Re:Weird pupils let octopuses see their colorful gardens
by kirykl on Thursday December 01, 2016 @07:18PM
(#13084912)
Maybe possible to use this technique on space probes to capture true colors in a single channel, saving on data transfer
[ Parent
| on HN ]
Re:Weird pupils let octopuses see their colorful gardens
by crazydoggers on Friday December 02, 2016 @01:05AM
(#13086432)
This was brilliant to me at first blush. After thinking about it, though, wouldn't the image have to be a higher resolution to compensate for the blurring of the chromatic aberration? When you process out the blur to extract the color information, you'll have less resolution left over. I'm guessing that would offset the gains on using a single channel. Ultimately for "full color", at a specific resolution, you need a specific number of bits of information, regardless if it's encoded in three color channels, or a single channel.
Re:Weird pupils let octopuses see their colorful gardens
by kirykl on Friday December 02, 2016 @01:14AM
(#13086480)
Yea just throwing it out there. I think you're right it would most likely require more data
Re:Weird pupils let octopuses see their colorful gardens
by taneq on Friday December 02, 2016 @02:10AM
(#13086747)
It might still be good for other reasons, though. Like if the colour filters used in space probes degrade over time (I don't know, do they?) then it would allow more consistent colour recording. Also it might be usable on very small probes that primarily take black and white images but can capture colour (at expense of resolution) if required.
Re:Weird pupils let octopuses see their colorful gardens
by crazydoggers on Friday December 02, 2016 @01:29AM
(#13086556)
Kudos for the idea ... getting people thinking about the possible engineering applications is awesome!
Re:Weird pupils let octopuses see their colorful gardens
by terravion on Friday December 02, 2016 @02:00AM
(#13086708)
Actually, the opposite, this is the same principle as a hyperspectral sensor. These line scan sensors (kind of like the pupil photos in the article) create "data cubes" (i.e. an 2D image with a 3rd spectral dimension for every spatial pixel) which are huge compared to red, green, & blue images. The article actually talks about how computationally intensive it is to handle and how that made squid so smart.