科尔沁沙地不同演替阶段差巴嘎蒿群落高光谱特征
邬亚娟1, 刘廷玺1,2, 童 新1,2, 罗艳云1,2, 段利民1,2, 曹文梅1

(1.内蒙古农业大学 水利与土木建筑工程学院, 呼和浩特 010018; 2.内蒙古农业大学 内蒙古自治区水资源保护与利用重点实验室, 呼和浩特 010018)

差巴嘎蒿; 群落演替; 连续统去除; 一阶导数; 马氏距离

Hyperspectral Characteristics of Artemisia halodeudrou Communities in Different Succession Stages in Horqin Sandy Land
WU Yajuan1, LIU Tingxi1,2, TONG Xin1,2, LUO Yanyun1,2, DUAN Limin1,2, CAO Wenmei1

(1.College of Water Conservancy and Civil Engineering, Inner Mongolia Agricultural University, Hohhot 010018, China; 2.Key Laboratory of Water Resources Protection and Utilization, Inner Mongolia Agricultural University, Hohhot 010018, China)

Artemisia halodeudrou; community succession; continuum removal; first derivation; Mahalanobis distance

备注

为了研究干旱半干旱地区植被资源的可持续利用,以科尔沁沙地建群种差巴嘎蒿为主要研究对象,将群落演替划分为4个阶段,从“同物异谱”特性着手,利用连续统去除和一阶微分对原始光谱进行变换处理,使用马氏距离法确定了冠层原始光谱和变换光谱的差异显著波段。结果 表明:同一时期的差巴嘎蒿群落随着演替进行,红边反射率持续上升,9月群落将进入凋萎期,总体曲线较5月降低; 对比不同演替阶段的差巴嘎蒿群落在导数变换以及连续统去除后的光谱特征,群落的一阶导数光谱在峰值、峰形、吸收峰位置存在显著差异,可以利用红边特征以及吸收深度进行识别,且9月特征识别效果好于5月; 马氏距离值能够科学快速地提取群落间差异较显著的特征波段,同时结合光谱变换达到了很好的降维效果,且提高了识别精度。其中9月份一阶导数变换降维效果最佳,且识别效果最优。研究利用一阶导数法与连续统去除法弱化了土壤背景的影响,为增强反射率特征及特征波段的差异性提供了较好的方法,同时马氏距离值能够反映不同光谱变换下差异显著的波段位置,有助于干旱半干旱地区植被识别研究。

As one of the main vegetation community species in arid and semi-arid areas, the rapid and accurate identification of Artemisia halodeudrou is beneficial to the effective utilization of vegetation resources in alpine areas. In order to study the sustainable utilization of vegetation resources in arid and semi-arid areas, the main objective of this study is to study the species diversity of Artemisia halodeudrou in Horqin Sandy Land. The community succession of Artemisia halodeudrou succession was divided into four stages and the characteristics of ‘same subject with different spectra' were studied. The method of continuum removal and first derivative reflectance were used to deal with the original spectral data. The Mahalanobis distance method was used to determine the significant difference between the original spectrum and the transform spectrum. The results showed that: at the same time, with the succession of Artemisia communities, the reflectivity of red edge continued to rise; in September, the community entered the wilting period, and the overall curve was lower than that in May; compared the spectral characteristics of Artemisia communities at different succession stages after first derivative transformation and continuum removal, the first derivative spectrum of the community showed significant differences in peak, peak shape and absorption peak positions; red edge and absorption depth could be used for recognition, and the best time for recognition was September; Mahalanobis distance could scientifically and quickly extract characteristic bands with significant differences between communities; the dimensional reduction effect was achieved by spectral transformation, and the recognition accuracy was improved; among them, the dimension reduction effect of the first derivative transformation in September was the best, and the identification effect was the best. The first derivative method and continuum removal method were used to weaken the influence of soil background, and provided the better method to enhance the reflectance characteristics and the difference of characteristic bands; at the same time, Mahalanobis distance range could express the positions of bands with significant differences under different spectral transforms, which was helpful to study vegetation identification in arid and semi-arid areas.