本文關(guān)鍵詞：稀疏植被區(qū)土壤表層介電特性分析及水鹽信息提取研究　出處：《新疆大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 介電常數(shù) 微波遙感 RadarSat-2 高分一號 BP神經(jīng)網(wǎng)絡(luò) AIEM模型

【摘要】：新疆維吾爾自治區(qū)地處我國西北地區(qū),因水資源空間分布極不均衡,導(dǎo)致荒漠多、綠洲少,加之新疆大部分區(qū)域的降雨較少、氣候干燥,生態(tài)環(huán)境極其脆弱。與此同時,新疆位于中緯度地區(qū),光照資源豐富,農(nóng)作物產(chǎn)量高且質(zhì)量優(yōu),作為我國糧食后備儲備基地,擔(dān)負(fù)著糧食安全、社會安全、國家安全等重大戰(zhàn)略。近年來,由于區(qū)域農(nóng)田的不合理開發(fā)、地下水資源的濫用等現(xiàn)象,區(qū)域生態(tài)環(huán)境日益惡化,尤其在綠洲-荒漠交錯區(qū)域,大面積土地發(fā)生退化、生態(tài)多樣性驟減。造成土地退化的因素較多,而土壤水分匱乏、鹽分含量高是綠洲-荒漠交錯區(qū)域土地退化的重要因素。目前,新疆境內(nèi)受鹽漬化脅迫的耕地面積約占耕地總面積的1/3,土壤鹽漬化現(xiàn)象已嚴(yán)重威脅到區(qū)域生態(tài)系統(tǒng)協(xié)調(diào)發(fā)展,抑制了農(nóng)作物的正常生長、生產(chǎn),帶給綠洲經(jīng)濟(jì)的損失不容忽視。本文選擇新疆典型綠洲——渭-庫綠洲為研究區(qū),該區(qū)域主要靠渭干河和庫車河水系補(bǔ)給形成綠洲,綠洲穩(wěn)定性受水資源影響較大。在西部大開發(fā)等相關(guān)政策的刺激下,渭-庫綠洲境內(nèi)縣鄉(xiāng)的社會經(jīng)濟(jì)水平逐年提高,農(nóng)用地面積迅猛擴(kuò)增,巨大的水量需求加劇了水資源空間分布的不均衡性,綠洲下游、綠洲周邊生態(tài)系統(tǒng)持續(xù)惡化、稀疏植被覆蓋區(qū)土地退化現(xiàn)象加劇,綠洲至荒漠區(qū)域間的緩沖區(qū)域日漸縮減�？臻g傳感手段的發(fā)展實(shí)現(xiàn)了大區(qū)域、多時相的土壤信息提取技術(shù),在相關(guān)技術(shù)手段快速發(fā)展的背景下,土壤信息監(jiān)測精度逐步增高,尤其是高分辨率(光譜、空間)遙感數(shù)據(jù)更是為土壤信息監(jiān)測精度提供了保障。本文以提取新疆渭-庫綠洲局部地表土壤水分、鹽分信息為研究目的,選擇多極RadarSat-2雷達(dá)影像和國產(chǎn)高分一號多光譜影像為數(shù)據(jù)源,(1)以實(shí)測介電常數(shù)實(shí)部和虛部為基礎(chǔ),運(yùn)用多種數(shù)學(xué)變換形式,建立土壤水分、鹽分的介電常數(shù)逐步回歸方程;(2)運(yùn)用AIEM模型模擬一定入射角下,地表粗糙度、土壤水分與雷達(dá)后向散射系數(shù)間的擬合研究,并建立裸露地表土壤含水量經(jīng)驗(yàn)?zāi)Ｐ?(3)優(yōu)選GF-1的多種光譜指數(shù),利用BP神經(jīng)網(wǎng)絡(luò)模型針對雷達(dá)數(shù)據(jù)和多光譜數(shù)據(jù)建立土壤鹽分模型,并進(jìn)行驗(yàn)證和精度評估。主要得出以下結(jié)論:1、土壤含水量對介電常數(shù)實(shí)部有較大的影響且相關(guān)性可達(dá)0.94,虛部不僅受含水量的影響,也受含鹽量的影響;在鹽分較大的情況下,虛部隨著含水量的增加而增大,虛部與土壤鹽分含量相關(guān)性較弱,僅0.45,但與含水量高度相關(guān),相關(guān)性可達(dá)0.90。此外,利用介電常數(shù)的5種變換形式,所建立土壤水分實(shí)部模型精度較高,其中,倒數(shù)模型為土壤水分最優(yōu)模型,建模精度的決定系數(shù)R2為0.82、驗(yàn)證精度R2為0.91;與此同時,運(yùn)用介電常數(shù)虛部綜合模型模擬土壤鹽分,鹽分指數(shù):(水分+鹽分)2為最優(yōu)模型,建模精度的決定系數(shù)R2為0.84、驗(yàn)證精度R2為0.91。水分與實(shí)部、虛部均呈良好的線性關(guān)系,而鹽分與虛部呈非線性關(guān)系。2、利用AIEM模型模擬了區(qū)域地表粗糙度和土壤水分情況,區(qū)域地表粗糙度與實(shí)際相符,綠洲-荒漠交錯區(qū)、河道附近的粗糙度較大;土壤水分精度較高,決定系數(shù)R2達(dá)0.86,土壤含水量呈西高東低分布,綠洲邊緣和河道附近土壤含水量較高。3、由于土壤鹽分線性經(jīng)驗(yàn)?zāi)Ｐ碗y以準(zhǔn)確評估土壤鹽分含量,因此結(jié)合多光譜輔助數(shù)據(jù)對后向散射信息進(jìn)行分割。利用RadarSat-2四極化后向散射系數(shù)、及地表粗糙度和土壤水分的雷達(dá)提取數(shù)據(jù)、GF-1提取的光譜指數(shù)作數(shù)據(jù)集,并構(gòu)建BP神經(jīng)網(wǎng)絡(luò)模型。經(jīng)多次試驗(yàn),BP神經(jīng)網(wǎng)絡(luò)模型較好地提取了區(qū)域土壤鹽分信息,模型模擬精度達(dá)到了78.95%,因此從一定程度上驗(yàn)證了光學(xué)遙感與主動微波遙感結(jié)合進(jìn)行鹽漬區(qū)信息提取的有效性。綜上所述,本研究通過主動微波遙感特征,利用AIEM物理模型建立了適用于裸露區(qū)土壤水分反演的經(jīng)驗(yàn)?zāi)Ｐ?同時,結(jié)合多光譜數(shù)據(jù),利用BP算法建立了土壤鹽分模型,且模擬精度較高,準(zhǔn)同步地表土壤水鹽研究可為干旱區(qū)土壤信息提取工作提供一定的參考。
[Abstract]:The Xinjiang Uygur Autonomous Region is located in the northwest of China, because of the spatial distribution of water resources is extremely uneven, resulting in much less in most areas of Xinjiang oasis, the less rainfall, dry climate, the ecological environment is extremely fragile. At the same time, Xinjiang is located in the middle latitudes, the light is rich in resources, high crop yield and quality, as our food reserve base, responsible for food safety, social security, major national security strategy. In recent years, due to the unreasonable development of regional farmland, abuse of groundwater resources, the regional ecological environment is deteriorating, especially in the oasis desert ecotone, a large area of land degradation, ecological diversity caused more sharply. The factors of land degradation, and soil water scarcity, high salt content is an important factor in the oasis desert ecotone of land degradation. At present, the territory of Xinjiang by saline stress The cultivated land area accounts for about 1/3 of the total farmland, soil salinization has been a serious threat to the sustainable development of eco system, inhibit the normal growth of crop production, to oasis economic losses can not be ignored. This paper chooses Xinjiang oasis - Wei library oasis as the research area, the region mainly rely on the Weigan River and Kuche river water supply forms oasis, oasis stability affected by water greatly. The policy of western development and stimulation, social economic level Wei library oasis in the county increased year by year, agricultural land area of rapid expansion, huge water demand exacerbated the imbalance, the