基于Triple-Collocation的地面观测与卫星遥感数据融合的雪深反演  

The Triple-Collocation-based Fusion of In-situ and Satellite Remote Sensing Data for Snow Depth Retrieval

作　　者：许剑辉[1] 舒红[1] 

机构地区：[1]武汉大学测绘遥感信息工程国家重点实验室

出　　处：《武汉大学学报（信息科学版）》2015年第4期469-473,486,共6页Geomatics and Information Science of Wuhan University

基　　金：国家自然科学基金资助项目(41171313;41331175);地理空间信息工程国家测绘地理信息局重点实验室开放研究基金资助项目(201329);湖北省自然科学基金资助项目(2014CFB725)~~

摘　　要：针对单一被动微波遥感反演雪深的精度和空间分辨率不足的问题,提出了一种星-地多源数据融合的雪深反演方法。以北疆每日站点观测雪深、AMSR-E遥感反演雪深和SSM/I遥感反演雪深数据为研究对象,首先利用地统计方法结合地面站点观测数据估计北疆区域的雪深,然后采用Triple-Collocation方法分别估计三个雪深数据的误差方差,最后结合最小二乘原理实现星-地雪深观测数据的融合。对融合雪深与地面雪深观测数据进行验证分析,结果显示,与AMSR-E和SSM/I遥感反演雪深相比,融合的雪深与地面观测雪深的相关性更高;融合的雪深的精度有一定程度的提高。实验结果证明了多源数据融合方法的有效性。Because of the insufficient accuracy and spatial resolution of snow depth products retrieved by passive microwave remote sensing,a new multi-sources data fusion approach is developed for retrieving snow depth.The data from different sources contains visible,passive microwave satellite and in-situ data.The daily in-situ,AMSR-E and SSM/I retrieved snow depth products are used in this study.First,combining in-situ snow depth,the snow depth of Northern Xinjiang is estimated through geostatistical analysis.Then the error variances of each product are calculated using a triple collocation（TC）method.Finally,the new snow depth products are obtained by merging in-situ,AMSR-E and SSM/I snow depth data in a least squares criterion where the optimal weights of each product are determined with the TC-based error variances.The merged snow depth is validated against in-situ snow depth and exhibits a higher correlation with in-situ observations than that with original AMSR-E and SSM/I snow depth.The results with higher accuracy demonstrate the effectiveness of our approach.

关 键 词：雪深 AMSR-E SSM/I Triple-Collocation 最小二乘原理 遥感反演 数据融合 

分 类 号：P237.9[天文地球—摄影测量与遥感]