real time video processing in FPGA xilinx virtex 5

i m doing project on real time video enhancement to remove effects of fog/haze from real time videos
i m using xilinx virtex 5.
its going to be so complicated because i m still facing problem with how to read real time video from camera and
and how to handle the image if converted to frames c
can anybody help me relating to this problem
how video is read in verilog? and how to apply algorithm on this video
plzzz help me asap

You dont read video as such with an HDL - all you will see is a stream of data (pixels in this case) and then you process it as a pipeline.

TrickyDicky said:

You dont read video as such with an HDL - all you will see is a stream of data (pixels in this case) and then you process it as a pipeline.

Click to expand...

i m donig like this, pixels are read but after that nothing is displayed or i m not aware where these pixels are stored and what they contained in xilinx

Hi

I am no expert but also am trying to do somewhat similar sort of thing, video but not camera.


And as tricky Dicky said, you dont read video, you read in form of Pixels, so 1 pixel is x-bit, store it somewhere in memory internal to FPGA or external like SDRAM or flash memory etc.

You need to explain a little more your system, your current techniques for capturing data etc.
I think break down your work into pieces. I am sorry I have an idea that you are in very initial phase. since you wrote following

i m donig like this, pixels are read but after that nothing is displayed or i m not aware where these pixels are stored and what they contained in xilinx

Click to expand...

How can you not know where your data is stored? This is among most essential things.

I think First you just try to read data as it is from camera etc. and just display or retrieve through FPGA. That internally you will do utliziing some buffer that can capture data. (find ping-pong or double buffer since in such applications this is used usually, although I have failed using this...:sad: ) But go ahead...

You need to know how many bits is 1-pixel, e.g. 8-bit 12 bit etc. this might depend on your ADC bit resolution.
Then how is ADC-sending-data-to-FPGA process starts, most commonly it is trigger based.

If you are able to read the video or play the video as it is then first step done...
Now start implementing your algorithm for processing.

Hope this serves,
Shan

zahra.asl said:

i m doing project on real time video enhancement to remove effects of fog/haze from real time videos
i m using xilinx virtex 5.
its going to be so complicated because i m still facing problem with how to read real time video from camera and
and how to handle the image if converted to frames c
can anybody help me relating to this problem
how video is read in verilog? and how to apply algorithm on this video
plzzz help me asap

Click to expand...

Hello I'm also working on same project and same kit. Can you provide source code?

If you don't know where the data is stored, logically I'm guessing you are using some others' source code provided either by Xilinx or third party IP. FIND THE SPEC FOR IT!

Camera output should have been sampled and converted to video data stream before you are able to process it.
Again, if you don't even know the basic of video stream format, how would you be able to enhance the video quality of it?