment
Technological means of the invention, creation characteristic, achieving the goal is easy to understand with effect in order to make, below in conjunction with
Specific embodiments of the present invention and attached drawing, are clearly and completely described technical solution of the present invention.
One, Preparatory work of experiment
(1) area is studied:
The crop used in the embodiment of the present invention is by taking maize leaves as an example, and experiment is in 2012 in national precision agriculture research
Demonstration Base is carried out, and base is located at the Changping District, Beijing northeast little Tang Shan Zhen, and 40 ° 00 '~40 ° 21 ' of north latitude, 116 ° of east longitude
34 '~117 ° 00 ', height above sea level 36m.Soil types is moisture soil.Climate type is the continental monsoon climate in temperate zone, and winter is cold
Cold, summer is hot, throughout the year drought.The corn of plantation is Nongda108 (osculant) and capital 8 (compact).3 nitrogens
Nitrogen (N0), normal nitrogen (N1), 2 times of nitrogen (N2), 3 repetitions are not applied in horizontal processing respectively.Thickness of sowing is about 882m2。2012
Sowing on June 21, in, on October 15th, 2012 harvest.
(2) spectroscopic assay
Spectroscopic assay uses ASD FieldSpec FR2500 type field spectroradiometer (U.S. Analytical Spectral
Device company), spectral region 350nm-2500nm is spaced 1.4nm, 1000nm- within the scope of 350nm-1000nm
1nm is spaced within the scope of 2500nm.Generally measured in Beijing time 10:00-14:00, when measurement fine, calm nothing
Cloud, probe vertical is downward when observation, from the ground height 1.3m, and 25 ° of field angle, measurement front and back is corrected with reference to version every time.
(3) measurement of the total nitrogen content LNC1 of crop canopies blade known to
Cauline leaf is separated in time after carrying out spectral measurement, and blade is fitted into paper bag.Blade is put into 105 ° of baking
30min is dried in case, 48h or more is then dried under 80 ° of constant temperature, until weight.The leaf samples for weighing drying, then by blade powder
It is broken, its nitrogen content is measured with kjeldahl apparatus (Buchi B-339, FOSS, Sweden).Totally 72 hillslope processes in this research,
48 cells are for training, and 24 cells are for verifying.
Two, principle and method
(1) high-spectral data pre-processes
To eliminate the partial noise in spectrum, spectrum denoising is carried out using Savitzky_Golay (SG) convolution exponential smoothing,
By preliminary experiment, moving window width is 17, and when the degree of polynomial is 2, denoising effect is preferable.All kinds of spectral variables herein are such as
First derivative, position feature and vegetation index etc., the spectral reflectivity after being all based on SG denoising are calculated, and wherein single order is led
Number (FD) formula is as follows:
Wherein, Rλ(j)It is the reflectivity of wave band j, Rλ(j+1)It is the reflectivity of wave band j+1, λ (j+1)-λ (j) is wave band j and j
Wavelength difference between+1；
The first derivative of first derivative FD reflectivity of wavelength midpoint i between wave band j and j+1.
(2) the position wave band used when crop EO-1 hyperion nitrogen is studied
The position wave band that 1 crop EO-1 hyperion nitrogen of table is used when studying
(3), typical vegetation index VIs
It has been reported that many different optical indexs in document, and verified they and vegetation parameter have well
Correlation.This research has chosen 34 typical vegetation indexs (VIs) to estimate maize canopy leaf N content (being shown in Table 2).Wherein,
Including 6 EO-1 hyperion vegetation indexs sensitive to nitrogen, such as best vegetation index (Viopt), normalized site attenuation
(NDVIg-b #), ratio vegetation index (RVI I#)、RVI II#, association index (MCARI/MTVI2), bimodal canopy nitrogen index
(DCNI#), there are also NDVI I#、RVI III#、DVI I#、SAVI I#, the normalization red side of difference (NDRE) etc.；And 23 typical cases
Index, such as anti-atmosphere vegetation index (ARVI), difference vegetation index (DVI II#), enhance vegetation index (EVI), green normalizing
Change vegetation index (GNDVI), improve non-linear vegetation index (MNLI), modified soil adjusts vegetation index 2 (MSAVI2), repairs
Just simply than (MSR), NDVI II#, non-linear vegetation index (NLI), optimization SAVI (OSAVI), RDVI, ratio vegetation index
(RVI IV), SAVI II#, TVI and improvement triangle vegetation index (MTVI2), index NDVI relevant to red sideRed-edge,
CIRed-edgeAnd MTCI, and with the associated index of moisture (WI, NDWI), normalized difference infrared index (NDII), disease moisture
It coerces index (DSWI) and standardization LAI determines index (sLAIDI*) etc., be related to the VIs of broadband information to above-mentioned, using pair
The spectral response functions of inductive sensing device are calculated by EO-1 hyperion.
