小麥、玉米葉片和植株氮營養(yǎng)高光譜診斷與應(yīng)用研究
發(fā)布時間：2018-08-03 18:48
【摘要】：氮素是作物生長發(fā)育所必須的營養(yǎng)元素,是保證作物長勢及產(chǎn)量的基本元素。但是氮肥的大量使用給環(huán)境帶來沉重負擔。因此,準確探測作物氮素營養(yǎng)狀況,在作物關(guān)鍵生育期給作物補充適量的養(yǎng)分以保證作物的生長是必要的。研究表明氮營養(yǎng)指數(shù)(nitrogen nutrition index, NNI)可用于準確探測作物氮素營養(yǎng)狀況。NNI的計算需要氮濃度和生物量這兩個生化參數(shù),常規(guī)測定方法費時費力,難以指導(dǎo)精準農(nóng)業(yè)生產(chǎn)。因此迫切需要用遙感技術(shù)來準確估算氮濃度和生物量,即實現(xiàn)NNI的實時遙感估算。本論文致力于圍繞葉片和植株的高光譜遙感數(shù)據(jù)準確估算氮濃度和生物量,建立關(guān)鍵氮濃度曲線以實現(xiàn)NNI的遙感估測,為及時監(jiān)測作物氮素營養(yǎng)狀況、指導(dǎo)變量施肥和產(chǎn)量預(yù)報提供有效手段。為此,本文基于北京小湯山國家精準農(nóng)業(yè)示范基地和北京農(nóng)林科學(xué)院小麥試驗基地田間試驗數(shù)據(jù),結(jié)合2014/2015年挑旗期和開花期2景無人機高光譜影像數(shù)據(jù),以建立葉片及植株的高光譜氮營養(yǎng)指數(shù)診斷模型為主要研究目標,把氮營養(yǎng)指數(shù)診斷模型應(yīng)用到無人機高光譜影像上,為大范圍氮素營養(yǎng)實時監(jiān)測和精確診斷、變量施肥和產(chǎn)量預(yù)報提供了技術(shù)支撐。論文主要工作如下：(1)從原始光譜反射特征、紅邊參數(shù)、連續(xù)統(tǒng)去除光譜特征、光譜指數(shù)及輻射傳輸模型等角度詳細介紹了作物氮素營養(yǎng)診斷研究進展、氮濃度稀釋模型及氮營養(yǎng)指數(shù)診斷作物氮素狀況的研究進展及不足；并提出了本文的研究思路。(2)基于北京小湯山國家精準農(nóng)業(yè)示范基地和北京農(nóng)林科學(xué)院試驗基地開展的多年小麥、玉米試驗,詳細介紹相關(guān)試驗設(shè)計方案、田間試驗數(shù)據(jù)測定方法、無人機數(shù)碼影像和高光譜影像獲取及處理方法。(3)通過回歸建模分析了小麥、玉米葉片原始光譜反射特征、紅邊參數(shù)、連續(xù)統(tǒng)光譜吸收特征、EFAST(extended fourier amplitude sensitivity, EFAST)方法和PROSPECT模型整合構(gòu)建的對氮素敏感的歸一化光譜指數(shù)(normalized difference spectral index, NDSI)和比值植被指數(shù)(ratio spectral index, RSI)光譜指數(shù)、氮素常用植被指數(shù)與作物葉片氮素營養(yǎng)狀況的關(guān)系,比較了不同植被指數(shù)估算作物葉片氮素營養(yǎng)狀況的精度,確定了對小麥、玉米氮素營養(yǎng)狀況敏感的光譜指數(shù)的順序。①利用PROSPECT模型隨機模擬葉片光譜反射率數(shù)據(jù),采用EFAST方法對PROSPECT模型中各個生理生化參數(shù)在400-2500nm波段范圍的葉片反射光譜進行敏感性分析。結(jié)果表明對葉綠素敏感的波段范圍是417-728nm,參照歸一化植被指數(shù)和比值植被指數(shù),本研究構(gòu)建了對小麥氮素敏感的歸一化光譜指數(shù)NDS 1(564,728)、 NDSI(543,728)、RSI(564,728)和RSI(543,728)、對玉米氮素敏感的光譜指數(shù)是NDSI(629,649)、NDSI(495,669、RSI(629,649)和RSI(495,669)分析了光譜指數(shù)NDSI和RSI與作物葉片氮含量及葉片氮累積量的相關(guān)性,研究表明除了玉米灌漿期外,其他生育期的光譜指數(shù)與葉片氮含量的相關(guān)性高于與葉片氮累積量的相關(guān)性,葉片氮含量比葉片氮累積量對葉片光譜參數(shù)更為敏感。②建立了光譜指數(shù)與葉片氮含量、葉片氮累積量的回歸模型,用決定系數(shù)(coefficient of determination,R2)、均方根誤差(root mean square error, RMSE)相對誤差(relative error, RE)作為評價確定最佳葉片氮素營養(yǎng)狀況光譜指數(shù)的指標。結(jié)果表明對小麥葉片氮含量較為敏感的前5個光譜指數(shù)是mmND705、ND705、SR705、GMI-2、RI-half;對小麥葉片氮累積量較為敏感的前5個光譜指數(shù)是R550/R800、GMI-1、RSI(564,728)、RSI(543,728)、RI-2dB;對玉米葉片氮含量較為敏感的前5個光譜指數(shù)是VOGb、VOGc、NDRE、VOGa、CIred edge;對玉米葉片氮累積量較為敏感的前5個光譜指數(shù)是NDRE、MTCI、RI-2dB、VOGa、VOGb。(4)通過回歸建模分析了植株在不同年份、不同生育期原始光譜反射特征、紅邊參數(shù)、連續(xù)統(tǒng)光譜吸收特征、EFAST方法和PROSPAIL模型整合構(gòu)建的對氮素敏感的光譜指數(shù)NDSI和RSI、氮素常用光譜指數(shù)與氮素營養(yǎng)狀況的關(guān)系,研究確定了對小麥植株敏感的光譜指數(shù)的順序。①利用PROSAIL模型隨機模擬植株光譜反射率數(shù)據(jù),采用EFAST方法對PROSAIL模型各個參數(shù)在400-2500nm波段范圍的植株反射光譜進行敏感性分析。結(jié)果表明對葉綠素敏感的波段范圍是515-745nm,參照歸一化植被指數(shù)和比值植被指數(shù),構(gòu)建了對小麥植株氮含量和植株氮累積量敏感的光譜指數(shù)是NDSI(546,698)、 NDSI(667,685)、NDSI(539,745)、RSI(546,694)、RSI(667,684)、RSI(539,745)、分析了光譜指數(shù)NDSI和RSI與植株氮含量及植株氮累積量的相關(guān)性。②基于經(jīng)驗統(tǒng)計關(guān)系建立了光譜指數(shù)與植株氮含量、植株氮累積量的回歸模型,用尺2、RMSE、RE作為評價植株氮素營養(yǎng)狀況的指標。結(jié)果表明對小麥植株氮含量敏感的前5個光譜指數(shù)是SRPI、NPCI、ND705、MCARI/MTVI2和MTCI,對小麥植株氮累積量敏感的前5個光譜指數(shù)是SR705、RI-half、NPCI、VOGb和mSR705。(5)通過回歸建模分析了葉片及植株生物量原始光譜反射特征、紅邊參數(shù)、連續(xù)統(tǒng)光譜吸收特征、EFAST方法和PROSPECT (PROSAIL)模型整合構(gòu)建的對生物量敏感的光譜指數(shù)、與生物量相關(guān)的常用光譜指數(shù)與生物量的關(guān)系,研究確定了對葉片及植株生物量敏感的光譜指數(shù)的順序。①利用PROSPECT (PROSAIL)模型隨機模擬葉片及植株光譜反射率數(shù)據(jù),采用EFAST方法對PROSPECT (PROSAIL)模型各個參數(shù)在400-2500nm波段范圍的植株反射光譜進行敏感性分析。結(jié)果表明對葉片及植株生物量敏感的波段范圍為749-2410nm,構(gòu)建了對小麥葉片及植株生物量敏感的光譜指數(shù)是NDSI(2126,2347、NDSI(1652,1686)、RSI(2126,2347)和RSI(1652,1686),分析了光譜指數(shù)NDSI和RSI與葉片及植株生物量的相關(guān)性。②基于經(jīng)驗統(tǒng)計關(guān)系建立了光譜指數(shù)與葉片及植株生物量的回歸模型,用R2、RMSE、RE作為評價葉片及植株生物量的指標。結(jié)果表明對小麥葉片生物量敏感的前5個光譜指數(shù)是mSR705、RI-1dB、VOGa、GNDVI和NDCI,對小麥植株生物量敏感的前5個光譜指數(shù)是VOGa、mSR705、REP、NDVI705和mNDVI705。(6)建立了研究區(qū)小麥葉片及植株關(guān)鍵氮稀釋曲線模型；建立了“遙感信息—農(nóng)學(xué)參數(shù)—氮營養(yǎng)指數(shù)”葉片及植株氮營養(yǎng)診斷模型。①建立的研究區(qū)小麥葉片關(guān)鍵氮濃度曲線模型為Ncl=4.42×W-0.18L,建立的研究區(qū)小麥植株關(guān)鍵氮濃度曲線模型為N印=5.81×W-0.54。②基于葉片“遙感信息—農(nóng)學(xué)參數(shù)—氮營養(yǎng)指數(shù)”估測的NNI與實際NNI之間的R2為0.77,基于植株“遙感信息—農(nóng)學(xué)參數(shù)—氮營養(yǎng)指數(shù)”估測的NNI與實測NNI之間的R2為0.83,研究表明建立的植株NNI診斷模型精度高于葉片NNI診斷模型精度。(7)采用小湯山國家精準農(nóng)業(yè)示范基地?zé)o人機高光譜影像提取作物NNI,挑旗期提取的NNI與實際NNI間的R2為0.66；開花期提取的NNI與實際NNI間的R2為0.69,均達到了顯著相關(guān)。結(jié)果表明,基于“遙感信息—農(nóng)學(xué)參數(shù)—氮營養(yǎng)指數(shù)”方法用于NNI估算是可行的,能夠得到與實際吻合的結(jié)果,為快速、準確實時監(jiān)測作物氮素狀況、變量施肥和產(chǎn)量預(yù)報提供了科學(xué)依據(jù)。
[Abstract]:Nitrogen is a necessary nutrient element for crop growth and development. It is the basic element to ensure crop growth and yield. However, the heavy use of nitrogen fertilizer brings a heavy burden to the environment. Therefore, it is necessary to accurately detect the nitrogen nutrition status of crops and to supplement the appropriate amount of nutrients to crops to ensure the growth of crops in the critical growth period of crops. The nitrogen nutrition index (NNI) can be used to accurately detect the nitrogen nutrition status of the crop,.NNI, which requires two biochemical parameters of nitrogen and biomass. The conventional method is time-consuming and difficult to guide the production of precision agriculture. Therefore, it is urgent to use remote sensing technology to accurately estimate nitrogen concentration and biomass, that is, to achieve NN. I's real-time remote sensing estimation. This paper is devoted to accurately estimating nitrogen concentration and biomass around the hyperspectral remote sensing data of leaves and plants, establishing a key nitrogen concentration curve to achieve remote sensing estimation of NNI, providing effective means for timely monitoring of nitrogen nutrition status of crops, guiding variable fertilization and yield forecasting. This paper is based on Beijing soup. The field test data of Mountain National precision agriculture demonstration base