A hyperspectral estimation method for nitrogen content in rice and wheat leaves based on nitrogen distribution theory

Abstract

The present invention proposes a hyperspectral estimation method for nitrogen content in rice and wheat leaves based on nitrogen distribution theory, comprising the following steps: constructing a leaf nitrogen distribution model, calibrating the model coefficients using a rice and wheat leaf scale data set, estimating the leaf chlorophyll content and dry matter content based on the canopy reflectance spectrum, and estimating the leaf nitrogen content using the calibrated leaf nitrogen distribution model. The method of the present invention constructs a leaf nitrogen distribution model at the leaf scale, and then estimates the leaf nitrogen content using biochemical parameters that are more easily estimated using remote sensing technology. The method has simple operating steps, and the model has cross-scale scalability. It is suitable for different rice and wheat varieties, different cultivation treatments, and different growth periods at the canopy level, and can be widely used for monitoring leaf nitrogen content at the rice and wheat canopy level.

Description

Technical Field

[0001] The invention belongs to the field of crop growth monitoring, and in particular to a method for estimating the nitrogen content of rice and wheat leaves by using a high-spectrum method based on nitrogen distribution theory. Background Art

[0002] Rice and wheat play an important role in my country's grain production. Timely monitoring of rice and wheat growth plays a fundamental role in ensuring national food security and sustainable agricultural development. Leaf nitrogen content is an important indicator of crop growth, which can indicate the photosynthesis capacity of plants and diagnose the nutritional status of crops. At present, hyperspectral remote sensing technology has been widely used to accurately and non-destructively estimate the nitrogen content of crop leaves.

[0003] In the past two decades, a large number of studies have used vegetation indices based on the visible-near infrared spectral range to estimate leaf nitrogen content. Commonly used vegetation indices include the ratio vegetation index (RVI) and the normalized difference vegetation index (NDVI). However, the study of the spectral response mechanism in the visible light range is relatively mature, and the main absorbing substances are pigment molecules such as chlorophyll. Therefore, the estimation of leaf nitrogen content based on the visible-near infrared spectral range is mainly attributed to the high correlation between chlorophyll and nitrogen content.

[0004] Total nitrogen content is commonly used in the agricultural field, while chlorophyll can only reflect photosynthetic nitrogen information related to photosynthesis, and cannot fully reflect total nitrogen information. Therefore, the relationship model between chlorophyll-sensitive vegetation index and leaf total nitrogen content may be affected by multiple factors such as crop type and growth period, which limits the application of such models in different species and growth periods. Therefore, a stable and universal method is needed to improve the estimation accuracy of leaf nitrogen content. Summary of the invention

[0005] The technical problem solved by the present invention is to provide a hyperspectral estimation method for nitrogen content in rice and wheat leaves based on nitrogen distribution theory, construct a leaf nitrogen distribution model at the leaf scale, then estimate the biochemical parameters that are relatively easy to obtain at the canopy scale, and finally estimate the leaf nitrogen content. The method has simple operation steps and high accuracy, is suitable for different rice and wheat varieties, different cultivation treatments, and different growth periods at the canopy level, and can be widely used for leaf nitrogen content monitoring at the rice and wheat canopy level.

[0006] The technical solution to achieve the purpose of the present invention is:

[0007] A hyperspectral estimation method for nitrogen content in rice and wheat leaves based on nitrogen distribution theory comprises the following steps:

[0008] Step 1: Construct leaf nitrogen distribution model and calibrate model coefficients;

[0009] Step 2: Estimate leaf chlorophyll content (LCC) and dry matter content (LMA) based on canopy reflectance spectra, and estimate leaf nitrogen content (LNC) using a calibrated leaf nitrogen allocation model M .

[0010] Preferably, the step of constructing and calibrating the blade nitrogen distribution model in step 1 includes:

[0011] Step 1-1: Split the total nitrogen content of the leaves into photosynthetic nitrogen and non-photosynthetic nitrogen, and characterize them by leaf chlorophyll content and leaf dry matter content, respectively. The specific formula is:

[0012] LNC A =PN+NPN=k PN ×LCC+k NPN ×LMA

[0013]

[0014] Among them, LNC A and LNC M are leaf nitrogen contents based on area and mass, PN and NPN are photosynthetic nitrogen and non-photosynthetic nitrogen, LCC and LMA are leaf chlorophyll content and dry matter content, respectively. PN and k NPN is the conversion factor;

[0015] Step 1-2: Determine the conversion factor k PN and k NPN The leaf nitrogen content

[0016] Preferably, in step 1-2: the conversion coefficient k PN and k NPN The value is determined by:

[0017] For the formula:

[0018]

[0019] The known leaf nitrogen content LNC M , chlorophyll content LCC, dry matter content LMA data, enter the formula to establish LNC M and The scatter plot of the regression model is the conversion coefficient k. PN , the intercept is the conversion coefficient k NPN .

