Remotely sensed data are typically sparse, which means that data have missing values, for instance due to cloud cover. This is problematic for applications and signal processing algorithms that require complete data sets. Hence, imputing the missing values is common and has been extensively discussed in the past, ranging from mechanistic filtering to classical geostatistical interpolation procedures. In the talk I will give some basic concepts of spatial and spatio-temporal prediction and present the new non-parametric gapfill algorithm. This latter method imputes each missing value separately based on data points in a spatio-temporal neighborhood around the missing data point. The imputation of the missing values and the estimation of the corresponding prediction uncertainties are based on sorting procedures and quantile regression. The procedure is implemented and available in the open-source R package gapfill. We illustrate the gapfill algorithm with MODIS NDVI data from Alaska with realistic cloud cover scenarios featuring up to 50% missing data. Validation against established software showed promising results.