and Beijing Academy of agricultural and Forestry Sciences, combined with the high spectral image data of 2 unmanned aerial vehicles (UAV) in the flag period and the flowering period of 2014/2015, the high light spectrum nitrogen nutrition index diagnosis model of leaves and plants was established as the main target, and the nitrogen nutrition index diagnosis model was applied to no one. The main work of this paper is as follows: (1) the nitrogen nutrition of crop is introduced in detail from the original spectral reflectance characteristics, red edge parameters, continuous spectral characteristics, spectral index and radiative transfer model. The research progress, the nitrogen concentration dilution model and the nitrogen nutrition index in the diagnosis of nitrogen status of the crop, and put forward the research ideas in this paper. (2) many years of wheat and corn test based on the Xiaotangshan national precision agriculture demonstration base and the experimental base of Beijing Academy of agricultural and Forestry Sciences in Beijing, the related experiment design was introduced in detail. Method, field test data determination method, UAV digital image and hyperspectral image acquisition and processing methods. (3) the original spectral reflectance characteristics of wheat and corn leaves, red edge parameters, continuous spectral absorption characteristics, EFAST (Extended Fourier amplitude sensitivity, EFAST) method and PROSPECT model are analyzed by regression modeling. The relationship between nitrogen sensitive spectral index (normalized difference spectral index, NDSI) and ratio vegetation index (ratio spectral index, RSI), nitrogen common vegetation index and nitrogen nutrition status of crop leaves was built, and the precision of nitrogen nutrition status of crop leaves was compared with different vegetation indices. The order of spectral indices sensitive to nitrogen nutrition status in wheat and maize. (1) the PROSPECT model was used to simulate the spectral reflectance of leaves, and the sensitivity analysis of the leaf reflectance spectra of the physiological and biochemical parameters in the 400-2500nm band was analyzed by the EFAST method. According to the normalized vegetation index and the ratio vegetation index, this study constructed the normalized spectral index NDS 1 (564728), NDSI (543728), RSI (564728) and RSI (543728) for wheat nitrogen sensitivity. The spectral index of nitrogen sensitivity to maize was NDSI (629649), NDSI (495669, RSI (629649) and RSI (495669) analyzed the spectral finger. The correlation of NDSI and RSI with nitrogen content of crop leaves and nitrogen accumulation of leaves showed that the correlation between the spectral index of other growth periods and leaf nitrogen content was higher than that of leaf nitrogen accumulation except for the grain filling period. The nitrogen