[0020] Preferably, in step 1-2: the known leaf nitrogen content LNC M, chlorophyll content LCC, and dry matter content LMA data were obtained from the rice and wheat leaf scale dataset.

[0021] Preferably, the specific formula for estimating the leaf dry matter content LMA in step 2 is:

[0022] LMA=14.81×DMI+46.6

[0023]

[0024] Among them, R 2140-2180 and R 2235-2270 They are the average values ​​of canopy reflectance between 2140-2180nm and 2235-2270nm respectively, and DMI is the vegetation index.

[0025] Preferably, the specific formula for estimating the leaf chlorophyll content LCC in step 2 is:

[0026] LCC=0.83×LCC top -1.1

[0027] Among them, LCC top Represents the leaf chlorophyll content of leaves at the top of the canopy.

[0028] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0029] 1. The hyperspectral estimation method for nitrogen content in rice and wheat leaves based on nitrogen distribution theory of the present invention has simple operation steps, fast calculation speed and good applicability in different scenarios;

[0030] 2. The hyperspectral estimation method of nitrogen content in rice and wheat leaves based on nitrogen distribution theory of the present invention is applicable to different rice and wheat varieties, different cultivation treatments and different growth periods, and can be widely used for leaf nitrogen content estimation;

[0031] 3. The hyperspectral estimation method of nitrogen content in rice and wheat leaves based on nitrogen distribution theory of the present invention is beneficial to improving the accuracy of leaf nitrogen content estimation based on spectroscopy in rice and wheat production. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of nitrogen distribution model coefficient calibration of the rice and wheat leaf nitrogen content hyperspectral estimation method based on nitrogen distribution theory of the present invention.

[0033] Figure 2 This is the estimation result of leaf nitrogen content proposed based on the method of the present invention.

[0034] Figure 3 is the estimated result of leaf nitrogen content based on traditional methods, (A) is VI 401,434,496 The result is (B) VI423,703,924 result. DETAILED DESCRIPTION

[0035] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0036] The embodiment of the present invention is based on rice and wheat experimental fields treated with different growth stages, different nitrogen application levels and different planting densities. The basic information and data collection time of the specific rice and wheat experimental fields are shown in Tables 1 to 3.

[0037] Table 1 Basic information of rice and wheat experimental fields

[0038]

[0039] Table 2 Growth period information when collecting rice canopy spectral data

[0040] date Days after transplanting Reproductive period July 30, 2015 50 Jointing stage August 14, 2015 55 Booting stage August 26, 2015 67 Heading period August 6, 2016 56 Jointing stage August 14, 2016 64 Booting stage August 28, 2016 78 Heading period

[0041] Table 3 Growth period information when wheat canopy spectral data were collected

[0042] date Days after sowing Reproductive period March 18, 2016 138 Jointing stage April 9, 2016 160 Booting stage April 22, 2016 173 Heading period March 15, 2017 135 Jointing stage March 27, 2017 147 Booting stage April 12, 2017 163 Heading period

[0043] The nitrogen content, leaf dry matter, leaf chlorophyll content and canopy reflectance spectrum data of rice and wheat were obtained at the jointing stage, booting stage and heading stage. The experimental data acquisition of the present invention is highly systematic, covering rice and wheat varieties of two plant types, including the main growth period, and multiple treatment factors (density and nitrogen application level), which can effectively verify the leaf nitrogen content estimation method based on nitrogen distribution theory of the present invention, and the accuracy and adaptability of monitoring the nitrogen content of rice and wheat leaves under different environmental conditions and treatments.

[0044] A hyperspectral estimation method for nitrogen content in rice and wheat leaves based on nitrogen distribution theory specifically comprises the following steps:

[0045] Step 1: Obtain leaf nitrogen content (based on mass), leaf chlorophyll content, and leaf dry matter content data for rice and wheat during their main growth periods, and establish LNC based on the following formula M and Scatter plot (such as Figure 1 ), where the slope and intercept of the regression model are the conversion coefficient k PN (2.31) and k NPN (0.71). Once the coefficients are determined, this model is also an estimation model for leaf nitrogen content:

[0046]

[0047] Among them, LNC M is the leaf nitrogen content based on mass, and LCC and LMA are the leaf chlorophyll content and dry matter content, respectively.

[0048] Step 2: Obtain canopy reflectance spectra and corresponding leaf nitrogen content data during the main growth periods of rice and wheat. The collected reflectance spectrum data include different growth periods, different nitrogen application levels, different planting density treatments and different variety types. The measured growth periods include jointing stage, booting stage and heading stage. The canopy reflectance spectrum was measured using the FieldSpec Pro FR2500 back-mounted field hyperspectral radiometer produced by the American Analytical Spectral Device (ASD), using a band range of 400 to 2500nm. The leaf nitrogen content was tested using the Kjeldahl nitrogen method using a SEALAutoAnalyzer 3HR flow analyzer (SEAL Analytical, Ltd., Germany) to measure LNC M .

