基于时序植被指数数据的地表覆盖聚类分析研究

摘要

Reference Land-cover classification is important to the land use and land management .As the land-cov-er is complex because of it is changing with the temperature and season ,it is difficult to land-cover classification .There-fore,this paper proposed a new method to clustering land-cover based on time series NDVI data.Through analyzing the different time series NDVI data and clustering the NDVI data ,we can achieve a well result about land-cover classification The result of the experiment proves this method has a high accuracy over 89 .76%that is better than traditional method .%地表覆盖分类是土地利用和土地资源管理的基础，由于地表覆盖情况十分复杂，且受温度、季节等影响较大，难以做到准确分类。为此，本文提取一种基于时序植被指数的地表覆盖聚类新方法，利用不同月份的影像，提取时序NDVI数据，并对其进行聚类分析，得到较好的地表覆盖分类结果。实验表明，该方法分类精度远可达89.76％，高于单幅影像分类结果。