Hacker News Reader

Comment by jvanderbot | Hacker News Reader

Technical SAR question: Is the sea surface removed post-hoc, or does the sea not reflect these radar waves in a sufficient amount to show up? I seem to remember NISAR [1] will map land and ice - so presumably water ice shows up.

I ask b/c if they are synthetically removing the sea surface at 1-5m resolution that seems really hard given tides, waves, etc.

1. https://en.wikipedia.org/wiki/NISAR_(satellite)

View on HN

Calm water surface is a specular scatterer, meaning radar energy will be reflected away, and proportionally so as the incidence angle increases.

Images collected at lower incidence (closer to nadir) might feature very bright surface water if it is calm, as more energy is reflected towards the radar.

Rough water surface is a diffuse scatterer, and will generally appear brighter than calm waters.

Any good source for SAR image formation? From antennas to image processing?

I am just a half-educated layman when it comes to SAR, which has a very heavy electrical engineering heritage. I hear that one very good technical resource is this book written by Iain Woodhouse [1], but I would be lucky to understand a tenth of it myself.

There is also a very good SAR vulgarisation book written by Tom Ager [2].

[1] https://www.routledge.com/Introduction-to-Microwave-Remote-S...

[2] https://www.amazon.com/Essentials-SAR-Conceptual-Remarkable-...

thank you very much

The liquid water absorbs most of the radar wave (at least for the C band, as is the case of sentinel-1). Thus, in the images it appears natively "black". When there are a lot of waves, the surface of the water forms some spurious reflectors that appear as a light texture in the surface, but the signal is definitely less powerful than metal/concrete reflectors of buildings and ships. Notice that in the "Humber" image of TFA you can appreciate some texture in the water. The contrast in this image has been exaggerated a lot, which saturates most of the land.

It actually does not absorb radar energy, rather it reflects it away (specular reflection). That's why rougher sea surface appears brighter on SAR as it turns from a specular to a diffuse scatterer. Furthermore, in some specific conditions (low incidence angle, closer to nadir), calm water can appear much brighter than other land surfaces.

thanks for the clarification! I'm just used to looking to the images without thinking too much about the water.

I've never seen "close to nadir" radar images... wouldn't the ground fold over itself?

Nadir-SAR would look very confusing, full of ambiguities, and probably unexploitable since the ranging part of RADAR would not be able to distinguish the left returns from the right returns -- that's why SAR is side-looking.

even with a strictly side-looking geometry, images collected closer to nadir (steep/low incidence) start exhibiting some artifacts such as nadir returns [1].

[1] https://www.researchgate.net/figure/Example-of-nadir-echo-in...

They're composite images, so there's definitely some post-processing going on. The sea surface will show up darker, but SAR is sufficiently sensitive to to detect oil slicks.

Reply on news.ycombinator.com