spatial distribution of water resources in Oasis downstream, continued to deteriorate around the oasis ecosystem, increasing land sparse vegetation degradation, a buffer zone between oasis to desert areas is increasingly reduced. The development space of the sensing region Time domain, soil information extraction phase, with the rapid development of related technology under the background of information monitoring accuracy of soil gradually increased, especially high resolution (spectral, spatial) remote sensing data is provided a guarantee for monitoring accuracy. Based on the soil information extraction of Xinjiang Wei library oasis local soil moisture, salinity information for the purpose of the study, select the multipole RadarSat-2 radar image and domestic high one multi spectral image as the data source, (1) to the measured dielectric constant of the real and imaginary parts as the foundation, using a variety of mathematical transformations, the establishment of the soil moisture, dielectric constant salinity regression equation; (2) using AIEM model a certain incident angle, surface roughness, soil moisture and radar to the fitting of scattering coefficient, and the establishment of bare soil moisture model; (3) multi spectral index optimization GF-1, using BP neural network Network model for radar data and multi spectral data to establish the model of soil salinity, and the validation and accuracy assessment. Main conclusions are as follows: 1. Soil moisture content on the real part of permittivity and the influence of correlation is the larger of the 0.94, the imaginary part is not only affected by water content, is also affected by salinity; in the situation of salt is large, the imaginary part increases with the increase of water content, and the imaginary part correlation of soil salt content is only 0.45, but is highly correlated with the water content, the correlation can reach 0.90. in addition, using 5 kinds of transformations of the dielectric constants of the established real soil moisture model is of high precision, the countdown model for optimal decision model of soil moisture, the modeling accuracy of coefficient R2 is 0.82, verify the accuracy of R2 is 0.91; at the same time, using the imaginary part of the dielectric constant of the integrated model simulation of soil salinity, salt index: (water + salt) 2 is the optimal model, modeling The accuracy of the determination coefficient of R2 is 0.84, R2 is 0.91. water and verify the accuracy of the real part, the imaginary part showed a good linear relationship, and the imaginary part of the non-linear relationship between salinity and.2, simulation of the regional surface roughness and soil moisture conditions using the AIEM model, the regional surface roughness and the actual match, the oasis desert crisscross District, near the river roughness larger; soil moisture with high precision, the coefficient of determination R2 reached 0.86, soil moisture in the East West High low distribution, the edge of oasis and the river near the soil moisture was higher than.3, because of the linear empirical model of soil salinity is difficult to accurately assess the soil salt content, so the combination of multi spectral data of auxiliary backscattering information segmentation. By using RadarSat-2 four polarization backscattering coefficient, data extraction and surface roughness and soil moisture on the radar, the spectral index GF-1 extracted data set, and the BP neural network model was built by multiple. Test, the BP neural network model used to extract regional soil salinity information model, simulation accuracy reached 78.95%, so to some extent, the verification of optical remote sensing and active microwave remote sensing combined with effective extraction of saline region information. In summary, this study through the characteristics of active microwave remote sensing, established empirical model exposed area soil moisture inversion using the physical model of AIEM; at the same time, combined with multi spectral data, established the soil salinity model by using BP algorithm, and the simulation precision, quasi synchronous study on soil water and salt surface can provide some reference for information extraction of soil in arid areas.