The typical vegetation index VIs of table 2 summarizes
(4) screening and modeling method
(1) successive projection converts algorithm SPA
In recent years, successive projection algorithm (successive projections algorithm, SPA) is being screened and is being extracted
Increasingly extensive in the application of sensitive variable, it is a kind of forward variable choosing that can effectively eliminate synteny problem in spectral information
Algorithm is selected, by reducing the redundancy between variable, representative characteristic parameter is selected for modeling, is greatly improved modeling
The efficiency of analysis.With root-mean-square error (root mean square error, RMSE) for evaluation index, work as root-mean-square error
Number when minimum is variable number.
(2) Partial Least Squares Regression
Partial Least Squares Regression (Partial Least Squares Regression, PLS) is one of statistics
Method, it integrates principal component analysis, canonical correlation analysis and multiple linear regression analysis method, be a kind of research multivariate response or
The modeling pattern of single dependent variable and more independents variable.It can be filtered out in the case where sample size is few synteny it is lesser at
Point.
Consider m dependent variable y1,y2,…,ym, n independent variable x1,x2,…,xn。E0,E1,…,Er,F0,…,FrIt is two groups
Data matrix is observed in the standardization of variables set, directly in E0,E1,…,Er-1Extract component t in matrix1,…,tr (r≤m),th(th
It is independent variable collection X=(x1,…,xm)TLinear combination) it is as much as possible carry X in information, meanwhile, thTo dependent variable system
F0There is maximum interpretability.If finally extracting r ingredient t to independent variable collection1,…,tr, Partial Least Squares Regression will be by building
Vertical y1,y2,…,ymWith t1,…,trRegression equation, be then expressed as y again1,y2,…,ymWith the regression equation of former independent variable, i.e.,
Partial Least Squares Regression equation.
(3) statistical analysis technique
Phase is carried out with SPSS 24.0 to sensitive band, position feature, vegetation index and maize canopy leaf-nitrogen (LNC)
The analysis of closing property.Verifying sample, coefficient of determination R are used as using 1/3 (24) sample for having neither part nor lot in modeling2, root-mean-square error RMSE
With normalization relative error NRMSE (normalized rootmeanaquareerror) 3 indexs as index explanation, quantify
Change the relationship between canopy leaves total nitrogen content.
Wherein xiIndicate the measured value of i-th of maize canopy leaf N content, i.e. LNC 1；yiIndicate i-th of maize canopy leaf
The predicted value of nitrogen content；It is expressed as the measured value mean value of the maize canopy leaf N content of forecast set or actual measurement collection, whereinN indicates sample size.
Three, a kind of modeling process of the total nitrogen content evaluation method of crop canopies blade
A kind of modeling process of the total nitrogen content evaluation method of crop canopies blade, includes the following steps:
The selection of S1, sensitivity spectrum characteristic data set
(1) arbitrarily choose the canopy leaves of corn in 48 training centers, measurement maize canopy blade 350 nm~
Reflectivity in 2500nm wave-length coverage, obtaining abscissa is wavelength, and ordinate is the spectrogram of reflectivity；Remove steam shadow
It rings, remaining wave band is 400nm~1350nm, 1400nm~1790nm, the canopy leaf of 1990nm~2350nm in spectrogram
Piece spectrum, as a result as shown in Figure 1.
(2) the crop canopies blade after above-mentioned spectral reflectivity measurement will be carried out and is put into 105 DEG C of baking oven to dry
Then 30min is dried to weight at 80 DEG C, obtains leaf samples；The leaf samples for weighing drying, then crush leaf samples,
Leaf samples nitrogen content is measured with azotometer again, obtains the total nitrogen content LNC 1 of known crop canopies blade.