content of leaves was more sensitive than leaf nitrogen accumulation to leaf spectral parameters. The number and leaf nitrogen content, the regression model of the nitrogen accumulation of leaves, the coefficient of determination, R2, the relative error (relative error, RMSE) relative error (relative error, RMSE) as the index of determining the optimum leaf nitrogen nutrient status spectral index. The results showed that the nitrogen content of Wheat leaves was more sensitive. The first 5 spectral indices of the sense are mmND705, ND705, SR705, GMI-2, RI-half; the first 5 spectral indices that are more sensitive to nitrogen accumulation in wheat leaves are R550/R800, GMI-1, RSI (564728), RSI (543728), RI-2dB, and the first 5 spectral indices that are more sensitive to the nitrogen content of maize leaves are VOGb. The first 5 spectral indices are NDRE, MTCI, RI-2dB, VOGa, VOGb. (4), through regression modeling, the original spectral reflectance characteristics of plants in different years, different growth stages, red edge parameters, continuous spectral absorption characteristics, and EFAST method and PROSPAIL model are integrated to construct nitrogen sensitive spectral index NDSI and RSI, nitrogen is commonly used for spectral index. The relationship between number and nitrogen nutrition status, the order of spectral index sensitive to wheat plants was determined. (1) PROSAIL model was used to simulate plant spectral reflectance data, and EFAST method was used to analyze the sensitivity analysis of plant reflectance spectra of each parameter of PROSAIL model in the range of 400-2500nm band. The results showed that the chlorophyll sensitivity was sensitive to chlorophyll sensitivity. The band range of sense is 515-745nm. According to the normalized vegetation index and ratio vegetation index, the spectral index of nitrogen content and plant nitrogen accumulation sensitivity to wheat plants is NDSI (546698), NDSI (667685), NDSI (539745), RSI (546694), RSI (667684), RSI (539745)). The spectral index NDSI and RSI and plant nitrogen content and plant are analyzed. The correlation of nitrogen accumulation, based on the empirical statistical relationship, established the spectral index and plant nitrogen content, the regression model of plant nitrogen accumulation, using scale 2, RMSE, RE as indicators to evaluate plant nitrogen nutrition status. The results showed that the first 5 spectral indices sensitive to nitrogen content in wheat plants were SRPI, NPCI, ND705, MCARI/MTVI2 and MTCI, to wheat. The first 5 spectral indices of plant nitrogen accumulation sensitivity are SR705, RI-half, NPCI, VOGb and mSR705. (5). By regression modeling, the original spectral reflectance characteristics of leaf and plant biomass, red edge parameters, continuum spectral absorption characteristics, EFAST method and PROSPECT (PROSAIL) model