note that i didn’t attempt to do super resolution, i just merged the images for noise reduction. when i compare against the image on the webpage you linked this seems to actually work.
not sure there’s much to gain from super res here, the lens looks incredibly blurry. has this plastic smartphone feel to it. i did apply a round of deconvolution (“capture sharpening” like), this helped a lot.
most of the highlights are blown out and, well, the colours are off, but you get the idea
hm sorry i know nothing about telephones, and usually don’t accept their sensors as “camera”. wasn’t aware this one is even cheap and comes with linux, which i am way more impressed by than by the image quality
i think the border comes from my boundary handling in the deconvolution (or lack thereof). possible that it interpolates leftovers from black border cropping. need to double check, but i’m still not too worried about these “final touches”.
how have i merged them: there’s a button for that in the gui, if that’s what you mean? it uses some custom alignment code (takes 9.129ms to align these 5 relatively lores shots).
how does one program this phone? does it have a GPU?
how have i merged them: there’s a button for that in the gui, if that’s what you mean? it uses some custom alignment code (takes 9.129ms to align these 5 relatively lores shots).
That’s what I wanted to know. (some custom alignment code)
how does one program this phone?
I don’t have one but afaik you can simply connect it to a USB-C docking bar, add a keyboard, mouse and monitor and start hacking. I just found this video:
does it have a GPU?
Yes, a Allwinner A64 Quad Core SoC with Mali 400 MP2 GPU
nice, thanks. the GPU is a 2008 build, and it does support opengl es 2.0. i haven’t done much oldschool opengl, but i think this one even doesn’t have compute shaders (afaik that’d be 3.0), so it’ll be, let’s say, tricky to implement anything on it.