本文選題：高光譜遙感 + 油菜　； 參考：《西北農(nóng)林科技大學》2017年碩士論文

【摘要】：農(nóng)業(yè)是一個國家的基本產(chǎn)業(yè),屬于第一產(chǎn)業(yè),是提供支撐國民經(jīng)濟建設(shè)與發(fā)展的基礎(chǔ)產(chǎn)業(yè)。油菜的平均產(chǎn)量在國產(chǎn)植物油中占比超過了一半。同時,油菜的產(chǎn)油效率較高,因此在我國植物油產(chǎn)業(yè)中具有重要地位。因此快速,準確,大范圍,實時地獲取油菜相關(guān)的生理參數(shù),檢測油菜長勢情況顯得尤為重要。本研究以油菜為研究對象,通過田間試驗和實驗室檢測,分別獲取了油菜冠層高光譜遙感數(shù)據(jù),不同尺度的油菜高光譜遙感影像以及油菜的生理參數(shù)。通過計算機和數(shù)理統(tǒng)計方法,構(gòu)建油菜最佳的生理參數(shù)預測模型,并通過不同尺度的高光譜遙感圖像反演驗證,為油菜的精細農(nóng)業(yè)發(fā)展奠定基礎(chǔ),推動了農(nóng)業(yè)數(shù)字信息化進程。取得的主要結(jié)論如下:(1)從整體趨勢上來看,不同生育期油菜冠層高光譜特征較為一致,在可見光藍紫光和紅光波段范圍有兩個吸收谷,在綠光范圍有明顯的反射峰,在近紅外光范圍形成極高反射區(qū)域。同時隨著冠層葉綠素(SPAD)含量以及氮素含量的增加,在可見光范圍內(nèi)油菜冠層高光譜反射率降低,在近紅外光范圍內(nèi)的極高反射區(qū)域升高。隨著冠層葉綠素(SPAD)含量以及含氮量的增加,紅邊特征也出現(xiàn)了“紅移”的現(xiàn)象。從苗期到開花期,紅邊特征的“紅移”現(xiàn)象明顯,等到了成熟期,紅邊特征逐步顯現(xiàn)出“藍移”的現(xiàn)象。由此可以看出,油菜冠層的生理參數(shù)與高光譜反射率之間存在這密不可分的聯(lián)系。(2)通過相關(guān)處理油菜冠層光譜數(shù)據(jù),提取了多個能夠反映植被生理參數(shù)的光譜特征參數(shù)和植被指數(shù)。同時分別提取與這些光譜參數(shù)對應的不同生育期油菜的葉綠素(SPAD)含量和葉片氮素含量。然后利用這些提取的數(shù)據(jù)進行相關(guān)性分析,結(jié)果證明絕大多數(shù)光譜特征參數(shù)和植被指數(shù)都與葉綠素(SPAD)含量和葉片氮素含量都是極顯著相關(guān),同時全生育期油菜構(gòu)建的葉綠素(SPAD)含量和葉片氮素含量預測模型不僅精度高于分生育期油菜,還避免了單生育期預測模型無法預測其他生育期油菜生理參數(shù)的弊端。(3)在對全生育期油菜的葉綠素(SPAD)含量和葉片氮素含量預測研究過程中發(fā)現(xiàn),通過隨機森林回歸算法,并選擇自變量影響力較好的光譜特征參數(shù)和植被指數(shù)作為構(gòu)建模型的自變量,構(gòu)建的預測模型精度好于普通回歸算法和多元逐步回歸算法。在前人研究的基礎(chǔ)上,提高了油菜的相關(guān)生理參數(shù)預測模型的精度。(4)利用隨機森林回歸算法構(gòu)建的油菜的葉綠素(SPAD)含量和葉片氮素含量預測模型,對不同尺度的油菜高光譜影像進行反演分析,結(jié)果發(fā)現(xiàn)反演結(jié)果基本符合油菜真實養(yǎng)分分布情況。
[Abstract]:Agriculture is the basic industry of a country, belongs to the primary industry, and provides the basic industry to support the construction and development of the national economy. The average yield of rapeseed accounted for more than half of the domestic vegetable oil. At the same time, rapeseed has high oil production efficiency, so it plays an important role in Chinese vegetable oil industry. Therefore, it is very important to obtain the physiological parameters of rapeseed quickly, accurately, in large range and in real time, and to detect the growth of rapeseed. In this study, the rape canopy hyperspectral remote sensing data, the hyperspectral remote sensing images of rape on different scales and the physiological parameters of rape were obtained by field experiments and laboratory tests. Through computer and mathematical statistics method, the best physiological parameter prediction model of rapeseed was constructed, and the inversion verification of hyperspectral remote sensing image of different scales was carried out, which laid a foundation for the development of fine agriculture of rape and promoted the process of agricultural digital information. The main conclusions are as follows: (1) from the overall trend, the hyperspectral characteristics of rape canopy at different growth stages are consistent, and there are