integrated and constructed by the EFAST method and PROSPECT (PROSAIL) model, are used to analyze the spectral index of biomass sensitive. The sequence of spectral indices sensitive to the biomass of leaves and plants was determined by the correlation of the related spectral index and biomass. (1) the PROSPECT (PROSAIL) model was used to simulate the spectral reflectance of leaves and plants, and the EFAST method was used to determine the plant parameters of the PROSPECT (PROSAIL) model in the range of 400-2500nm band. The spectral sensitivity analysis showed that the range of sensitivity to leaf and plant biomass was 749-2410nm. The spectral index of wheat leaf and plant biomass sensitive was NDSI (21262347, NDSI (16521686), RSI (21262347) and RSI (16521686). The spectral index NDSI and RSI and leaf and plant biomass were analyzed. " The regression model of spectral index and leaf and plant biomass was established based on the empirical statistical relationship. R2, RMSE and RE were used as indicators to evaluate the biomass of leaves and plants. The results showed that the first 5 spectral indices sensitive to wheat leaf biomass were mSR705, RI-1dB, VOGa, GNDVI and NDCI, and the first 5 sensitive to wheat plant biomass. The spectral index is VOGa, mSR705, REP, NDVI705 and mNDVI705. (6) to establish the key nitrogen dilution curve model of wheat leaves and plants in the study area. A model of "remote sensing information - Agronomy parameter nitrogen nutrition index" leaf and plant nitrogen nutrition diagnosis model was established. (1) the key nitrogen concentration curve model of wheat leaves was established in the study area Ncl=4.42 x W-0.18L The key nitrogen concentration curve model of the wheat plant in the study area is N =5.81 x W-0.54. (2) based on the =5.81 of the leaf "remote sensing information - Agronomy parameter - nitrogen nutrition index" and the R2 between the actual NNI and the actual NNI. The R2 between the NNI and the measured NNI based on the plant "remote sensing information - agronomic parameter nitrogen nutrition index" is 0.83. The results showed that the accuracy of the NNI diagnostic model was higher than that of the blade NNI diagnosis model. (7) the crop NNI was extracted from the UAV hyperspectral image of the national precision agricultural demonstration base in Xiaotangshan, the R2 between the NNI and the actual NNI was 0.66, and the R2 between the NNI and the actual NNI during the flowering period was 0.69, and the results reached significant correlation. The results show that it is feasible to estimate NNI based on the "remote sensing information - agronomic parameter nitrogen nutrition index" method. It can be consistent with the actual results. It provides a scientific basis for fast, accurate and real-time monitoring of crop nitrogen status, variable fertilization and yield prediction.
