Spatial statistics is traditionally concerned with data collected in space. As one field of application that has developed in the recent past, we consider satellite images. Starting with using such images as collateral information for supporting the handling and interpretation of spatial data, we later explored the images themselves, like for predicting pixel values below the clouds. The issue of up- and downscaling was addressed, as well as the integration of images, which was followed by including images into the modeling of spatial phenomena. At present, there is more and more attention to time series of images, with new challenges and opportunities emerging. Both optical and radar images require a solid understanding of the quality of the data and a fine-tuning to properly relate successive images. In this presentation, I will present some of the recent developments, like those developed in Zhang et al. [9], Mohammadi et al. [4], and Kulshrestha et al. [3]. Attention is given to issues of deep learning in the context of modern AI.