14 October 2014 Sea slicks classification by synthetic aperture radar
Author Affiliations +
Abstract
An automatic system called OSAD (Oil Spill Automatic Detector), able to discriminate oil spills (OS) from similar features (look-alikes – LA) in SAR images, was developed some years ago. Slick detection is based on a probabilistic method (tuned with a training dataset defined by an expert photointerpreter) evaluating radiometric and geometric characteristics of the areas of interest. OSAD also provides wind field by analyzing SAR images. With the aim to completely classify sea slicks, recently a new procedure has been added. Dark areas are identified on the image and the wind is computed inside and outside for every area: if outside wind value is less than a threshold of 2 m/s it is impossible to evaluate if damping is due to a slick. On the other hand, if outside wind is higher than the threshold and the difference between inside and outside the dark area is lower than 1 m/s we consider this reduction as wind fluctuation. Wind difference higher than 1 m/s is interpreted as damping effect due to a slick; therefore the remaining dark spots are split in OS and LA by OSAD. LA are then analyzed and separated in “biogenic” or “anthropogenic” slicks following an analogous procedure. The system performances has been tested on C-band SAR images, in particular on images having spatial resolution so high to examine details near the coastline; the obtained results confirm the efficiency of the algorithm in the classification of four types of signatures usually found on the sea surface.
© (2014) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
P. Trivero, P. Trivero, W. Biamino, W. Biamino, M. Borasi, M. Borasi, M. Cavagnero, M. Cavagnero, L. Di Matteo, L. Di Matteo, D. Loreggia, D. Loreggia, } "Sea slicks classification by synthetic aperture radar", Proc. SPIE 9240, Remote Sensing of the Ocean, Sea Ice, Coastal Waters, and Large Water Regions 2014, 92400C (14 October 2014); doi: 10.1117/12.2066158; https://doi.org/10.1117/12.2066158
PROCEEDINGS
9 PAGES


SHARE
RELATED CONTENT

Texture classification based on evolution feature selection
Proceedings of SPIE (September 24 2001)
Laser radar system for obstacle avoidance
Proceedings of SPIE (October 11 2005)
Identification of built up areas using SAR data a...
Proceedings of SPIE (October 10 2008)
An experimental evaluation of endmember generation algorithms
Proceedings of SPIE (November 04 2005)

Back to Top