夏玉米不同生育時期生理生態(tài)參數(shù)的高光譜遙感監(jiān)測模型
發(fā)布時間：2018-09-05 11:28
【摘要】：高光譜遙感技術(shù)是現(xiàn)代農(nóng)業(yè)信息技術(shù)研究的重要方向,高效低耗的作物營養(yǎng)診斷及長勢監(jiān)測技術(shù)對精準(zhǔn)作物管理及合理施肥具有重要意義。高光譜遙感技術(shù)具有分辨率高、連續(xù)性強、信息量大等特點,可對作物進行實時、快速、無損的營養(yǎng)診斷、長勢監(jiān)測、產(chǎn)量和品質(zhì)預(yù)測。本試驗在不同夏玉米品種、不同氮磷營養(yǎng)水平及不同生育時期的大田試驗條件下,測定了冠層光譜反射率、葉片氮磷含量、葉面積指數(shù)、地上干物質(zhì)量、葉片水分含量、籽粒品質(zhì)及產(chǎn)量等生理生態(tài)參數(shù),分析不同生育時期及不同氮磷水平對夏玉米冠層光譜反射率的影響、光譜反射率與生理生態(tài)參數(shù)的相關(guān)性,綜合前人已有的研究成果,建立生理生態(tài)參數(shù)與光譜敏感波段、光譜特征參量、植被指數(shù)的相關(guān)模型并進行模型精度的檢驗。最終篩選出適用于不同夏玉米品種及不同生育時期的高光譜診斷與監(jiān)測生理生態(tài)參數(shù)模型,為夏玉米營養(yǎng)診斷及長勢監(jiān)測提供理論基礎(chǔ)及現(xiàn)實依據(jù)。1.在夏玉米拔節(jié)期、大喇叭口期、吐絲期和灌漿期,葉片氮素含量隨施氮量的增加而增加。可見光波段夏玉米冠層光譜反射率隨施氮量的增加而降低,在近紅外波段,光譜反射率隨施氮量的增加而增加。在可見光波段,從拔節(jié)期到吐絲期,光譜反射率逐漸降低,進入灌漿期后,光譜反射率增強;在近紅外區(qū)域,從拔節(jié)期到吐絲期,冠層光譜反射率逐漸增加,進入灌漿期后,光譜反射率降低。夏玉米葉片全氮含量與原始光譜在350~725nm為負相關(guān),在740~1400nm的近紅外波段為顯著正相關(guān),725~735nm為相關(guān)系數(shù)轉(zhuǎn)變的位置。拔節(jié)期(Dy、Ro、SDr)、大喇叭口期(Rg、SDr/SDb、GNDVI)、吐絲期(DI、DVI、NDVI1)和灌漿期(SDr/SDb、GNDVI、NDVI1)的參量與葉片全氮含量具有較高的相關(guān)系數(shù)。拔節(jié)期、大喇叭口期、吐絲期和灌漿期玉米葉片氮素含量的最適反演模型參量分別為SDr、GNDVI、NDVI1和NDVI1。2.夏玉米葉片全磷含量隨著施磷量增加而升高,在不同生育時期均有相似規(guī)律。在可見光波段,夏玉米冠層光譜反射率與不同磷肥水平響應(yīng)不一致。在近紅外波段(740~1380nm),光譜反射率隨施磷量的增加而增加,但光譜響應(yīng)曲線差異不明顯;在此區(qū)域,相關(guān)系數(shù)較為穩(wěn)定,9個歸一化光譜波段(830nm、880 nm、940 nm、1100 nm、1430 nm、1580 nm、1650 nm、1740 nm、2200 nm)為敏感波段,雙波段組合參量具有更高的相關(guān)系數(shù)。拔節(jié)期(R830、R880、R830+880、R830+940)、大喇叭口期(R830、R880、R830+880、R830+940)、灌漿期(R940、R1100、R880+1100、R940+1100)歸一化參量具有較高的相關(guān)系數(shù),灌漿期的參量均未達到顯著相關(guān)。拔節(jié)期、大喇叭口期和灌漿期的R830+880、R830+940和R880+1100歸一化參量的模型為玉米葉片磷素含量反演最適模型。3.在夏玉米大喇叭口期、吐絲期、灌漿期和成熟期,在同一施磷水平下,冠層葉面積指數(shù)隨施氮量的增加而增加;在同一施氮水平下,隨施磷量的增加,葉面積指數(shù)也明顯的增加。冠層葉面積指數(shù)和光譜反射率的相關(guān)性在可見光波段(420~680nm)為負相關(guān),在近紅外波段(740~1120nm)為正相關(guān),各生育時期相關(guān)系數(shù)曲線走勢基本一致,但相關(guān)系數(shù)普遍不高。利用原始冠層反射光譜構(gòu)建21個植被指數(shù),大喇叭口期(PSSRb、NDVI1、N DVI2)、吐絲期(PSSRc、MTCI、MSR705)、灌漿期(MSR705)和成熟期(DI)構(gòu)建的擬合方程具有較高的決定系數(shù)和F值。大喇叭口期、吐絲期、灌漿期和成熟期的植被指數(shù)NDVI2、MSR705、MSR705、DI可較好的應(yīng)用于夏玉米冠層葉面積指數(shù)監(jiān)測。4.從大喇叭口期到成熟期,不同氮磷水平下夏玉米地上干物質(zhì)量均表現(xiàn)為逐漸增加的趨勢,且都符合典型的“S”型生長曲線,生長速率為“慢-快-慢”。在夏玉米的各生育時期,在同一施磷水平下,地上干物質(zhì)量隨施氮量的增加而增加;在同一施氮水平下,隨施磷量的增加,干物質(zhì)積累量也有明顯的升高。在可見光波段,夏玉米冠層原始光譜反射率與地上干物質(zhì)量為較高且穩(wěn)定的負相關(guān);在近紅外波段,夏玉米冠層原始光譜反射率與地上干物質(zhì)量具有較高的正相關(guān)性。大喇叭口期、吐絲期、灌漿期和成熟期的植被指數(shù)GNDVI、PSSRc、NDVI4和DI可較好的用于估算夏玉米地上干物質(zhì)量。5.從拔節(jié)期到灌漿期,葉片水分含量呈逐漸下降趨勢,在吐絲期水分含量差異最大,拔節(jié)期水分含量差異最小。夏玉米的拔節(jié)期、大喇叭口期、吐絲期和灌漿期的葉片光譜反射率與葉片水分含量為負相關(guān),不同生育時期具有較大差異。拔節(jié)期和大喇叭口期,葉片光譜反射率與水分含量在740~1340nm為穩(wěn)定負相關(guān),但相關(guān)系數(shù)不高;大喇叭口期、吐絲期和灌漿期的1370~2500nm波段,除1900nm波段附近,其他波段為穩(wěn)定負相關(guān),且相關(guān)系數(shù)較高。選擇9個敏感波段和18個水分植被指數(shù)與葉片水分含量進行相關(guān)性分析。在大喇叭口期、吐絲期和灌漿期,敏感波段1450nm和1650nm、植被指數(shù)NDWI2和SIWSI的相關(guān)系數(shù)較高。大喇叭口期、吐絲期和灌漿期的R1450、NDWI2和NDWI2建立的回歸擬合模型可用于夏玉米葉片水分含量的監(jiān)測。6.夏玉米籽粒蛋白含量及產(chǎn)量均隨施肥量的增加而升高,且不同氮磷肥力處理間差異多為顯著。選擇灌漿期的14個植被指數(shù)與籽粒蛋白含量和產(chǎn)量進行回歸分析。在籽粒蛋白含量擬合方程中,植被指數(shù)RVI和OSAVI擬合方程具有較高的決定系數(shù);在籽粒產(chǎn)量擬合方程中,植被指數(shù)RVI、GRVI和NDVI具有較高的決定系數(shù)及F值。模型精度精度檢驗結(jié)果表明,植被指數(shù)RVI可較好的預(yù)測夏玉米籽粒蛋白含量和產(chǎn)量,具有較好的穩(wěn)定性和抗背景干擾性。