【學(xué)位授予單位】：中國礦業(yè)大學(xué)(北京)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2016
【分類號】：S512.1;S513
，
本文編號：2162676
[Abstract]:Nitrogen is a necessary nutrient element for crop growth and development. It is the basic element to ensure crop growth and yield. However, the heavy use of nitrogen fertilizer brings a heavy burden to the environment. Therefore, it is necessary to accurately detect the nitrogen nutrition status of crops and to supplement the appropriate amount of nutrients to crops to ensure the growth of crops in the critical growth period of crops. The nitrogen nutrition index (NNI) can be used to accurately detect the nitrogen nutrition status of the crop,.NNI, which requires two biochemical parameters of nitrogen and biomass. The conventional method is time-consuming and difficult to guide the production of precision agriculture. Therefore, it is urgent to use remote sensing technology to accurately estimate nitrogen concentration and biomass, that is, to achieve NN. I's real-time remote sensing estimation. This paper is devoted to accurately estimating nitrogen concentration and biomass around the hyperspectral remote sensing data of leaves and plants, establishing a key nitrogen concentration curve to achieve remote sensing estimation of NNI, providing effective means for timely monitoring of nitrogen nutrition status of crops, guiding variable fertilization and yield forecasting. This paper is based on Beijing soup. The field test data of Mountain National precision agriculture demonstration base and Beijing Academy of agricultural and Forestry Sciences, combined with the high spectral image data of 2 unmanned aerial vehicles (UAV) in the flag period and the flowering period of 2014/2015, the high light spectrum nitrogen nutrition index diagnosis model of leaves and plants was established as the main target, and the nitrogen nutrition index diagnosis model was applied to no one. The main work of this paper is as follows: (1) the nitrogen nutrition of crop is introduced in detail from the original spectral reflectance characteristics, red edge parameters, continuous spectral characteristics, spectral index and radiative transfer model. The research progress, the nitrogen concentration dilution model and the nitrogen nutrition index in the diagnosis of nitrogen status of the crop, and put forward the research ideas in this paper. (2) many years of wheat and corn test based on the Xiaotangshan national precision agriculture demonstration base and the experimental base of Beijing Academy of agricultural and Forestry Sciences in Beijing, the related experiment design was introduced in detail. Method, field test data determination method, UAV digital image and hyperspectral image acquisition and processing methods. (3) the original spectral reflectance characteristics of wheat and corn leaves, red edge parameters, continuous spectral absorption characteristics, EFAST (Extended Fourier amplitude sensitivity, EFAST) method and PROSPECT model are analyzed by regression modeling. The relationship between nitrogen sensitive spectral index (normalized difference spectral index, NDSI) and ratio vegetation index (ratio spectral index, RSI), nitrogen common vegetation index and nitrogen nutrition status of crop leaves was built, and the precision of nitrogen nutrition status of crop leaves was compared with different vegetation indices. The order of spectral indices sensitive to nitrogen nutrition status in wheat and maize. (1) the PROSPECT model was used to simulate the spectral reflectance of leaves, and the sensitivity analysis of the leaf reflectance spectra of the physiological and biochemical parameters in the 400-2500nm band was analyzed by the EFAST method. According to the normalized vegetation index and the ratio vegetation index, this study constructed the normalized spectral index NDS 1 (564728), NDSI (543728), RSI (564728) and RSI (543728) for wheat nitrogen sensitivity. The spectral index of nitrogen sensitivity to maize was NDSI (629649), NDSI (495669, RSI (629649) and RSI (495669) analyzed the spectral finger. The correlation of NDSI and RSI with nitrogen content of crop leaves and nitrogen accumulation of leaves showed that the correlation between the spectral index of other growth periods and leaf nitrogen content was higher than that of leaf nitrogen accumulation except for the grain filling period. The nitrogen content of leaves was more sensitive than leaf nitrogen accumulation to leaf spectral parameters. The number and leaf nitrogen content, the regression model of the nitrogen accumulation of leaves, the coefficient of determination, R2, the relative error (relative error, RMSE) relative error (relative error, RMSE) as the index of determining the optimum leaf nitrogen nutrient status spectral index. The results showed that the nitrogen content of Wheat leaves was more sensitive. The first 5 spectral indices of the sense are mmND705, ND705, SR705, GMI-2, RI-half; the first 5 spectral indices that are more sensitive to nitrogen accumulation in wheat leaves are R550/R800, GMI-1, RSI (564728), RSI (543728), RI-2dB, and the first 5 spectral indices that