基于不同植被指数的植被-土壤混合像元高光谱偏振信息与模型研究

发布时间：2018-04-29 23:24

本文选题：高光谱 + 偏振　；参考：《光谱学与光谱分析》2017年11期

【摘要】：高光谱遥感被越来越多的应用于确定混合像元的地物组分和比例。将不同面积比例的植被-土壤混合像元作为研究对象,使用偏振装置和ASD FieldSpec~汶3光谱仪得到植被—土壤组成的混合像元的偏振反射光谱曲线,计算得到八种植被指数值,讨论不同面积比例,不同偏振角度下植被-土壤混合像元的高光谱偏振特性。研究发现,随着叶片占混合像元面积比例的增大,植被-土壤光谱曲线越来越明显地表现出植被光谱"五谷四峰"的特性,且峰值与谷底的位置与植被光谱基本相同。偏振角越大,混合像元的光谱偏振反射比越大;混合像元条件下,植被所占混合像元的面积比例越大,光谱受偏振角的影响越大。各植被指数与混合像元中植被面积大小呈线性关系,其中植被衰减指数和改进红边归一化植被指数的相关系数最大,可以达到98%左右,适合用于建立植被指数与植被占混合像元面积比例之间的相关模型。在植被面积发生变化时,改进红边比值植被指数的灵敏性更好。在利用光谱吸收特征参数进行植被指数估算时,发现吸收谷深度与光化学植被指数的二次函数模型拟合度最强,决定系数R2为0.963 3;光谱吸收指数与光化学植被指数的二次函数模型拟合度最强,决定系数R2为0.960 5。
[Abstract]:Hyperspectral remote sensing is increasingly used to determine the composition and proportion of ground objects in mixed pixels. The polarization reflectance spectra of vegetation-soil mixed pixel with different area ratio were obtained by using polarization device and ASD FieldSpec-3 spectrometer, and eight vegetation index values were calculated. The hyperspectral polarization characteristics of vegetation-soil mixed pixels with different area ratios and different polarization angles are discussed. It is found that with the increase of the proportion of the leaf to the mixed pixel area, the vegetation-soil spectral curve shows the characteristic of "four peaks" of the vegetation spectrum more and more obviously, and the position of the peak value and the bottom of the valley are basically the same as the vegetation spectrum. The larger the polarization angle, the greater the spectral polarization reflectance of mixed pixel, and the greater the area proportion of mixed pixel is, the greater the spectrum is affected by polarization angle under mixed pixel condition. There is a linear relationship between the vegetation index and the vegetation area in the mixed pixel. The correlation coefficient between the vegetation attenuation index and the improved normalized vegetation index of the red edge is the largest, which can reach about 98%. It is suitable for establishing the correlation model between vegetation index and the proportion of vegetation to mixed pixel area. When the vegetation area changes, the sensitivity of improving the red edge ratio vegetation index is better. When the spectral absorption characteristic parameters are used to estimate the vegetation index, it is found that the quadratic function model of the depth of absorption valley and the photochemical vegetation index has the strongest fitting degree. The quadratic function model of spectral absorption index and photochemical vegetation index is the best, and the determining coefficient R2 is 0.960 5.
【作者单位】：东北师范大学地理科学学院;
【基金】：国家自然科学基金项目(41301364) 吉林省教育厅“十三五”科学研究规划项目,中央高校基本科研业务费项目(2412016) 高等学校博士学科点专项科研基金项目(20120043120013) 吉林省科技发展计划项目(20150520067JH)资助
【分类号】：S154.4;TP79

【相似文献】