中文摘要：

针对植被指数和植被覆盖度与遥感影像空间尺度的关系对高分辨率影像在植被分析中的运用具有重要影响的问题,该文提出一种确定最佳分辨率的方法。以云南鲁甸县茨院乡为研究区,先对原始影像（Worldview-2）进行重采样处理,生成一系列低分辨率影像;计算13组数据的NDVI并利用像元二分模型反演出植被覆盖度;运用信息熵理论和按照局部方差的思路等两种方法,定量分析该区植被分析的最佳分辨率。实验结果表明,随着空间分辨率的降低,NDVI值域区间呈现收敛趋势;植被覆盖度分级图斑斑块数量急剧减少,而所占面积比则保持相对平稳;信息熵中的信息量也随之减少。分析结果认为,该区进行基于Worldview-2影像植被分析的最佳空间分辨率为12m。

英文摘要：

Research on the relationship among vegetation index,vegetation coverage and remote sensing image resolution is significant to the rational use of high-resolution image in vegetation analysis.In the light of the relationship among the three which are mentioned above,a new way is proposed to determine the best resolution in this paper.The paper selects Ciyuan village,Ludian County of Yunnan Province as the study region.The original image,Worldview-2,with fine resolution is degraded to successively coarser resolutions by resampling to calculate the Normalized Difference Vegetation Index（NDVI）and vegetation coverage of 13 groups data respectively.Based on the methods described above,referencing the two methods：the theory of entropy and the thought of local variance,this paper runs the quantitative analysis of the best fitting Worldview-2image resolution in this study region.According to the research,the paper comes to the following conclusions.Firstly,as the resolution becoming coarser,the statistics interval of NDVI presents a narrowing trend,numbers of map patches in each classification of vegetation coverage decrease rapidly,area proportion of each classification remains stable and information entropy contents come reducing.Secondly,also the most important conclusion of the quantitative analysis in this paper is the best fitting Worldview-2image resolution of vegetation analysis in the study area is 12 m.