As footage of different events continues to come about, I wanted to raise some awareness about an enhancement technique known as multi-frame super resolution (or super-resolution imaging), it's existed as a possible forensic technique for a number of years that can be applied to CCTV footage of lesser quality, but these days it is also used for astrophotography, microscopy, and better digital zoom.

A limitation of the technique is it is only useful for multiple frames that deviate only slightly in movement, so its uses are more limited.

What matter for us is the various implementations in the wild on places such as GitHub that might be used and if there is existing footage out there that ot would be worth applying it to.

I have seen a few implementations in my searching, but the more promising looking ones require MatLab which I don't have currently, so as such I haven't been able to try it myself. Some implementations are also only meant for raw imagery as the Bayer filter data (Which is needed for the most optimal results) is preserved, for forensic use these aren't particularly useful as such data won't exist in compressed images or video frames.

So far the most promising looking of these I saw is at the link below, but certainly more research is warranted.

https://github.com/rafaelmaeuer/MultiFrameSuperResolution

Edit: Made a mistake and found out this implementation above just needs the MatLab runtime (which is apparently freely available) as mentioned in it's documentation. I have started to fiddled with it a little, but now looking for some suitable footage to use.