(3) algorithm SPA is converted using successive projection, from 400nm~1350nm, 1400nm~1790nm, 1990nm~
It is selected in the canopy leaves spectral region of 2350nm and the smallest 6 sensitive bands of LNC1 synteny, composition sensitivity spectrum feature
Data set；
6 sensitive bands include for 4 canopy leaves reflectivity Ref wave bands and 2 first derivative FD wave bands；
4 canopy leaves reflectivity Ref wave bands are respectively 412nm, 724nm, 1084nm, 1343nm；2 first derivatives
FD wave band is 658nm and 937nm；
The selection of S2, position feature data set
Algorithm SPA is converted using successive projection, the position wave band that the crop EO-1 hyperion nitrogen shown in the table 1 is used when studying
In, the smallest two positions wave band SDb and Dr of synteny is chosen, position feature data set is constituted；
SDb is the sum of the canopy leaves spectral reflectivity first derivative FD of wave band at 490nm~530nm；
Dr is the canopy leaves spectral reflectivity first derivative FD maximum value of wave band at 680nm~760nm；
The selection of S3, vegetation index characteristic data set
SPA algorithm is converted using successive projection, it is minimum that synteny is chosen from typical vegetation index VIs shown in table 2
Two vegetation indexs, i.e. normalized differential vegetation index NDVIg-b #With difference vegetation index DVI II#, constitute vegetation index characteristic
According to collection；
The NDVIg-b #=(R573-R440)/(R573+R440)；
Wherein, R573The canopy leaves spectral reflectivity for being wavelength at 573nm；
R440The canopy leaves spectral reflectivity for being wavelength at 440nm；
DVI II#=RNIR-RR；
Wherein, RNIRFor the mean value of the canopy leaves spectral reflectivity of wave band at 841nm~876nm；
RRFor the canopy leaves spectral reflectivity of wave band at 680nm~760nm；
The selection of S4, sensitivity spectrum variables set
SPA algorithm is converted using successive projection, from sensitivity spectrum characteristic data set, position feature data set and vegetation index
Characteristic concentration filters out the small the smallest sensitivity spectrum variables set of synteny；
Wherein, sensitivity spectrum variables set include 2 canopy leaves reflectivity Ref wave bands, 1 first derivative FD wave band and
Normalized differential vegetation index NDVIg-b #；
2 canopy leaves reflectivity Ref wave bands are respectively 724nm and 1343nm；
1 first derivative FD wave band is 658nm；
Sensitivity spectrum variables set is subjected to Partial Least Squares Regression PLS modeling again, the full nitrogen for obtaining crop canopies blade contains
Measure the appraising model of LNC: Y=3.6344-5.7351X1-4.4275X2+1236.6X3+2.0886X4；
Wherein, X1: wave band is the canopy leaves reflectivity Ref at 1343nm；
X2: normalized differential vegetation index NDVIg-b #；
X3: wave band is the first derivative values of reflectivity at 658nm；
X4: wave band is the canopy leaves reflectivity Ref at 724nm.
In order to further verify the application crop canopies blade total nitrogen content LNC appraising model stability and standard
True property, has also done following test:
(1) it is tested by ASD FieldSpec FR2500 type field spectroradiometer in 48 training centers (modeling collection) and 24
It demonstrate,proves and arbitrarily takes one plant of plant in each area of area (test set), while measuring 48 training centers (modeling collection) and being tested with 24
The reflectivity of area (test set) is demonstrate,proved, each plant of plant measures 10 times, average, then remove the influence of steam part, use
SPA algorithm filters out spectral reflectance of each plant of maize canopy blade at 440nm, 573nm, 658nm, 724nm and 1341nm
Rate is denoted as R respectively440、R573、R658、R724And R1341。
And then obtain NDVIg-b #=(R573-R440)/(R573+R440) specific value；In addition, passing throughR is calculated658First derivative FD658。
By R1341、NDVIg-b #、FD658And R724It is brought into the estimation mould of the total nitrogen content LNC of crop canopies blade respectively
Type: Y=3.6344-5.7351X1-4.4275X2+1236.6X3+2.0886X4In obtain modeling collection or test set maize canopy
The prediction total nitrogen content LNCY of blade1~Y24。
(2) the practical total nitrogen content of the maize canopy blade of modeling collection or test set is measured respectively with kjeldahl apparatus respectively
LNC'Y'1~Y'24。
(3) by the practical total nitrogen content LNC'Y' of the above-mentioned maize canopy blade for obtaining modeling collection1~Y'24With modeling collection
The prediction total nitrogen content LNC Y of maize canopy blade1~Y24It substitutes intoIn
Obtain coefficient of determination R2；It substitutes intoIn obtain root-mean-square error RMSE；It substitutes intoIn obtain normalization relative error NRMSE；Test set is obtained by above-mentioned
Maize canopy blade practical total nitrogen content LNC'Y'1~Y'24With the prediction total nitrogen content of the maize canopy blade of test set
LNC Y1~Y24It substitutes intoIn obtain coefficient of determination R2；It substitutes intoIn obtain root-mean-square error RMSE；It substitutes intoIn obtain normalization relative error NRMSE, concrete outcome such as table 3.