are more sensitive to the nitrogen content of maize leaves are VOGb. The first 5 spectral indices are NDRE, MTCI, RI-2dB, VOGa, VOGb. (4), through regression modeling, the original spectral reflectance characteristics of plants in different years, different growth stages, red edge parameters, continuous spectral absorption characteristics, and EFAST method and PROSPAIL model are integrated to construct nitrogen sensitive spectral index NDSI and RSI, nitrogen is commonly used for spectral index. The relationship between number and nitrogen nutrition status, the order of spectral index sensitive to wheat plants was determined. (1) PROSAIL model was used to simulate plant spectral reflectance data, and EFAST method was used to analyze the sensitivity analysis of plant reflectance spectra of each parameter of PROSAIL model in the range of 400-2500nm band. The results showed that the chlorophyll sensitivity was sensitive to chlorophyll sensitivity. The band range of sense is 515-745nm. According to the normalized vegetation index and ratio vegetation index, the spectral index of nitrogen content and plant nitrogen accumulation sensitivity to wheat plants is NDSI (546698), NDSI (667685), NDSI (539745), RSI (546694), RSI (667684), RSI (539745)). The spectral index NDSI and RSI and plant nitrogen content and plant are analyzed. The correlation of nitrogen accumulation, based on the empirical statistical relationship, established the spectral index and plant nitrogen content, the regression model of plant nitrogen accumulation, using scale 2, RMSE, RE as indicators to evaluate plant nitrogen nutrition status. The results showed that the first 5 spectral indices sensitive to nitrogen content in wheat plants were SRPI, NPCI, ND705, MCARI/MTVI2 and MTCI, to wheat. The first 5 spectral indices of plant nitrogen accumulation sensitivity are SR705, RI-half, NPCI, VOGb and mSR705. (5). By regression modeling, the original spectral reflectance characteristics of leaf and plant biomass, red edge parameters, continuum spectral absorption characteristics, EFAST method and PROSPECT (PROSAIL) model integrated and constructed by the EFAST method and PROSPECT (PROSAIL) model, are used to analyze the spectral index of biomass sensitive. The sequence of spectral indices sensitive to the biomass of leaves and plants was determined by the correlation of the related spectral index and biomass. (1) the PROSPECT (PROSAIL) model was used to simulate the spectral reflectance of leaves and plants, and the EFAST method was used to determine the plant parameters of the PROSPECT (PROSAIL) model in the range of 400-2500nm band. The spectral sensitivity analysis showed that the range of sensitivity to leaf and plant biomass was 749-2410nm. The spectral index of wheat leaf and plant biomass sensitive was NDSI (21262347, NDSI (16521686), RSI (21262347) and RSI (16521686). The spectral index NDSI and RSI and leaf and plant biomass were analyzed. " The regression model of spectral index and leaf and plant biomass was established based on the empirical statistical relationship. R2, RMSE and RE were used as indicators to evaluate the biomass of leaves and plants. The results showed that the first 5 spectral indices sensitive to wheat leaf biomass were mSR705, RI-1dB, VOGa, GNDVI and NDCI, and the first 5 sensitive to wheat plant biomass. The spectral index is VOGa, mSR705, REP, NDVI705 and mNDVI705. (6) to establish the key nitrogen dilution curve model of wheat leaves and plants in the study area. A model of "remote sensing information - Agronomy parameter nitrogen nutrition index" leaf and plant nitrogen nutrition diagnosis model was established. (1) the key nitrogen concentration curve model of wheat leaves was established in the study area Ncl=4.42 x W-0.18L The key nitrogen concentration curve model of the wheat plant in the study area is N =5.81 x W-0.54. (2) based on the =5.81 of the leaf "remote sensing information - Agronomy parameter - nitrogen nutrition index" and the R2 between the actual NNI and the actual NNI. The R2 between the NNI and the measured NNI based on the plant "remote sensing information - agronomic parameter nitrogen nutrition index" is 0.83. The results showed that the accuracy of the NNI diagnostic model was higher than that of the blade NNI diagnosis model. (7) the crop NNI was extracted from the UAV hyperspectral image of the national precision agricultural demonstration base in Xiaotangshan, the R2 between the NNI and the actual NNI was 0.66, and the R2 between the NNI and the actual NNI during the flowering period was 0.69, and the results reached significant correlation. The results show that it is feasible to estimate NNI based on the "remote sensing information - agronomic parameter nitrogen nutrition index" method. It can be consistent with the actual results. It provides a scientific basis for fast, accurate and real-time monitoring of crop nitrogen status, variable fertilization and yield prediction.
【學(xué)位授予單位】：中國礦業(yè)大學(xué)(北京)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2016
【分類號】：S512.1;S513
，
本文編號：2162676
本文鏈接：http://sikaile.net/shoufeilunwen/nykjbs/2162676.html
最近更新
教材專著
