摘要
鉴于MODIS积雪产品存在的空间不连续降低了数据的应用潜力,该文提出一种时空自适应加权的去云方法,同时顾及积雪在空间和时间上的相关性,自适应地衡量两者对积雪分布的影响。以新疆伊犁河流域为实验区,进行了模拟验证和基于气象台站数据的真实验证。结果表明:该方法能完全去除云覆盖,去云精度可达到90%以上,且对不同云量的修复都有较好的鲁棒性,能真实地反映地面积雪覆盖情况,可为积雪监测研究提供数据保障。
Due to the cloud obscuration, MODIS snow cover products are subjected to spatially discontinuous issues, which decrease the application potential. Therefore,this work proposed an adaptively spatio-temporal weighted method for alleviating the trouble. This method simultaneously took the spatial and temporal correlations of snow cover into consideration, and weighted their contributions to snow distribution automatically. Based on simulated experiments and daily snow depth data at climate stations,we evaluated its accuracy in Ill river basin of Xinjiang. The results indicated that our proposed method can remove cloud obscuration completely, with overall accuracy above 90% under different cloud coverages. Since the cloud free snow cover maps reflect the real distribution of snow cover,they can be widely used for monitoring snow.
出处
《遥感信息》
CSCD
北大核心
2016年第2期36-43,共8页
Remote Sensing Information
基金
863课题(2013AA12A301)
国家自然科学基金(41271376
41271358)