The practical total nitrogen content LNC''s and prediction total nitrogen content LNC of the maize canopy blade of the modeling of table 3 collection or test set
Statistical analysis
, it can be seen that the canopy obtained using the total nitrogen content evaluation method of the crop canopies blade of the application from table 3
Leaf total nitrogen content is not much different compared to the practical total nitrogen content LNC' of maize canopy blade, illustrates the crop canopies of the application
The total nitrogen content evaluation method of blade can accurately estimate canopy leaves total nitrogen content.
(4) in addition, with the practical total nitrogen content LNC'Y' of test set maize canopy blade1~Y'24For abscissa, with test
Collect the prediction total nitrogen content LNC Y of maize canopy blade1~Y24It for ordinate, and is modeled, is obtained by Partial Least Squares Regression PLS
To result such as Fig. 2, dotted line is the straight line of y=x in Fig. 2, and solid line is the practical total nitrogen content LNC' of test set maize canopy blade
With test set maize canopy blade prediction total nitrogen content LNC simulation curve y=0.8422x+0.2868, from Fig. 2 more into
One step illustrates that the total nitrogen content evaluation method of the crop canopies blade of the application can accurately estimate crop canopies leaf total nitrogen
Content.
In conclusion SPA algorithm is converted using successive projection, from sensitivity spectrum characteristic data set, position feature data set
The small the smallest sensitivity spectrum variables set of synteny is filtered out with vegetation index characteristic concentration；Again by sensitivity spectrum variables set into
Row Partial Least Squares Regression PLS modeling, obtains the appraising model of the total nitrogen content LNC of crop canopies blade, which being capable of essence
True estimation crop canopies leaf total nitrogen content, is not much different with measured value.
When embodiment provides numberical range, it should be appreciated that except non-present invention is otherwise noted, two ends of each numberical range
Any one numerical value can be selected between point and two endpoints.Unless otherwise defined, the present invention used in all technologies and
Scientific term is identical as the normally understood meaning of those skilled in the art of the present technique.Except specific method, equipment used in embodiment,
Outside material, grasp and record of the invention according to those skilled in the art to the prior art can also be used and this
Any method, equipment and the material of the similar or equivalent prior art of method described in inventive embodiments, equipment, material come real
The existing present invention.
Disclosed above is only presently preferred embodiments of the present invention, and still, the embodiment of the present invention is not limited to this, Ren Heben
What the technical staff in field can think variation should all fall into protection scope of the present invention.
