Shadow Detection and Segmentation on Satellite Images: a Survey

Abstract

Shadow detection and segmentation are widely used in many computer vision and image processing applications. Shadows on various types of images can provide both positive and negative traits so a researcher can retrieve some useful information or, on the contrary, must get rid of or mitigate some predicaments. In satellite imagery, the problem of shadow detection is of special importance as far as shadows can give useful insights into objects, landscapes, and dynamics of a captured scene, as well as pose some obscurity about objects of a researcher’s interest. This survey paper provides a comprehensive exploration of the state-of-the-art techniques and methodologies in the domain of shadow detection and segmentation within satellite imagery. We give descriptions and analysis for ten method and algorithm categories. We also compare them based on the selected aspects: accuracy, complexity, robustness, ability to work with different types of images, and data processing requirements