Stripe segmentation of oceanic internal waves in SAR images based on SegNet

Abstract

The development of ocean remote sensing makes it possible to obtain valuable information from a large amount of data. Deep learning is a powerful tool that is beneficial for obtaining ocean information from remote sensing data. Oceanic internal waves play an essential role in ocean activities. To obtain information on irregular stripes from Synthetic Aperture Radar (SAR) images, a stripe segmentation algorithm for oceanic internal waves is proposed based on SegNet. The research results show that the proposed method can identify whether the SAR images contain oceanic internal waves and obtain the respective locations of the light and dark stripes of the oceanic internal waves from SAR images. Furthermore, because this method can accurately determine the relative locations of the light and dark stripes, it can distinguish the moment when oceanic internal waves undergo polarity conversion.

Keywords:

• Deep learning
• oceanic internal waves
• SAR
• stripe segmentation
• SegNet

Acknowledgements

The authors are grateful for the websites of Envisat, Sentinel, and ASF which are used to collect global SAR images, and the support of SNAP, QGIS, OpenCV, and Python.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Natural Science Foundation of China (grant numbers 51679132), and the Science and Technology Commission of Shanghai Municipality (grant numbers 21ZR1427000 and 17040501600).