[Abstract]:Hyperspectral remote sensing technology is an important research direction of modern agricultural information technology. High-efficiency and low-consumption crop nutrition diagnosis and growth monitoring technology is of great significance to accurate crop management and rational fertilization. Nutrition diagnosis, growth monitoring, yield and quality prediction. The physiological and ecological parameters, such as canopy spectral reflectance, leaf nitrogen and phosphorus content, leaf area index, aboveground dry matter, leaf water content, grain quality and yield, were measured under different summer maize varieties, different nitrogen and phosphorus nutrition levels and different growth stages. The effects of different growth stages and nitrogen and phosphorus levels on the spectral reflectance of summer maize canopy and the correlation between spectral reflectance and physiological and ecological parameters were analyzed. Hyperspectral diagnostic and monitoring physiological and ecological parameters models suitable for different summer maize varieties and different growth stages were screened to provide theoretical and practical basis for nutritional diagnosis and growth monitoring of summer maize. The spectral reflectance of summer maize canopy decreased with the increase of nitrogen application in the light band, and increased with the increase of nitrogen application in the near infrared band. The total nitrogen content in summer maize leaves was negatively correlated with the original spectrum from 350 nm to 725 nm, positively correlated with the near infrared spectrum from 740 nm to 1400 nm, and positively correlated with the position of correlation coefficient change from 725 nm to 735 nm. The parameters of I, NDVI1 and grain filling (SDr/SDb, GNDVI, NDVI1) had higher correlation coefficients with leaf total nitrogen content. The optimum parameters of the model were SDr, GNDVI, NDVI1 and NDVI1.2 at jointing stage, big trumpet stage, silking stage and grain filling stage, respectively. In the near infrared band (740 ~ 1380 nm), the spectral reflectance increased with the increase of phosphorus application, but the spectral response curves were not significantly different; in this region, the correlation coefficient was relatively stable, and nine normalized spectral bands (830 n) M, 880 nm, 940 nm, 1100 nm, 1430 nm, 1580 nm, 1650 nm, 1740 nm, 2200 nm are sensitive bands, and the combined parameters of the two bands have higher correlation coefficients. The normalized parameters of jointing stage (R830, R880, R830 + 880, R830 + 940), large trumpet stage (R830, R880, R830 + 880, R830 + 940, R830 + 940), grouting stage (R940, R1100, R880 + 1100, R940) have higher correlation coefficients. The models of R830+880, R830+940 and R880+1100 normalized parameters at jointing stage, big trumpet stage and filling stage were the most suitable models for