Claims (6)
1. a kind of total nitrogen content evaluation method of crop canopies blade, which is characterized in that comprising steps of
S1, algorithm SPA is converted using successive projection, is selected from canopy leaves spectral region complete with known crop canopies blade
The smallest 6 sensitive bands of nitrogen content LNC1 synteny constitute sensitivity spectrum characteristic data set；
6 sensitive bands are 4 canopy leaves reflectivity Ref wave bands and 2 first derivative FD wave bands；
4 canopy leaves reflectivity Ref wave bands are respectively 412nm, 724nm, 1084nm, 1343nm；
2 first derivative FD wave bands are 658nm and 937nm；
S2, algorithm SPA is converted using successive projection, in the position wave band used when being studied from crop EO-1 hyperion nitrogen, chosen altogether
Linear the smallest two positions wave band SDb and Dr, constitute position feature data set；
SDb is the sum of the canopy leaves spectral reflectivity first derivative FD of wave band at 490nm~530nm；
Dr is the canopy leaves spectral reflectivity first derivative FD maximum value of wave band at 680nm~760nm；
S3, SPA algorithm is converted using successive projection, the smallest two vegetation of synteny is chosen from typical vegetation index VIs
Index, i.e. normalized differential vegetation index NDVIg-b #With difference vegetation index DVI II#, constitute vegetation index characteristic data set；
NDVIg-b #=(R573-R440)/(R573+R440)；
Wherein, R573The canopy leaves spectral reflectivity for being wavelength at 573nm；
R440The canopy leaves spectral reflectivity for being wavelength at 440nm；
DVI II#=RNIR-RR；
Wherein, RNIRFor the mean value of the canopy leaves spectral reflectivity of wave band at 841nm~876nm；
RRFor the canopy leaves spectral reflectivity of wave band at 680nm~760nm；
S4, SPA algorithm is converted using successive projection, from sensitivity spectrum characteristic data set, position feature data set and vegetation index
Characteristic concentration filters out the small the smallest sensitivity spectrum variables set of synteny；
Wherein, sensitivity spectrum variables set includes 2 canopy leaves reflectivity Ref wave bands, 1 first derivative FD wave band and normalization
Vegetation index NDVIg-b #；
2 canopy leaves reflectivity Ref wave bands are respectively 724nm and 1343nm；
1 order derivative FD wave band is 658nm；
Sensitivity spectrum variables set is subjected to Partial Least Squares Regression PLS modeling again, obtains the total nitrogen content LNC of crop canopies blade
Appraising model: Y=3.6344-5.7351X1-4.4275X2+1236.6X3+2.0886X4；
Wherein, X1: wave band is the canopy leaves reflectivity Ref at 1343nm；
X2: normalized differential vegetation index NDVIg-b #；
X3: wave band is the first derivative values of canopy leaves reflectivity at 658nm；
X4: wave band is the canopy leaves reflectivity Ref at 724nm.
2. the total nitrogen content evaluation method of crop canopies blade as described in claim 1, which is characterized in that the canopy leaves
Spectral region is 400nm~1350nm, 1400nm~1790nm, 1990nm~2350nm, and the canopy leaves spectral region
It chooses and obtains by the following method:
Reflectivity of the crop canopies blade in 350nm~2500nm wave-length coverage is measured, obtaining abscissa is wavelength, ordinate
For the spectrogram of reflectivity；Removing steam influences, in spectrogram remaining wave band be 400nm~1350nm, 1400nm~
The canopy leaves spectrum of 1790nm and 1990nm~2350nm.
3. the total nitrogen content evaluation method of crop canopies blade as described in claim 1, which is characterized in that the known crop
The total nitrogen content LNC1 of canopy leaves detects to obtain by azotometer.
4. the total nitrogen content evaluation method of crop canopies blade as described in claim 1, which is characterized in that the known crop
The total nitrogen content LNC1 of canopy leaves is obtained by the following method:
By carry out spectral reflectivity measurement after crop canopies blade be put into 100 DEG C~110 DEG C of baking oven dry 25min~
35min is dried to weight at then 75 DEG C~85 DEG C, obtains leaf samples；The leaf samples for weighing drying, then by leaf samples
It crushes, then measures leaf samples total nitrogen content with azotometer, obtain LNC1.
5. the total nitrogen content evaluation method of crop canopies blade as described in claim 1, which is characterized in that the first derivative
The formula of FD is as follows:
Wherein, Rλ(j)It is the reflectivity of wave band j, Rλ(j+1)The reflectivity of wave band j+1, λ (j+1)-λ (j) be wave band j and j+1 it
Between wavelength difference；
The first derivative of first derivative FD reflectivity of wavelength midpoint i between wave band j and j+1.
