基于無(wú)人機(jī)高光譜數(shù)據(jù)的多類(lèi)型混合作物L(fēng)AI反演及尺度效應(yīng)分析
發(fā)布時(shí)間：2018-10-26 20:33
【摘要】：葉面積指數(shù)(Leaf Area Index,LAI)作為表征不同作物生長(zhǎng)狀況的基本參數(shù),是農(nóng)業(yè)精細(xì)化管理及農(nóng)田生態(tài)系統(tǒng)建模的關(guān)鍵。我國(guó)農(nóng)田作物種植比較離散,受地表空間結(jié)構(gòu)非均一性和反演模型非線(xiàn)性等因素影響,不同尺度遙感數(shù)據(jù)估算的作物L(fēng)AI存在一定的差異,即農(nóng)田作物L(fēng)AI的遙感反演普遍存在尺度效應(yīng)問(wèn)題。以包頭遙感綜合驗(yàn)證場(chǎng)農(nóng)業(yè)示范區(qū)為研究區(qū),利用無(wú)人機(jī)高光譜數(shù)據(jù)結(jié)合PROSPECT+SAIL模型構(gòu)建典型農(nóng)作物區(qū)多類(lèi)型作物的查找表(Look-Up-Table,LUT)反演農(nóng)田L(fēng)AI,研究查找表用于玉米、馬鈴薯、向日葵、瓜地等不同作物L(fēng)AI反演的適用性和精度;通過(guò)無(wú)人機(jī)高光譜數(shù)據(jù)聚合獲得多尺度遙感數(shù)據(jù)源,結(jié)合Taylor展開(kāi)理論和計(jì)算幾何模型,提出了一種既考慮類(lèi)間差異又考慮類(lèi)內(nèi)異質(zhì)性的尺度轉(zhuǎn)換模型,定量描述多種作物混合的非均一地表LAI反演過(guò)程中的尺度效應(yīng)特征。結(jié)果表明:基于分類(lèi)和參數(shù)敏感性分析的LUT方法能很好地應(yīng)用于包頭典型農(nóng)作物區(qū)多類(lèi)型混合作物L(fēng)AI反演,總估算精度為相關(guān)系數(shù)R~2=0.82、均方根誤差RMSE=0.43m~2/m~2。隨著反演尺度的增加,作物類(lèi)間差異造成的反演偏差明顯高于類(lèi)內(nèi)異質(zhì)性,利用本文所提出的尺度轉(zhuǎn)換模型均能較好糾正低分辨率LAI反演的尺度效應(yīng)問(wèn)題。
[Abstract]:Leaf area index (Leaf Area Index,LAI), as the basic parameter of different crop growth status, is the key to agricultural fine management and farmland ecosystem modeling. Crop cultivation in China is relatively discrete and affected by the heterogeneity of surface spatial structure and nonlinear inversion model. There are some differences in crop LAI estimation based on remote sensing data of different scales in China. In other words, the scale effect problem exists in the remote sensing inversion of crop LAI. Taking the agricultural demonstration area of Baotou remote sensing comprehensive verification farm as the research area, using UAV hyperspectral data combined with PROSPECT SAIL model to construct the lookup table (Look-Up-Table,LUT) of multi-type crops in typical crop area to invert farmland LAI,. The applicability and accuracy of LAI inversion of corn, potato, sunflower, sunflower and melon field were studied. Based on the Taylor expansion theory and computational geometry model, a scale conversion model considering both intra-class heterogeneity and inter-class difference is proposed by aggregation of hyperspectral data from UAV to multi-scale remote sensing data source. The scale effect characteristics of heterogeneous surface LAI inversion with multiple crop mixtures are quantitatively described. The results show that the LUT method based on classification and parameter sensitivity analysis can be well applied to the LAI inversion of mixed crops in Baotou typical crop area. The total accuracy of the estimation is the correlation coefficient RG20.82, and the root mean square error (RMSE=0.43m~2/m~2.). With the increase of inversion scale, the inversion deviation caused by crop difference is obviously higher than that of intra-class heterogeneity. Using the scale conversion model proposed in this paper, the scale effect of low-resolution LAI inversion can be well corrected.
