1

I have a rather general question about supervised classification for oil spill detection. I want to apply supervised classification on the ESA SNAP dark spot detection tool results. My question is: if I create spectral signatures from an Envisat image (containing verified oil spills) will they perform well on a Sentinel - 1 image? As a pre-processing step I would apply radiometric correction (calibration) on both images. Any suggestion is welcome. Efi

  • Probably, SAR processing tends to be non-trivial, see: sciencedirect.com/science/article/pii/S2214241X15002357. Are you planning on using polarimetry or just amplitude? I don't know enough about how oil affects backscatter, or how wavelength differences between different sensors behave, however if you have verified oil spills, then you ought to be able to figure this out. – John Powell Jun 26 '18 at 12:17
  • Hello John, I will use amplitude only. Both sensors operate at C-band. Thanks for replying. Efi – user123038 Jun 26 '18 at 12:35
  • I would recommend using Sentinel-1 VV pole if it is available for your study area. VV polarization tends to be more effective in detecting oil spills on water. – AWGIS Jul 27 '18 at 14:59
1

A critical point to keep in mind when looking at oil spills on water is wind speed. Oil has the effect of damping the amplitude of waves. Back in the day, when I used to process SAR imagery for both natural and anthropogenic-sourced oil on water - we would always get wind speeds for the area. There are "ranges" of wind speeds most appropriate for various volumes of oil on water to maximize the radar return difference between water and oil on water.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.