Nanowerk: Olympus and Tokyo Institute of Technology RGB-IR solution presented at EI 2016 claims to be able effectively separate IR and RGB images taken by a mixed CFA sensor:
What is the pixel size? Does it exist color resists that can correctly filter NIR? Or narrower around 850nm or 950nm? Does is exist color resist for R, G or B that is can filter NIR?
Well, I don't know how Olympus does this but you could use the "N" NIR channel signal to subtract NIR from R,G, and B channel signals. This is what we did with the RGBZ sensor at Samsung (reported at ISSCC a few years ago). Not so good for noise, but effective at removing NIR fog.
The NIR subtraction method works only if the NIR channel isn't saturated which is often the case when there's a lot if IR illumination (e.g. surveillance). This proposed technology (and similar) is often used in low-cost CCTV cameras to eliminate the need for an expensive mechanical IRCF switcher.
What is the pixel size?
ReplyDeleteDoes it exist color resists that can correctly filter NIR? Or narrower around 850nm or 950nm?
Does is exist color resist for R, G or B that is can filter NIR?
(ok we need a color filter specialist here...)
Well, I don't know how Olympus does this but you could use the "N" NIR channel signal to subtract NIR from R,G, and B channel signals. This is what we did with the RGBZ sensor at Samsung (reported at ISSCC a few years ago). Not so good for noise, but effective at removing NIR fog.
DeleteThe NIR subtraction method works only if the NIR channel isn't saturated which is often the case when there's a lot if IR illumination (e.g. surveillance). This proposed technology (and similar) is often used in low-cost CCTV cameras to eliminate the need for an expensive mechanical IRCF switcher.
Deletedakloifarwa
Eric,
ReplyDeleteHow strong was the NIR signal compared to RGB?