【作者單位】： 中國(guó)科學(xué)院定量遙感信息技術(shù)重點(diǎn)實(shí)驗(yàn)室中國(guó)科學(xué)院光電研究院;中國(guó)科學(xué)院大學(xué);
【基金】：國(guó)家863計(jì)劃項(xiàng)目“遙感載荷性能與數(shù)據(jù)質(zhì)量檢測(cè)技術(shù)”(2013AA122102)
【分類(lèi)號(hào)】：S127;TP79
本文編號(hào)：2296870
[Abstract]:Leaf area index (Leaf Area Index,LAI), as the basic parameter of different crop growth status, is the key to agricultural fine management and farmland ecosystem modeling. Crop cultivation in China is relatively discrete and affected by the heterogeneity of surface spatial structure and nonlinear inversion model. There are some differences in crop LAI estimation based on remote sensing data of different scales in China. In other words, the scale effect problem exists in the remote sensing inversion of crop LAI. Taking the agricultural demonstration area of Baotou remote sensing comprehensive verification farm as the research area, using UAV hyperspectral data combined with PROSPECT SAIL model to construct the lookup table (Look-Up-Table,LUT) of multi-type crops in typical crop area to invert farmland LAI,. The applicability and accuracy of LAI inversion of corn, potato, sunflower, sunflower and melon field were studied. Based on the Taylor expansion theory and computational geometry model, a scale conversion model considering both intra-class heterogeneity and inter-class difference is proposed by aggregation of hyperspectral data from UAV to multi-scale remote sensing data source. The scale effect characteristics of heterogeneous surface LAI inversion with multiple crop mixtures are quantitatively described. The results show that the LUT method based on classification and parameter sensitivity analysis can be well applied to the LAI inversion of mixed crops in Baotou typical crop area. The total accuracy of the estimation is the correlation coefficient RG20.82, and the root mean square error (RMSE=0.43m~2/m~2.). With the increase of inversion scale, the inversion deviation caused by crop difference is obviously higher than that of intra-class heterogeneity. Using the scale conversion model proposed in this paper, the scale effect of low-resolution LAI inversion can be well corrected.
【作者單位】： 中國(guó)科學(xué)院定量遙感信息技術(shù)重點(diǎn)實(shí)驗(yàn)室中國(guó)科學(xué)院光電研究院;中國(guó)科學(xué)院大學(xué);
【基金】：國(guó)家863計(jì)劃項(xiàng)目“遙感載荷性能與數(shù)據(jù)質(zhì)量檢測(cè)技術(shù)”(2013AA122102)
【分類(lèi)號(hào)】：S127;TP79
【相似文獻(xiàn)】
相關(guān)期刊論文 前4條
1 汪小欽;葉煒;江洪;;基于光譜歸一化的闊葉林LAI遙感估算模型適用性分析[J];福州大學(xué)學(xué)報(bào)(自然科學(xué)版);2011年05期
2 彭哠,張樹(shù)文;基于NDVI與LAI的水稻生長(zhǎng)狀況研究[J];東北測(cè)繪;2002年04期
3 崔要奎;趙開(kāi)廣;范聞捷;徐希孺;;機(jī)載Lidar數(shù)據(jù)的農(nóng)作物覆蓋度及LAI反演[J];遙感學(xué)報(bào);2011年06期
4 ;[J];;年期
相關(guān)碩士學(xué)位論文 前1條
1 邵田田;玉米葉綠素和LAI的高光譜遙感模型研究[D];中國(guó)科學(xué)院研究生院(東北地理與農(nóng)業(yè)生態(tài)研究所);2012年
，本文編號(hào)：2296870
本文鏈接：http://sikaile.net/kejilunwen/zidonghuakongzhilunwen/2296870.html
最近更新
教材專(zhuān)著
