I know that all of us have watched detective series on tv. What kind of computer software do detectives use to blow up and reconstruct images. Can these programs be downloaded off the internet? Any links??
You mean where they zoom in to an image and it’s all blocky pixels and the detective says “can you enhance that”, the tech guy types a few lines of obscure code and the detail appears out of nowhere, then they zoom in again and it’s all blocky pixels and the detective says…
It’s fiction. There’s no way to restore lost detail to an image.
There are some clever techniques that can work on defocused images, turning circular features back into points, but there’s no way to get back what is actually gone.
OK all digital images are composed of dots.
As Mangetout pointed out - clever software uses educated guesses, but.
If you expand too far all you get is Big guess or Big dots.
However - this is only with current video metgods, who knows what tomoz holds.
All that is said so far holds true for still images.
I’d like to know what you can do with a piece of video, though. I’d guess that by averaging over several frames, you’d be able to pick out some more detail, just like you do with a digital oscilloscope, where you can take a very noisy signal, get a good trigger, and average out a lot of the noise.
Of course, this assumes that the object isn’t moving in relation to the camera[sup]*[/sup], and that the noise has a reasonable distribution.
Does any such software exist?
[sup]*[/sup] Although, if the camera moved in a very controlled way, this could be used to increase the resolution! Have the camera move in steps of half a pixel, up, right, down and left, and then treat the corresponding images separately , before merging them back together as an image of twice the resolution!
Yes, the software exists and, as Popup says, it ises the movement over several frames. I remember a courtcase where the use was contested. The video showed a car used by someone who had held up a convenience store. In the video the tags were unreadable but after some processing the plates were readable. The defense argued to have the evidence thrown out saying the processing was putting information which was not there initially. The guy who did the work explained the information was already there spread over several frames.
This same idea occurred to me a while ago. Seems to me like it would be a good way to get very high resolution images, either of static microscopic subjects, or of the earth (taken from a geostationary satellite). But does it actually work? I fiddled around with the idea in my brain for a while, and couldn’t figure out a way to combine the multiple, slightly different images of one resolution into a single image of much higher resolution. I’d be very interested to know if a system like this already exists somewhere, and how it’s done.
I’ve acutally done some image enhancement for a local police deparment. Used photoshop, but results not too helpful. Even the cop thought it would be like TV fiction.
Apart form photoshop, anybody know any good program where one can be downloaded