6. the total nitrogen content evaluation method of crop canopies blade as described in claim 1, which is characterized in that the crop canopies
Blade is maize canopy blade or wheat canopy blade.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910305370.9A CN110160967A (en) | 2019-04-16 | 2019-04-16 | A kind of total nitrogen content evaluation method of crop canopies blade |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910305370.9A CN110160967A (en) | 2019-04-16 | 2019-04-16 | A kind of total nitrogen content evaluation method of crop canopies blade |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110160967A true CN110160967A (en) | 2019-08-23 |
Family
ID=67639621
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910305370.9A Pending CN110160967A (en) | 2019-04-16 | 2019-04-16 | A kind of total nitrogen content evaluation method of crop canopies blade |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110160967A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110567385A (en) * | 2019-09-19 | 2019-12-13 | 廊坊师范学院 | Hyperspectral technology-based construction thickness detection method for building reflective insulation coating |
| CN110987830A (en) * | 2019-12-23 | 2020-04-10 | 陕西地建土地工程技术研究院有限责任公司 | Model, method and application for rapidly determining chlorophyll content of plant canopy leaves |
| CN111650129A (en) * | 2020-03-05 | 2020-09-11 | 广州地理研究所 | Litchi leaf calcium content prediction method and prediction device |
| CN112129709A (en) * | 2020-09-16 | 2020-12-25 | 西北农林科技大学 | Apple tree canopy scale nitrogen content diagnosis method |
| CN112287886A (en) * | 2020-11-19 | 2021-01-29 | 安徽农业大学 | Wheat plant nitrogen content estimation method based on hyperspectral image fusion map features |
| CN112461773A (en) * | 2020-11-18 | 2021-03-09 | 西南大学 | Method and system for determining nitrogen content of tree leaf in fruit expanding period |
| CN112557393A (en) * | 2020-11-19 | 2021-03-26 | 南京农业大学 | Wheat leaf layer nitrogen content estimation method based on hyperspectral image fusion map features |
| CN112881309A (en) * | 2021-02-06 | 2021-06-01 | 内蒙古农业大学 | Establishment method of potato leaf nitrogen detection model and detection method of potato leaf nitrogen |
| CN112986158A (en) * | 2021-02-23 | 2021-06-18 | 内蒙古农业大学 | Beet nitrogen nutrition detection method and system based on multispectral data of unmanned aerial vehicle |
| CN113008890A (en) * | 2021-03-09 | 2021-06-22 | 塔里木大学 | Cotton leaf nitrogen content monitoring method and system based on hyperspectrum of unmanned aerial vehicle |
| CN113670913A (en) * | 2021-08-18 | 2021-11-19 | 沈阳农业大学 | Construction method for inverting hyperspectral vegetation index by using nitrogen content of rice |
| CN114332589A (en) * | 2021-12-31 | 2022-04-12 | 中国科学院紫金山天文台 | Method for accurately detecting surface water or hydroxyl of atmospheric celestial body |
| CN115266720A (en) * | 2022-08-09 | 2022-11-01 | 黑龙江八一农垦大学 | Multi-scale spectral imaging corn growth period nutrient information detection method |
| CN116026772A (en) * | 2023-03-30 | 2023-04-28 | 黑龙江省农业科学院农业遥感与信息研究所 | Corn leaf nitrogen content prediction method based on hyperspectral remote sensing |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101762463A (en) * | 2009-12-16 | 2010-06-30 | 中国烟草总公司郑州烟草研究院 | Method for measuring chlorophyll content of fresh tobacco leaf of flue-cured tobacco based on canopy multi-spectra |
| CN102313699A (en) * | 2011-05-26 | 2012-01-11 | 北京农业信息技术研究中心 | Estimation method of total nitrogen content in crop canopy leaf |
| CN106033052A (en) * | 2015-03-13 | 2016-10-19 | 中国人民解放军沈阳军区司令部工程科研设计所 | A vegetation information extracting method based on hyperspectral data sensitive waveband |
| CN106226246A (en) * | 2016-07-26 | 2016-12-14 | 新疆农垦科学院 | A kind of live monitoring method building garden jujube tree plant nitrogen content |
