Hi! I am replying to this post a few months late...
I am the CEO and Founder of Amped Software, producer of the software Amped FIVE mentioned above.
In general, for video surveillance, it's way better to upgrade the camera than to use an video enhancement software.
How are our products used?
When a crime happens, law enforcement usually tries to acquire the footage from a multitude of cameras in the area. They can come from a single network or from various sources.
You don't have control on the quality of the system, you just have videos of various formats, quality and content.
What's the purpose of our software:
- convert videos from proprietary/native CCTV video formats with minimal loss of quality
- identification of useful footage
- identifying and correcting quality issues (interlacing, darkness, blur, low resolution, noise...)
- analyzing issues of the compression (frame type, macroblocks...)
- analyzing image content (for example people height or distances...)
- annotating frames, redaction, reporting...
As you see, enhancement is a small, but important part of the processing. But this is useful mostly in an investigative/forensic/post-mortem scenario.
If you check our examples you can see some pretty good results. Where's the catch?
First of all, the system, while easy, is not automatic.
Second, with our algorithms we are just showing better what's already there, attenuating the defects. We can not (and must not) add new data to the video. Sorry, it's not possible to get a readable license plate from 3 white pixels, nor never will be. The data is simply not there.
If you are planning a new system, plan for better cameras, not for post-processing (or at least not post-processing alone).
Post-processing may be extremely useful in specific situations, but in most of the cases cannot replace a better camera.