园林植物滞尘能力分析及其叶面滞尘高光谱定量遥感模型

ANALYSIS OF DUST DETENTION CAPABILITY OF LANDSCAPE PLANTS AND THE HYPERSPECTRAL REMOTE SENSING QUANTITATIVE MODELS CONSTRUCTION OF FOLIAGEDUST DETENTION

园林植物对于改善大气污染具有重要意义。通过研究武汉地区不同大气粉尘污染下园林植物滞尘能力,筛选出了不同大气粉尘污染下滞尘能力最强的园林植物,在华中农业大学试验区(无污染区)滞尘能力最强的前3种植物分别是紫薇>枇杷>夹竹桃,青山试验区(中度污染区)是杜鹃>紫薇>枇杷,而在武钢试验区(重度污染区)则是法国冬青>云南黄馨>青桐。随着粉尘污染的加重,吉祥草是滞尘量增量最大的植物,其在武钢的滞尘量是华农的185倍;而紫薇和紫叶李的滞尘量则是增量最小的。分别以光谱反射率、光谱反射率一阶导数及光谱吸收深度作为自变量,滞尘量作为因变量,采用多元线性回归方程分别对乔木和灌草建立滞尘量-光谱反演模型,其中以光谱反射率一阶导数作为自变量得到的反演模型精度最高,乔木和灌草的滞尘量反演模型的判定系数分别达到了0.896和0.928,模型预测精度分别达到了57.87%和64.51%。这为采用高光谱遥感技术定量预测植物叶面滞尘量提供了理论依据。

The effect of dust detention capability of landscape plants to the environment improvement is significance.In this paper,we selected the landscape plants with higher accumulation ability of dust based on the quantitative research of dust detention capability at different atmospheric pollution areas of Wuhan.Results indicated that:the three plants with highest dust detention capability were crape myrtle>loquat> oleander at Huanong experiment area(no pollution area);those plants with highest dust detention capability were rhododendron > crape myrtle > loquat at Qingshan experiment area(medium pollution area),however,at Wugang area(severe pollution area),those plants with highest dust detention capability were viburnum odoratissimum >jasminum mesnyi hance> Phoenix tree.With the atmospheric pollution aggravating,the increment of dust detention capability of reineckia carnea was the largest compared with other plants,of which the dust detention content at Wugang was 185-fold than that at Huanong,however,the increment of dust detention abilities of crape myrtle and purple-leaf plum were least.This investigation constructed the spectral inversion models via multiple linear regression equation for arbors or shrubs and herbals by using reflectance spectrum characteristic values as independent variables and dust detention volume as the dependent variables.The spectral inversion models which were constructed via the first derivative of the reflectance spectrum as independent variables had the highest precision,and the determination coefficients were 0.896 and 0.928,the prediction accuracy were 57.87%and 64.51%for arbors or shrubs and herbals,respectively.This research could provide a theoretical basis for the quantitative prediction of the dust detention via hyperspectrum remote sensing technology.