| CN106290197A (en) * | 2016-09-06 | 2017-01-04 | 西北农林科技大学 | The estimation of rice leaf total nitrogen content EO-1 hyperion and estimation models construction method |
| CN108169141A (en) * | 2017-11-28 | 2018-06-15 | 华南农业大学 | A kind of tea tree LAI and nitrogen content evaluation method based on high spectrum image |
-
2019
- 2019-04-16 CN CN201910305370.9A patent/CN110160967A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101762463A (en) * | 2009-12-16 | 2010-06-30 | 中国烟草总公司郑州烟草研究院 | Method for measuring chlorophyll content of fresh tobacco leaf of flue-cured tobacco based on canopy multi-spectra |
| CN102313699A (en) * | 2011-05-26 | 2012-01-11 | 北京农业信息技术研究中心 | Estimation method of total nitrogen content in crop canopy leaf |
| CN106033052A (en) * | 2015-03-13 | 2016-10-19 | 中国人民解放军沈阳军区司令部工程科研设计所 | A vegetation information extracting method based on hyperspectral data sensitive waveband |
| CN106226246A (en) * | 2016-07-26 | 2016-12-14 | 新疆农垦科学院 | A kind of live monitoring method building garden jujube tree plant nitrogen content |
| CN106290197A (en) * | 2016-09-06 | 2017-01-04 | 西北农林科技大学 | The estimation of rice leaf total nitrogen content EO-1 hyperion and estimation models construction method |
| CN108169141A (en) * | 2017-11-28 | 2018-06-15 | 华南农业大学 | A kind of tea tree LAI and nitrogen content evaluation method based on high spectrum image |
Non-Patent Citations (2)
| Title |
|---|
| 武旭梅: "宁夏引黄灌区水稻生理生化参数高光谱估算研究", 《中国优秀硕士学位论文全文数据库 农业科技辑》 * |
| 白丽敏等: "结合SPA 和PLS 法提高冬小麦冠层全氮高光谱估算的精确度", 《植物营养与肥料学报》 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110567385A (en) * | 2019-09-19 | 2019-12-13 | 廊坊师范学院 | Hyperspectral technology-based construction thickness detection method for building reflective insulation coating |
| CN110987830A (en) * | 2019-12-23 | 2020-04-10 | 陕西地建土地工程技术研究院有限责任公司 | Model, method and application for rapidly determining chlorophyll content of plant canopy leaves |
| CN111650129A (en) * | 2020-03-05 | 2020-09-11 | 广州地理研究所 | Litchi leaf calcium content prediction method and prediction device |
| CN112129709A (en) * | 2020-09-16 | 2020-12-25 | 西北农林科技大学 | Apple tree canopy scale nitrogen content diagnosis method |
| CN112461773B (en) * | 2020-11-18 | 2021-07-06 | 西南大学 | Method and system for determining nitrogen content of tree leaf in fruit expanding period |
| CN112461773A (en) * | 2020-11-18 | 2021-03-09 | 西南大学 | Method and system for determining nitrogen content of tree leaf in fruit expanding period |
| CN112287886B (en) * | 2020-11-19 | 2023-09-22 | 安徽农业大学 | Wheat plant nitrogen content estimation method based on hyperspectral image fusion map features |
| CN112557393A (en) * | 2020-11-19 | 2021-03-26 | 南京农业大学 | Wheat leaf layer nitrogen content estimation method based on hyperspectral image fusion map features |
| CN112287886A (en) * | 2020-11-19 | 2021-01-29 | 安徽农业大学 | Wheat plant nitrogen content estimation method based on hyperspectral image fusion map features |
| CN112557393B (en) * | 2020-11-19 | 2024-02-20 | 南京农业大学 | Wheat leaf layer nitrogen content estimation method based on hyperspectral image fusion map features |
| CN112881309A (en) * | 2021-02-06 | 2021-06-01 | 内蒙古农业大学 | Establishment method of potato leaf nitrogen detection model and detection method of potato leaf nitrogen |
| CN112986158A (en) * | 2021-02-23 | 2021-06-18 | 内蒙古农业大学 | Beet nitrogen nutrition detection method and system based on multispectral data of unmanned aerial vehicle |
| CN113008890A (en) * | 2021-03-09 | 2021-06-22 | 塔里木大学 | Cotton leaf nitrogen content monitoring method and system based on hyperspectrum of unmanned aerial vehicle |
| CN113008890B (en) * | 2021-03-09 | 2023-11-03 | 塔里木大学 | Unmanned aerial vehicle hyperspectral-based cotton leaf nitrogen content monitoring method and system |
| CN113670913A (en) * | 2021-08-18 | 2021-11-19 | 沈阳农业大学 | Construction method for inverting hyperspectral vegetation index by using nitrogen content of rice |