inversion of phosphorus content in Maize leaves. The correlation between canopy leaf area index and spectral reflectance was negative in the visible band (420 ~ 680 nm), positive in the near infrared band (740 ~ 1120 nm), and the trend of correlation coefficient curve was basically the same in all growth stages, but the correlation coefficient of canopy leaf area index and spectral reflectance was negative in the visible band (420 ~ 680 nm) and positive in the near infrared band (740 ~ 1120 nm). The fitting equations of 21 vegetation indices, PSSRb, NDVI1, NDVI2, PSSRc, MTCI, MSR705, MSR705, MSR705 and DI constructed by the original canopy reflectance spectra have higher determinant coefficients and F values. SR705, MSR705, DI can be better used to monitor the canopy leaf area index of summer maize. 4. From the big trumpet stage to the mature stage, the above-ground dry matter of summer maize showed a gradual increase trend under different nitrogen and phosphorus levels, and all accorded with the typical "S" growth curve, the growth rate was "slow-fast-slow". Under the same phosphorus application level, the above-ground dry matter increased with the increase of nitrogen application rate, and the dry matter accumulation also increased with the increase of phosphorus application rate. The vegetation indices GNDVI, PSSRc, NDVI4 and DI of big trumpet stage, silking stage, filling stage and mature stage could be used to estimate the above-ground dry matter of summer maize. 5. From jointing stage to filling stage, the leaf water content decreased gradually, and the water content of summer maize during silking stage was decreased. The spectral reflectance of summer maize leaves at jointing stage, big trumpet stage, silking stage and grain filling stage was negatively correlated with leaf moisture content, and there were significant differences at different growth stages. But the correlation coefficient is not high; the 1370-2500 nm band of big trumpet stage, silking stage and filling stage, except near 1900 nm band, the other bands are stable negative correlation, and the correlation coefficient is high. The correlation coefficients of NDWI2 and SIWSI were higher at 1450nm and 1650nm. The regression models of R1450, NDWI2 and NDWI2 at Bell stage, silking stage and grain filling stage could be used to monitor the water content of summer maize leaves. 6. The protein content and yield of summer maize grain increased with the increase of fertilizer application, and the nitrogen and phosphorus fertility treatments were different. In the grain protein content fitting equation, the vegetation index RVI and OSAVI fitting equation have higher determinant coefficient; in the grain yield fitting equation, the vegetation index RVI, GRVI and NDVI have higher determinant coefficient and F value. The precision test results showed that the vegetation index RVI could better predict the protein content and yield of summer maize grain, and had better stability and anti-background interference.