[0049] Step 2: Estimate the LCC of leaves at the top of the canopy using the leaf area index insensitive chlorophyll index (LICI) top , the specific formula is:

[0050] LCC top =52.85×LICI-16.71

[0051]

[0052] Among them, R 573 ,R 680 ,R 720 and R 735 The canopy reflectance at 573nm, 680nm, 720nm and 735nm respectively. M represents the average leaf nitrogen content of the canopy, so LCC is further top Converted to canopy average LCC, the specific formula is:

[0053] LCC=0.83×LCC top -1.1

[0054] Step 2: Use the vegetation index DMI to estimate the leaf dry matter content. The specific formula is:

[0055] LMA=14.81×DMI+46.6

[0056]

[0057] Among them, R 2140-2180 and R 2235-2270 They are the average values ​​of canopy reflectance between 2140-2180nm and 2235-2270nm respectively.

[0058] Step 2: Substitute the estimated LCC and LMA into the calibrated model to estimate the leaf nitrogen content LNC M .

[0059] The performance of the proposed method in leaf nitrogen content estimation is evaluated below. The estimation accuracy is measured using the coefficient of determination (R 2 ) description, the calculation formula is as follows:

[0060]

[0061] Among them, y i and y' i are the measured and estimated LNC M , For the measured LNC M The estimated results are consistent with the traditional vegetation index VI 401,434,496 and VI 423,703,924 For comparison, the vegetation index calculation formula is as follows:

[0062]

[0063]

[0064] The calibration coefficients of the leaf nitrogen distribution model based on leaf scale data are as follows Figure 1 As shown, the model calibration accuracy R 2 The LNC estimated by the nitrogen distribution model of the present invention is 0.82. M Compared with the measured LNC M The relationship scatter plot is as follows Figure 2 As shown, it can be seen that the estimation accuracy R 2 is 0.63. However, the LNC estimated by the traditional method M Compared with the measured LNC M The relationship scatter plot is as follows Figure 3 As shown, the estimated accuracy R 2 The results show that the method of the present invention can improve LNC M Estimation accuracy.

[0065] The above descriptions are only some embodiments of the present invention. It should be pointed out that a person skilled in the art can make several improvements without departing from the principles of the present invention. These improvements should be regarded as within the protection scope of the present invention.

Claims

1. A method for estimating nitrogen content in rice and wheat leaves using a hyperspectral method based on nitrogen distribution theory, characterized in that: The following steps are involved: Step 1: construct a leaf nitrogen distribution model and calibrate the model coefficients; the steps of constructing and calibrating the leaf nitrogen distribution model in step 1 include: Step 1-1: Split the total nitrogen content of the leaves into photosynthetic nitrogen and non-photosynthetic nitrogen, and characterize them by leaf chlorophyll content and leaf dry matter content, respectively. The specific formula is: LNCA=PN+NPN=k PN ×LCC+k NPN ×LMA Among them, LNC A and LNC M are leaf nitrogen contents based on area and mass, PN and NPN are photosynthetic nitrogen and non-photosynthetic nitrogen, LCC and LMA are leaf chlorophyll content and dry matter content, respectively. PN and k NPN is the conversion factor; Step 1-2: Determine the conversion factor k PN and k NPN The leaf nitrogen content Step 2: Estimate leaf chlorophyll content (LCC) and dry matter content (LMA) based on canopy reflectance spectra, and estimate leaf nitrogen content (LNC) using a calibrated leaf nitrogen allocation model M .

2. The method for estimating nitrogen content in rice and wheat leaves by using a hyperspectral method based on nitrogen distribution theory according to claim 1, characterized in that: In step 1-2: conversion factor k PN and k NPN The value is determined by: For the formula: The known leaf nitrogen content LNC M , chlorophyll content LCC, dry matter content LMA data, enter the formula to establish LNC M and The scatter plot of the regression model is the conversion coefficient k. PN , the intercept is the conversion coefficient k NPN .

3. The method for estimating nitrogen content in rice and wheat leaves by using a hyperspectral method based on nitrogen distribution theory according to claim 1, characterized in that: In step 1-2: Known leaf nitrogen content LNC M , chlorophyll content LCC, and dry matter content LMA data were obtained from the rice and wheat leaf scale dataset.

4. The method for estimating nitrogen content in rice and wheat leaves by using a hyperspectral method based on nitrogen distribution theory according to claim 1, characterized in that: The specific formula for estimating leaf dry matter content LMA in step 2 is: LMA=14.81×DMI+46.6 Among them, R 2140-2180 and R 2235-2270 They are the average values ​​of canopy reflectance between 2140-2180nm and 2235-2270nm respectively, and DMI is the vegetation index.

5. The method for estimating nitrogen content in rice and wheat leaves by hyperspectral method based on nitrogen distribution theory according to claim 1, characterized in that: The specific formula for estimating leaf chlorophyll content LCC in step 2 is: LCC=0.83×LCC top -1.1 Among them, LCC top Represents the leaf chlorophyll content of leaves at the top of the canopy.

Images