A Review on Applicability of Deep Learning for Remote Sensing Applications

Abstract

Remote sensing data are being utilized to understand the features of earth surface. Due to the extensive spatial and temporal coverage of satellites the information received from them have many advantages. However, the data received from satellite’s optical sensors are susceptible of errors because of the presence of clouds. Further, the data may be of low quality due to technical glitches and get a miss for some unavoidable human errors. Attempts were made in the literature to bridge the missing data. They may not, however, perform well with degraded observations of quality and are restricted by inadequate temporal data. This paper, presents the analysis and comparison of these methods, followed by Deep Learning technology and its applicability to remote sensing applications.