Latest Trends on Satellite Image Segmentation

For the last few decades satellite imaging technology has taken massive strides towards higher spatial resolution, larger swath coverage and almost real-time data delivery. Satellite imaging or remote sensing is extensively used in optical imaging of the globe, military surveillance, deforestation detection, etc. Multispectral imaging by satellite has enabled an exceptional understanding of the earth by imaging beyond the realm of the visible spectrum. Against this backdrop image segmentation using machine learning, deep learning models and image processing has prompted several new approaches and techniques for satellite image segmentation. This survey provides a comprehensive review of the recent literature, covering the novel approaches in image segmentation using satellite imaging as well as others. There is a broad coverage of segmentation algorithms of UNET, TSVM and Random Walker. It investigates the strengths, challenges and novel aspects and compares precisions and deliberate potential research outlooks.