Leveraging Potential of Deep Learning for Remote Sensing Data: A Review

Remote sensing has witnessed impressive progress of computer vision and state of art deep learning methods on satellite imagery analysis. Image classification, semantic segmentation and object detection are the major computer vision tasks for remote sensing satellite image analysis. Most of work in literature is concentrated on utilization of optical satellite data for the aforementioned tasks. There remains a lot of potential in usage of Synthetic Aperture Radar (SAR) data and its fusion with optical data which is still at its nascent stage. This paper reviews, state of the art deep learning methods, recent research progress in Deep learning applied to remote sensing satellite image analysis, related comparative analysis, benchmark datasets and evaluation criteria. This paper provides in depth review of satellite image analysis with the cutting edge technologies and promising research directions to the budding researchers in the field of remote sensing and deep learning.