摘要
开展了时间序列Landsat TM/ETM遥感影像定量化处理与相对辐射校正,提取了陕西神木县不同地物光谱和NDVI物候特征,结合时间序列NDVI物候特征和多时相光谱信息,采用了地表覆盖的决策树分类算法,实现了陕西神木县地物的高精度遥感分类,包括水体、沙地、城镇、耕地、林地、草地及灌丛等7类地物,分类总体精度达95.77%,Kappa系数达0.93。研究结果表明,基于多时相光谱和物候特征的决策树分类算法能够有效集成多时相、多光谱信息,从而克服了单时相影像分类的缺陷,实现了地物的分类。论文研究方法和结果能够为三北防护林区域的生态环境监测与评估提供技术支持。
Multi-temporal and multi-spectral information is very important for land cover mapping. In this paper, there were five Landsat TM/ETM images of Shenmu city in shanxi province acquired and processed,and the spectral and NDVI phonologi-cal features of different land-covers were extracted from the Landsat imagery according to ground survey data. According to the spectral and phenological information, a decision tree was used to classify the land-covers, including water, sand, city, cropland, for-est, grassland and brush. The classification result was validated by the ground survey data,with an overall precision of 95.77%, and a Kappa coefficient of 0.93. The result shows that the land cover can be mapped by using the decision tree algorithm, which can integrate the spectral phonological information from the multi-temporal satellite imagery. The presented method can also be applied for the ecological environment monitoring in the Three-North Shelter region.
出处
《遥感信息》
CSCD
2013年第2期76-81,共6页
Remote Sensing Information
基金
国家973课题(2009CB723902)
中国科学院对外合作重点项目(GHJ21123)
关键词
多时相
多光谱
决策树
NDVI
分类
multi-temporal
multi-spectral
decision tree
NDVI
classification
