Hello, I'm new in this forum, maybe my question had been answered before. Is there a good utility or procedure to enhance and/or add more detail to a surveillance video? The problem is that its already compressed in DVD, and the image is awful. I'm trying to clearly identify someone in the images, but is very grainy and dark. I've tried photoshop filters, premiere resources, and other utilities, but the source is very bad. Any insight will be much appreciated. Thank you.
+ Reply to Thread
Results 1 to 8 of 8
-
-
Well you could adjust the gamma and contrast to make it a little brighter, but it sounds like you already tried that. Criminal investigators use temporal filters to clean up survellance footage. Basically, all 6 frames showing a guys face may be blurry and out of focus, but each frame contains bits and pieces that are better than another so you can composite them into a new frame that is higher quality than any of your existing ones. This software is very expensive and probably not sold to ordinary consumers but I bet you can roughly approximate its function in photoshop.
I'd adjust the gamma and such as best you can to the entire footage, then extract a series of frames for a particular feature you want to advance, like his t-shirt or his eyes. Put all the frames of this sequence as separate layers on top of one another in photoshop and line them up as best as possible. Then just play with the layers' blending properties. Like try setting them to lighten or overlay and play with the opacity. It can easily become a mess with so many layers blended together, but some features may come into focus better. -
But on CSI I saw them program a new filter in just a few keystrokes and 20 seconds later they could clearly see the reflection of the killer's face in the womans diamond ring. Surely I can do this with virtualdub ?
Read my blog here.
-
ROFFLE WAFFLES!
What you see on TV is... wait for it... NOT REAL. LOL. -
True Grum, but I was doing some programming work (ok back in the day). Got to know a guy who was doing work with/for some government agency (he was always deliberately vague) at anyrate it was essentially image enhacement denoise stuff for satellite images. Obviously not up to Mission Impossible or James Bond standards, but still some awfully impressive stuff. Things would absolutely leap into focus and details that you would swear couldn't possibly be recovered would appear. As noted it took the information from a series of pictures and constructed a single composite image from that information.
Who knows how it might look on moving images and if anything even vaguely like that would ever work it's way out into mainstream (and wouldn't require a rendering farm to work), but something even half as good as I saw 10 years ago would make you forget tweaking the gamma or MSharpen forever. -
I frequently do enhancements for surveillance video. For VHS tapes I receive, the biggest problem is the quality of the image. Most convenience stores reuse video tapes for over a year and rotate between two tapes. The tape itself has limited magnetic properties remaining. In any case, I dub the tape to DV. The reason for this is that I can get the most clear single frame images this way. If I capture the video directly into the computer (I typically use in-sync Speed Razor, but have also used Adobe Premiere Pro), the single frame images are distorted. If I dub to DV, the single frame is rock steady. I then capture the single frame to the computer. If the image is a split screen, I will use my editing program to crop and expand the screen I am interested in. I will export this frame into Photoshop and play with it there. Contrast, brightness and gamma are all things to play with. If the image is color, try changing it to B&W. You can see a lot more detail without the effects of color bleeding. If you are looking for a face, try using a soft blur rather than the sharpen tool in Photoshop. When you use the sharpen tool on an image that has already been blown up, you increase the pixel effect and edges actually become jagged. If you use the blur, you smooth out the edges. The human eye can recognize a face from a pisture that is slightly blurred with smooth edges more easily than it can recognize an image with jagged edges. Also, keep in mind that even though you might not be able to make a positive identification from the image produced, someone that knows the person often can. I had one person who was breaking into pharmacies at night. The image was dark and totally unrecognizable. I turned it into B&W, lightened and increased the contrast. I was able to get the picture to where you could tell that the suspect was a white male of thin build. That was about it. When a suspect was caught from another burglary, he denied doing the previous burglaries. The detective tossed the photograph I had enhanced towards the suspect and asked, "What about this?" The suspect looked at it and confessed.
Basically, TV is just that. I have seen what NASA can do and it's amazing, but still not quite like TV. You do the best you can with what you've got. Often it's enough. If not, well, the bad guys eventually get caught anyway.
Similar Threads
-
surveillance h264 video problem
By rix133 in forum Video ConversionReplies: 6Last Post: 19th Dec 2011, 17:39 -
Video Enhancement
By rckowal in forum RestorationReplies: 20Last Post: 13th Dec 2011, 13:54 -
Video Surveillance, DIY or leave to professionals?
By lildeee in forum Newbie / General discussionsReplies: 0Last Post: 29th Jul 2009, 11:02 -
Video enhancement software
By terryxpress in forum MacReplies: 4Last Post: 30th Aug 2007, 11:37 -
Need help choosing Video Surveillance System and software
By mowermanEd in forum Newbie / General discussionsReplies: 3Last Post: 16th Aug 2007, 16:03