摘要
针对全球范围内只能依据单季相Landsat影像数据提取耕地的现状,在现有基于影像光谱与纹理特征的监督分类方法基础上,引入NDVI时间序列数据,解决部分单季相影像造成的耕地与其他类型的混分问题。实验表明,利用NDVI时间序列数据能有效地反映耕地的物候特征,弥补使用单季相Landsat影像提取耕地的缺陷,较好地完善全球耕地的提取方案,提高了耕地提取精度。
According to the present situation of doing cropland classification using single-season Landsat image, this paper introduces NDVI time series data and resolves the confusion between cropland other land cover types in single-season Landsat images. Experiments show that this method could reflect the cropland＇ s phonological characteristics, complete the global cropland classification scheme and improve the accuracy of cropland classification.
出处
《地理信息世界》
2013年第2期66-69,88,共5页
Geomatics World
基金
国家"863计划"项目(2009AA12200)资助
