With the continuous development of space and sensor technologies during the last 40 years, ocean remote sensing has entered into the big-data era with typical five-V (volume, variety, value, velocity and veracity) characteristics. Ocean remote-sensing data archives reach several tens of petabytes and massive satellite data are acquired worldwide daily. To precisely, efficiently and intelligently mine the useful information submerged in such ocean remote-sensing data sets is a big challenge. Deep learning—a powerful technology recently emerging in the machine-learning field—has demonstrated its more significant superiority over traditional physical- or statistical-based algorithms for image-information extraction in many industrial-field applications and starts to draw interest in ocean remote-sensing applications. In this review paper, we first systematically reviewed two deep-learning frameworks that carry out ocean remote-sensing-image classifications and then presented eight typical applications in ocean internal-wave/eddy/oil-spill/coastal-inundation/sea-ice/green-algae/ship/coral-reef mapping from different types of ocean remote-sensing imagery to show how effective these deep-learning frameworks are. Researchers can also readily modify these existing frameworks for information mining of other kinds of remote-sensing imagery.

We developed a deep learning model to forecast the sea surface temperature evolution associated with tropical instability waves.

In 2008, the Canadian Space Agency sponsored the Radarsat Hurricane Applications Project (RHAP), for researching new developments in the application of Radarsat-1 synthetic aperture radar (SAR) data and innovative mapping approaches to better understand the dynamics of tropical cyclone genesis, morphology, and movement. Although tropical cyclones can be detected by many remote sensors, SAR can yield high-resolution (subkilometer) and low-level storm information that cannot be seen below the clouds by other sensors. In addition to the wind field and tropical cyclone eye information, structures associated with atmospheric processes can also be detected by SAR. We have acquired 161 Radarsat-1 SAR images through RHAP between 2001 and 2007. Among these, 73 images show clear tropical cyclone eye structure. In addition, we also acquired 10 images from the European Space Agency's Envisat SAR between 2004 and 2010. Both Atlantic hurricanes and Pacific typhoons are included. In this study, we analyze these 83 (73 Radarsat-1 and 10 Envisat) images with tropical cyclone eye information along with ancillary tropical cyclone intensity information from the archive to generate tropical cyclone morphology statistics. Histograms of wave-number asymmetry and intensity are presented. The statistics show that when the storm has higher intensity, the tropical cyclone eye tends to become more symmetric, and the area of the tropical cyclone eye, defined by the minimum wind area, tends to be smaller. Examples of finescale structures within the tropical cyclone (i.e., eye/eyewall mesovortices, arc clouds, double eyewalls, and abnormally high wind or rain within eyes) are presented and discussed.