【學(xué)位授予單位】：西北農(nóng)林科技大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2016
【分類號】：S513
，
本文編號：2224135
[Abstract]:Hyperspectral remote sensing technology is an important research direction of modern agricultural information technology. High-efficiency and low-consumption crop nutrition diagnosis and growth monitoring technology is of great significance to accurate crop management and rational fertilization. Nutrition diagnosis, growth monitoring, yield and quality prediction. The physiological and ecological parameters, such as canopy spectral reflectance, leaf nitrogen and phosphorus content, leaf area index, aboveground dry matter, leaf water content, grain quality and yield, were measured under different summer maize varieties, different nitrogen and phosphorus nutrition levels and different growth stages. The effects of different growth stages and nitrogen and phosphorus levels on the spectral reflectance of summer maize canopy and the correlation between spectral reflectance and physiological and ecological parameters were analyzed. Hyperspectral diagnostic and monitoring physiological and ecological parameters models suitable for different summer maize varieties and different growth stages were screened to provide theoretical and practical basis for nutritional diagnosis and growth monitoring of summer maize. The spectral reflectance of summer maize canopy decreased with the increase of nitrogen application in the light band, and increased with the increase of nitrogen application in the near infrared band. The total nitrogen content in summer maize leaves was negatively correlated with the original spectrum from 350 nm to 725 nm, positively correlated with the near infrared spectrum from 740 nm to 1400 nm, and positively correlated with the position of correlation coefficient change from 725 nm to 735 nm. The parameters of I, NDVI1 and grain filling (SDr/SDb, GNDVI, NDVI1) had higher correlation coefficients with leaf total nitrogen content. The optimum parameters of the model were SDr, GNDVI, NDVI1 and NDVI1.2 at jointing stage, big trumpet stage, silking stage and grain filling stage, respectively. In the near infrared band (740 ~ 1380 nm), the spectral reflectance increased with the increase of phosphorus application, but the spectral response curves were not significantly different; in this region, the correlation coefficient was relatively stable, and nine normalized spectral bands (830 n) M, 880 nm, 940 nm, 1100 nm, 1430 nm, 1580 nm, 1650 nm, 1740 nm, 2200 nm are sensitive bands, and the combined parameters of the two bands have higher correlation coefficients. The normalized parameters of jointing stage (R830, R880, R830 + 880, R830 + 940), large trumpet stage (R830, R880, R830 + 880, R830 + 940, R830 + 940), grouting stage (R940, R1100, R880 + 1100, R940) have higher correlation coefficients. The models of R830+880, R830+940 and R880+1100 normalized parameters at jointing stage, big trumpet stage and filling stage were the most suitable models for inversion of phosphorus content in Maize leaves. The correlation between canopy leaf area index and spectral reflectance was negative in the visible band (420 ~ 680 nm), positive in the near infrared band (740 ~ 1120 nm), and the trend of correlation coefficient curve was basically the same in all growth stages, but the correlation coefficient of canopy leaf area index and spectral reflectance was negative in the visible band (420 ~ 680 nm) and positive in the near infrared band (740 ~ 1120 nm). The fitting equations of 21 vegetation indices, PSSRb, NDVI1, NDVI2, PSSRc, MTCI, MSR705, MSR705, MSR705 and DI constructed by the original canopy reflectance spectra have higher determinant coefficients and F values. SR705, MSR705, DI can be better used to monitor the canopy leaf area index of summer maize. 4. From the big trumpet stage to the mature stage, the above-ground dry matter of summer maize showed a gradual increase trend under different nitrogen and phosphorus levels, and all accorded with the typical "S" growth curve, the growth rate was "slow-fast-slow". Under the same phosphorus application level, the above-ground dry matter increased with the increase of nitrogen application rate, and the dry matter accumulation also increased with the increase of phosphorus application rate. The vegetation indices GNDVI, PSSRc, NDVI4 and DI of big trumpet stage, silking stage, filling stage and mature stage could be used to estimate the above-ground dry matter of summer maize. 5. From jointing stage to filling stage, the leaf water content decreased gradually, and the water content of summer maize during silking stage was decreased. The spectral reflectance of summer maize leaves at jointing stage, big trumpet stage, silking stage and grain filling stage was negatively correlated with leaf moisture content, and there were significant differences at different growth stages. But the correlation coefficient is not high; the 1370-2500 nm band of big trumpet stage, silking stage and filling stage, except near 1900 nm band, the other bands are stable negative correlation, and the correlation coefficient is high. The correlation coefficients of NDWI2 and SIWSI were higher at 1450nm and 1650nm. The regression models of R1450, NDWI2 and NDWI2 at Bell stage, silking stage and grain filling stage could be used to monitor the water content of summer maize leaves. 6. The protein content and yield of summer maize grain increased with the increase of fertilizer application, and the nitrogen and phosphorus fertility treatments were different. In the grain protein content fitting equation, the vegetation index RVI and OSAVI fitting equation have higher determinant coefficient; in the grain yield fitting equation, the vegetation index RVI, GRVI and NDVI have higher determinant coefficient and F value. The precision test results showed that the vegetation index RVI could better predict the protein content and yield of summer maize grain, and had better stability and anti-background interference.
【學(xué)位授予單位】：西北農(nóng)林科技大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2016
【分類號】：S513
，
本文編號：2224135
本文鏈接：http://sikaile.net/shoufeilunwen/nykjbs/2224135.html
最近更新
教材專著
