Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Strange Kind of Noise in Captured Video Image

Status
Not open for further replies.

IndiJones

Full Member level 2
Joined
Aug 13, 2001
Messages
132
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,296
Activity points
786
We have built a prototype of a 3CCD digital video camera. When we take live images and view them on an LCD monitor, strange kind of noise signals appear on the image. These are scattered red spots having the size of 1 pixel and are always moving. They appear only on green colored objects. They can be seen on the edges of the object. If you hold a green colored cloth, noise spots can be seen on the borders and in places where the cloth surface is not even. For instance, if you cover the tip of your index finger with the cloth, it'll have a shape like a mountain and there will be some bending on the cloth, thereby creating some edges. Noise can be observed on these edges too. Please note that the image that I'm talking about is the final image that doesn't have any other noises. Noise correction has already been done. These noises don't appear on objects having colors other than green! Increasing the gain of the red channel makes them disappear. However, doing the same for green and blue channel increases the noise.

Does anyone have any such experiences? It would be very helpful if someone could advise me on how to get rid of this thing.

Thanks in advance!
 

Did you design your-self the camera PCB ?
May it be due to crosstalk of a green transition on the Red analog channel ?

in digital domain, I had already seen such strange spots due to a short between an MSB & LSB on different channel.....

... and you're sure that 's not due to the display ?....
 

We designed the optical side. The PCB was developed by the manufacturer of the camera. They are also looking into this matter. The display don't have any problem cause we have used various displays and the noise appears in all of them. Could you plz explain a little bit about the short that you've seen between MSB & LSB on different channels.

By the way, do you know of any books on digital camera design that explains everything starting from the filter design to the hardwere design.

Thanks!
 

Could be some sort of color encoding arithmetic overflow bug. Can you tell us what video processing and/or encoding steps occur between your three (R,G,B) sensors and your LCD monitor? Can you view the intermediate images to see where the problem appears?
 

The digitized output from the CCD (10bits/pixel) is captured by the computer using CamLink. It is then converted to so called XYZ format by using a 3X3 matrix and then to RGB format by another 3X3 matrix. Noise correction is done by averaging 16 frames. Strangely enough, the noise spots that I'm talking about don't appear on the noise corrected CCD output image or XYZ image. They only appear on the final noise corrected RGB image!
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top