Satellite Image Processing Systems: An Architectural Perspective

Abstract

Satellite image processing systems include satellite image classification, long ranged data processing, yield prediction systems, etc. All of these systems require a large quantity of images for effective processing, and thus they are directed towards big-data applications. All these applications require a series of highly complex image processing and signal processing steps, which include but are not limited to image acquisition, image pre-processing, segmentation, feature extraction & selection, classification and post processing. Numerous researchers globally have proposed a large variety of algorithms, protocols and techniques in order to effectively process satellite images. This makes it very difficult for any satellite image system designer to develop a highly effective and application-oriented processing system. In this paper, we aim to categorize these large number of researches w.r.t. their effectiveness and further perform statistical analysis on the same. This study will assist researchers in selecting the best and most optimally performing algorithmic combinations in order to design a highly accurate satellite image processing system.