基于NDVI的西昌市紫茎泽兰入侵监测模型初探

目的]为紫茎泽兰生物入侵的遥感监测提供参考。方法]以ASTER影像数据为主信息源,结合西昌市紫茎泽兰样地调查,提取3种植被指数分别为归一化植被指数NDVI、垂直植被指数PVI和比值植被指数RVI,通过与紫茎泽兰表征因子盖度、高度和生物量进行相关分析,筛选敏感植被指数。在紫茎泽兰单个表征因子回归分析基础上,建立敏感植被指数与重要值的回归模型。结果]3种植被指数中,NDVI最能反映紫茎泽兰表征因子的变化信息。NDVI与紫茎泽兰表征因子及IV显著正相关,对IV的复相关系数(R2=0.72)大于单个表征因子。基于NDVI与紫茎泽兰IV的线性模型模拟误差较小。结论]利用NDVI与IV的线性模型对研究区紫茎泽兰入侵进行监测是较为可行的。

A Primary Study on Modeling for Eupatorium adenophorum Spreng. Invasion Monitoring in Xichang City Based on NDVI

Objective] The study aimed to provide references for monitoring biological invasion of Eupatorium adenophorum Spreng.by means of remote-sensing.Method] On the base of ASTER images,coupling with sample plots investigation of E.adenophorum Spreng.in Xichang city,Normalized Difference Vegetation Index(NDVI),Ratio Vegetation Index(RVI) and Perpendicular Vegetation Index(PVI) were extracted,and correlation analyzing was conducted between vegetation indices and three community indices(including coverage,height,and biomass) to screen out sensitive vegetation index.After regression analyzing between the sensitive vegetation index and single community index,a regression model was established based on the sensitive vegetation index and important value(IV) of E.adenophorum Spreng.Result] NDVI was the most sensitive among 3 vegetation indices.The correlations between NDVI and community indices were highly significant,and the correlation between NDVI and IV too.But the multiple correlation coefficient(R2=0.72) of NDVI-IV regression model was higher than that of others.The error between the observation value and the simulation value based on the NDVI-IV regression model was low.Conclusion] Monitoring biological invasion of E.adenophorum Spreng.in research area with NDVI-IV regression model was feasible.