| CN113670913B (en) * | 2021-08-18 | 2023-05-16 | 沈阳农业大学 | Construction method for inversion hyperspectral vegetation index of nitrogen content of rice |
| CN114332589A (en) * | 2021-12-31 | 2022-04-12 | 中国科学院紫金山天文台 | Method for accurately detecting surface water or hydroxyl of atmospheric celestial body |
| CN115266720A (en) * | 2022-08-09 | 2022-11-01 | 黑龙江八一农垦大学 | Multi-scale spectral imaging corn growth period nutrient information detection method |
| CN116026772B (en) * | 2023-03-30 | 2023-06-02 | 黑龙江省农业科学院农业遥感与信息研究所 | Corn leaf nitrogen content prediction method based on hyperspectral remote sensing |
| CN116026772A (en) * | 2023-03-30 | 2023-04-28 | 黑龙江省农业科学院农业遥感与信息研究所 | Corn leaf nitrogen content prediction method based on hyperspectral remote sensing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110160967A (en) | A kind of total nitrogen content evaluation method of crop canopies blade | |
| Song et al. | Wavelength selection and spectral discrimination for paddy rice, with laboratory measurements of hyperspectral leaf reflectance | |
| CN107796764B (en) | Method for constructing wheat leaf area index estimation model based on three-band vegetation index | |
| Abdel-Rahman et al. | Estimation of sugarcane leaf nitrogen concentration using in situ spectroscopy | |
| CN101762463B (en) | Method for measuring chlorophyll content of fresh tobacco leaf of flue-cured tobacco based on canopy multi-spectra | |
| CN111855589A (en) | Remote sensing inversion model and method for rice leaf nitrogen accumulation | |
| CN106442338A (en) | Hyperspectral inversion method for content of chlorophyll in apple leaves based on SVR (support vector regression) algorithm | |
| CN111855590A (en) | Remote sensing inversion model and method for rice leaf starch accumulation | |
| CN110082309B (en) | Method for establishing SPAD value comprehensive spectrum monitoring model of winter wheat canopy | |
| Li et al. | Comparison of four chemometric techniques for estimating leaf nitrogen concentrations in winter wheat (Triticum Aestivum) based on hyperspectral features | |
| CN104778349B (en) | One kind is used for rice table soil nitrogen application Classified Protection | |
| CN109409350A (en) | A kind of Wavelength selecting method based on PCA modeling reaction type load weighting | |
| CN111044516A (en) | Remote sensing estimation method for chlorophyll content of rice | |
| CN112270131A (en) | Remote sensing inversion model and method for rice leaf area index based on ARD regression algorithm | |
| CN112129709A (en) | Apple tree canopy scale nitrogen content diagnosis method | |
| Min et al. | Design of a hyperspectral nitrogen sensing system for orange leaves | |
| CN113065230B (en) | High-spectrum inversion model for establishing rice leaf SPAD based on optimized spectral index | |
| Xie et al. | Spectral reflectance response to nitrogen fertilization in field grown corn | |
| CN102207453A (en) | Method for determining aboveground fresh biomass of flue-cured tobacco based on canopy multi-spectra | |
| CN113655027A (en) | Method for rapidly detecting tannin content in plant by near infrared | |
| CN112362812A (en) | Remote sensing inversion model and method for rice leaf chlorophyll carotenoid content ratio based on Lars algorithm | |
| Keskin et al. | Reflectance-based sensor to predict visual quality ratings of turfgrass plots | |
| CN106404699A (en) | Nondestructive measurement method for contents of nitrogen elements of pear tree leaf blades | |
| Wang et al. | Monitoring model for predicting maize grain moisture at the filling stage using NIRS and a small sample size | |
| Li et al. | Chlorophyll assessment and sensitive wavelength exploration for tea (Camellia sinensis) based on reflectance spectral characteristics |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190823 |
|
| RJ01 | Rejection of invention patent application after publication |