two absorption valleys in the band of visible blue violet light and red light, and obvious reflection peaks in the green light range. An extremely high reflectance region is formed in the near infrared region. At the same time, with the increase of chlorophyll (Spad) content and nitrogen content in the canopy, the hyperspectral reflectance of rape canopy decreased in the visible light range, and the extremely high reflectance region increased in the near infrared region. With the increase of chlorophyll (Spad) content and nitrogen content in the canopy, "red shift" appeared in the red edge. From seedling stage to flowering stage, the "red shift" phenomenon of red edge characteristics was obvious, and the "blue shift" phenomenon appeared gradually at maturity stage. It can be seen that there is a close relationship between the physiological parameters of rape canopy and hyperspectral reflectance. (2) processing rape canopy spectral data by correlation, Several spectral characteristic parameters and vegetation indices which can reflect the physiological parameters of vegetation were extracted. At the same time, the chlorophyll (Spad) content and leaf nitrogen content of rape at different growth stages were extracted respectively. The results showed that most of the spectral characteristic parameters and vegetation index were significantly correlated with chlorophyll (Spad) content and leaf nitrogen content. At the same time, the prediction precision of chlorophyll (Spad) content and leaf nitrogen content of rapeseed at the whole growth stage was higher than that of the rapeseed at different growth stage. It also avoids the disadvantage that the single growth stage prediction model can not predict the physiological parameters of other growth stages of rape. (3) in the process of prediction of chlorophyll (Spad) content and leaf nitrogen content in rapeseed during the whole growth period, it was found that the stochastic forest regression algorithm was used. The spectral characteristic parameters and vegetation index which have better influence on independent variables are chosen as independent variables to construct the model, and the prediction model is better than ordinary regression algorithm and multivariate stepwise regression algorithm. On the basis of previous studies, the precision of the prediction model of relative physiological parameters of rapeseed was improved. (4) the prediction model of chlorophyll (Spad) content and leaf nitrogen content of rapeseed was constructed by using stochastic forest regression algorithm. The inversion analysis of different scales of rape hyperspectral images shows that the inversion results basically accord with the true nutrient distribution of rapeseed.
【學位授予單位】：西北農(nóng)林科技大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：S565.4

