Nonuniform sampling and reconstruction for high resolution satellite images

The objective of nonuniform sampling methods in image processing is to represent an image with a low number of samples while preserving its content. This paper proposes an adaptive nonuniform sampling algorithm for high-resolution satellite images. The idea is that more samples are selected around the edges thanks to the measure of the local skewness, and that few samples are placed in regions containing low-frequency features. The Delaunay triangulation method has been developed to reconstruct the image from nonuniform samples. The results of the reconstruction show that the proposed algorithm can obtain a similar image quality with fewer samples, which is valuable for high-resolution satellite image transmission and storage system.