本文關(guān)鍵詞： 鹽漬化土壤 微生物修復(fù) 光譜變換 偏最小二乘法　出處：《光譜學(xué)與光譜分析》2017年05期 　論文類型：期刊論文

【摘要】：基于鹽漬土修復(fù)過程中鹽分含量和同步實(shí)測光譜數(shù)據(jù),通過對(duì)原始光譜數(shù)據(jù)、平滑光譜數(shù)據(jù)及平滑后的不同變換光譜數(shù)據(jù)等八種光譜數(shù)據(jù)集,分別以相關(guān)系數(shù)的極值和不同相關(guān)系數(shù)范圍兩種方法分析其最佳敏感波段范圍,深入分析了不同變換下土壤的光譜響應(yīng)特征。在此基礎(chǔ)上,運(yùn)用偏最小二乘回歸方法,以全波段(400~1 650nm)和分析獲得的最佳敏感波段建立了基于修復(fù)過程的土壤鹽含量和光譜反射率的關(guān)系模型。結(jié)果表明:針對(duì)八種光譜數(shù)據(jù)集,采用兩種方法提取的土壤最佳敏感波段,均集中在947.11~949.31,1 340.27,1 394.11,1 419,1 457.81~1 461.31,1 537.68~1 551.39和1 602.32nm;且最佳波段的土壤鹽含量反演模型,以模型評(píng)價(jià)參數(shù)的決定系數(shù)(R2)和均方根誤差(RMSE),以及赤池信息量準(zhǔn)則(akaike’s information criterion,AIC)作為選擇最佳模型的標(biāo)準(zhǔn),均以SGSD(Log R)模型的建模和預(yù)測結(jié)果比其他光譜變換的模型更為顯著�；谌ǘ蔚腜LSR建模效果總體上稍優(yōu)于最佳波段的模型,其中以SGSD的預(yù)測精度最為突出,其模型的決定系數(shù)R2與標(biāo)準(zhǔn)差RMSEP分別為0.673和1.256;基于兩種方法獲得的最佳波段的PLSR模型與全波段對(duì)比在模型精度方面雖有一定差距,但從模型的復(fù)雜程度比較,具有模型簡單、變量更少及運(yùn)算量小的特點(diǎn)。該研究可在土壤鹽含量及其光譜特征的研究中,為實(shí)現(xiàn)土壤鹽漬化定量、快速、便捷的監(jiān)測和檢測提供參考。
[Abstract]:Based on the salt content and synchronous measured spectral data in the process of saline soil remediation, eight spectral data sets, such as original spectral data, smoothing spectral data and different transformed spectral data after smoothing, are analyzed. The optimum sensitive band range is analyzed by the extreme value of correlation coefficient and the range of correlation coefficient, and the spectral response characteristics of soil under different transformation are analyzed in depth. On this basis, partial least square regression method is used. The relationship model of soil salt content and spectral reflectance based on remediation process was established by using the whole wave band (1 650 nm) and the best sensitive band. The results show that: for the eight kinds of spectral data sets, the relationship between soil salt content and spectral reflectivity is obtained. The best sensitive bands extracted by the two methods are concentrated at 1 340.27 ~ 1 394.11 ~ 1 4191.457.81 ~ 1 ~ 1 461.68 ~ 1 537.68 ~ 1 551.39 and 1 602.32 nm, respectively, and the inversion model of soil salt content in the optimum band is obtained. The determination coefficient of model evaluation parameters (R2) and root mean square error (RMSE), as well as the red pool information quantity criterion (AICs information criteria) are used as the criteria for selecting the best model. The modeling and prediction results of SGSD(Log R) model are more remarkable than those of other spectral transformation models. The modeling effect of PLSR based on the whole band is slightly better than that of the best band model, and the prediction accuracy of SGSD is the most prominent. The determination coefficient R2 and standard deviation RMSEP of the model are 0.673 and 1.256, respectively. Although there are some differences in the accuracy of the model between the best band PLSR model and the full-band model, the model is simple compared with the complexity of the model. This study can be used as a reference for quantitative, rapid and convenient monitoring and detection of soil salinization in the study of soil salt content and spectral characteristics.
【作者單位】： 上海交通大學(xué)農(nóng)業(yè)與生物學(xué)院低碳農(nóng)業(yè)研究中心;農(nóng)業(yè)部都市農(nóng)業(yè)(南方)重點(diǎn)實(shí)驗(yàn)室;上海交通大學(xué)船舶海洋與建筑工程學(xué)院;
【基金】：高分國土資源遙感應(yīng)用示范系統(tǒng)(一期)項(xiàng)目(04-Y30B01-9001-12/15) 國家自然科學(xué)基金項(xiàng)目(41471120) 社科重大項(xiàng)目(14ZDB139) 上海交大農(nóng)工交叉項(xiàng)目(Agri-X2015004)資助
【分類號(hào)】：S156.4

【相似文獻(xiàn)】

相關(guān)期刊論文 前2條

1 ;五、P23 攝影測量與遙感學(xué)[J];測繪文摘;2000年01期

2 ;[J];;年期

相關(guān)會(huì)議論文 前1條

1 唐宏;楊新;方濤;施鵬飛;;基于核方法的光譜角制圖模型及其在高光譜圖像分割中的應(yīng)用[A];中國地理信息系統(tǒng)協(xié)會(huì)第八屆年會(huì)論文集[C];2004年

相關(guān)碩士學(xué)位論文 前2條

1 何磊;基于實(shí)測光譜的昆承湖水質(zhì)遙感模型及反演[D];成都理工大學(xué);2010年

2 鄭標(biāo);基于GIS的青海湖流域典型地物光譜數(shù)據(jù)庫研究[D];青海師范大學(xué);2015年

，

本文編號(hào)：1506398