I came across this article on EE-times. A quarter gigapixels on a chip smaller than an old 35mm negative is a fantastic accomplishment—but what does it really mean?
The array of 19,580 by 12,600 pixels measures 29.2mm by 20.2mm—a fraction over APS-H format; that gives it a pixel pitch of about 1.5µm. Those pixels are somewhat larger than for commodity mobile phone 8-16 Mpixel imager chips (1.2µm) but a lot smaller than for your typical consumer DSLR (about 3µm), let alone professional DSLR (over 5µm).
To illustrate the performance of the new chip it’s claimed it can “distinguish lettering on the side of an airplane flying at a distance of approximately 18km”. I would argue that you can do the same with your consumer DSLR if you put an extreme telephoto lens on it. Time for a small calculation: If we assume the plane lettering is about a metre high, in order to distinguish characters we would need to be able to resolve to about 20cm at worst. At about 20km, meaning we need an angular resolution of arctan(0.00001) or 2 arcseconds, which is within reach of a good amateur telescope. Have a look at this article by Nikon. I’ve found prices around 650$, so it’s not that expensive.
The problem is that in order to get enough light onto each 1.5µm pixel over the entire APS-H sensor size you’ll need a lens with a very large aperture, but in order to get high angular resolution you want a small aperture. You’ll need a very big, heavy and expensive lens to take full advantage of the quarter gigapixels!
But what if you could get rid of the lens? Could this chip be a great match for lensless imaging systems?