【學(xué)位授予單位】：新疆大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：S156.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 趙昕;黃妮;宋現(xiàn)鋒;李增元;牛錚;;基于Radarsat2與Landsat8協(xié)同反演植被覆蓋地表土壤水分的一種新方法[J];紅外與毫米波學(xué)報;2016年05期

2 馬紅章;劉素美;彭愛華;孫林;孫根云;;L波段主被動微波協(xié)同反演裸土土壤水分[J];農(nóng)業(yè)工程學(xué)報;2016年19期

3 劉全明;成秋明;王學(xué);李相君;;河套灌區(qū)土壤鹽漬化微波雷達(dá)反演[J];農(nóng)業(yè)工程學(xué)報;2016年16期

4 丁建麗;陳文倩;陳蕓;;干旱區(qū)土壤鹽漬化災(zāi)害預(yù)警——以渭-庫綠洲為例[J];中國沙漠;2016年04期

5 曹雷;丁建麗;牛增懿;;基于TVDI的艾比湖地區(qū)土壤水分時空變化分析[J];水土保持研究;2016年03期

6 曹雷;丁建麗;玉米提·哈力克;蘇雯;寧娟;繆琛;李煥;;基于國產(chǎn)高分一號衛(wèi)星數(shù)據(jù)的區(qū)域土壤鹽漬化信息提取與建模[J];土壤學(xué)報;2016年06期

7 曹雷;丁建麗;于海洋;;渭-庫綠洲多尺度景觀格局與鹽度關(guān)系[J];農(nóng)業(yè)工程學(xué)報;2016年03期

8 牛增懿;丁建麗;李艷華;王爽;王璐;馬成霞;;基于高分一號影像的土壤鹽漬化信息提取方法[J];干旱區(qū)地理;2016年01期

9 張東;塔西甫拉提·特依拜;張飛;阿爾達(dá)克·克里木;;分?jǐn)?shù)階微分在高光譜數(shù)據(jù)估算礦區(qū)土壤砷含量中的應(yīng)用[J];中國礦業(yè);2015年11期

10 于海云;王志軍;李彪;于海立;;內(nèi)蒙古河套灌區(qū)融解期土壤鹽分多極化雷達(dá)響應(yīng)分析[J];長江科學(xué)院院報;2015年11期

相關(guān)博士學(xué)位論文 前3條

1 尼格拉·塔什甫拉提;渭干河—庫車河三角綠洲土壤鹽漬化遙感與近感協(xié)同監(jiān)測方法研究[D];新疆大學(xué);2014年

2 李洋洋;集成光學(xué)與微波遙感蘇打鹽堿地水鹽含量反演方法研究[D];中國科學(xué)院研究生院（東北地理與農(nóng)業(yè)生態(tài)研究所）;2014年

3 韓桂紅;干旱區(qū)鹽漬地極化雷達(dá)土壤水分反演研究[D];新疆大學(xué);2013年

相關(guān)碩士學(xué)位論文 前3條

1 劉增燦;微波散射測量及土壤水分反演研究[D];電子科技大學(xué);2009年

2 買買提·沙吾提;微波遙感數(shù)據(jù)在鹽漬地信息提取中的應(yīng)用研究[D];新疆大學(xué);2008年

3 熊文成;含水含鹽土壤介電特性及反演研究[D];中國科學(xué)院研究生院（遙感應(yīng)用研究所）;2005年

，

本文編號：1400726