【參考文獻】

相關(guān)期刊論文 前10條

1 喻俊;李曉敏;張權(quán);侍昊;褚軍;;基于高光譜遙感的植被冠層氮素反演方法研究進展[J];陜西林業(yè)科技;2016年06期

2 秦占飛;常慶瑞;謝寶妮;申健;;基于無人機高光譜影像的引黃灌區(qū)水稻葉片全氮含量估測[J];農(nóng)業(yè)工程學報;2016年23期

3 田明璐;班松濤;常慶瑞;由明明;羅丹;王力;王爍;;基于低空無人機成像光譜儀影像估算棉花葉面積指數(shù)[J];農(nóng)業(yè)工程學報;2016年21期

4 劉京;常慶瑞;劉淼;殷紫;馬文君;;基于SVR算法的蘋果葉片葉綠素含量高光譜反演[J];農(nóng)業(yè)機械學報;2016年08期

5 李媛媛;常慶瑞;劉秀英;嚴林;羅丹;王爍;;基于高光譜和BP神經(jīng)網(wǎng)絡(luò)的玉米葉片SPAD值遙感估算[J];農(nóng)業(yè)工程學報;2016年16期

6 楚萬林;齊雁冰;常慶瑞;蒲潔;;棉花冠層葉片葉綠素含量與高光譜參數(shù)的相關(guān)性[J];西北農(nóng)林科技大學學報(自然科學版);2016年09期

7 李仲;劉明地;吉守祥;;基于紅外光譜和隨機森林的枸杞產(chǎn)地鑒別[J];計算機與應用化學;2016年07期

8 馮海寬;李振海;金秀良;楊貴軍;萬鵬;郭建華;于海洋;楊福芹;李偉國;王衍安;;基于EFAST和PLS的蘋果葉片等效水厚度高光譜估算[J];農(nóng)業(yè)工程學報;2016年12期

9 李粉玲;常慶瑞;申健;王力;;基于GF-1衛(wèi)星數(shù)據(jù)的冬小麥葉片氮含量遙感估算[J];農(nóng)業(yè)工程學報;2016年09期

10 王輝;田海清;李哲;徐琳;李斐;史樹德;;基于高光譜的甜菜SPAD值估算研究[J];農(nóng)機化研究;2016年05期

相關(guān)博士學位論文 前5條

1 劉秀英;玉米生理參數(shù)及農(nóng)田土壤信息高光譜監(jiān)測模型研究[D];西北農(nóng)林科技大學;2016年

2 劉冰峰;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[D];西北農(nóng)林科技大學;2016年

3 李粉玲;關(guān)中地區(qū)冬小麥葉片氮素高光譜數(shù)據(jù)與衛(wèi)星影像定量估算研究[D];西北農(nóng)林科技大學;2016年

4 秦占飛;西北地區(qū)水稻長勢遙感監(jiān)測研究[D];西北農(nóng)林科技大學;2016年

5 王淵;不同水平油菜氮素含量遙感信息提取方法研究[D];浙江大學;2008年

相關(guān)碩士學位論文 前10條

1 查海涅;基于衛(wèi)星遙感的水稻生長監(jiān)測與氮素營養(yǎng)診斷系統(tǒng)[D];安徽科技學院;2016年

2 殷紫;不同生育期冬小麥生理生化參數(shù)高光譜估測研究[D];西北農(nóng)林科技大學;2016年

3 劉淼;不同營養(yǎng)水平冬小麥長勢高光譜遙感監(jiān)測[D];西北農(nóng)林科技大學;2016年

4 高雨茜;夏玉米葉綠素、葉面積指數(shù)高光譜估測研究[D];西北農(nóng)林科技大學;2016年

5 馬文君;棉花生理生化參數(shù)高光譜反演模型研究[D];西北農(nóng)林科技大學;2016年

6 王曉星;夏玉米冠層光譜特征及其生理生態(tài)參量的高光譜估算模型[D];西北農(nóng)林科技大學;2015年

7 李金夢;基于高光譜成像技術(shù)的柑橘葉片含氮量快速檢測方法研究[D];浙江大學;2014年

8 賈雯晴;基于高光譜的小麥水分狀況監(jiān)測研究[D];南京農(nóng)業(yè)大學;2013年

9 李軍偉;基于高光譜遙感的樟樹幼樹生理生化參量的反演[D];華中農(nóng)業(yè)大學;2011年

10 劉秀英;樟樹幼林主要生物化學參數(shù)高光譜遙感估算模型研究[D];中南林業(yè)科技大學;2006年

，

本文編號：2068246