摘要:
遥感反演土壤含水量是干旱遥感监测中必不可少的环节,但遥感传感器往往受到云雪或自身性能等因素影响而存在数据缺失.现有基于时间序列数据的滤波插补法对数据要求较高而难以推广,而空间插值方法在成块缺失区域效果较差.针对此问题,提出利用最优插值法,根据气象站点实测数据对插值点的贡献定权,进而实现缺失像元的插值与填充.选取宁夏回族自治区为实验区,应用植被条件反照率干旱指数(vegetation condition albdedo drought index,VCADI)反演多个时期的土壤含水量结果,结合分布在研究区的16个国家级气象台站观测数据,利用最优插值法对缺失数据进行插补,结果表明该方法对不同程度数据缺失都具有较好的效果.通过人工模拟具有成块缺失与不同缺失率的数据,对比反向距离加权插值、Kriging空间插值方法和本文方法的插补效果,验证了本文方法具有更高的插补精度.
Abstract:
Remote sensing - based soil moisture content inversion is an indispensable procedure in drought monitoring;however, the image acquisition process is often influenced by bad weather such as cloud cover and snowfall,or sensor performance defects, which causes missing data. The existing filtering interpolation methods based on time series images have a high requirement on input data and thus are difficult to be widely applied,while the spatial interpolation methods do not work well for the images with missing blocks. In view of the above problems,this paper proposes a missing data filling method based on optimum interpolation, which predicts and fills missing data with the ground observation data as a reference. The authors selected Ningxia as the study area and obtained the soil moisture content in multiple periods using the VCADI index, and conducted missing pixel interpolation using the proposed method with the ground observation data of 16 national meterological stations. Experimental results show that the proposed method performs well in all regions with different levels of missing data. The authors simulated the images with missing blocks and different levels of missing data,and compared the performances between the inverse distance weighted interpolation method,the Kriging interpolation method and the optimum interpolation method. Experimental results show that the method proposed by the authors can obtain more accurate interpolation results.
下载: