01-24-2012 03:43 AM
I am indending to connect two Dual image sensor FMC cards to the Spartan 6 Evaluation board in order to connect a total of 4 cameras for multi-view purposes. I know that connecting 2 cameras to one FMC is possible by duplicating the image processing pipeline. However, I am not sure whether this general scheme can be extended for 4 cameras on two FMC cards. Please help.
01-24-2012 05:45 AM
This hardware setup is definitely possible.
The FMC-IMAGEOV software driver will also support multiples instances.
You will need to change the example application for these multiple instances.
Also, make sure you plan out your FPGA resources.
If you will be using the Spartan-6 LX150T development board, you have one memory controller, supporting up to 4 bidirectionnal 32 bit ports.
If you need four video frame buffers (all camera inputs are asynchronous to each other),
then you will consume all of the memory controller ports for this task.
05-01-2012 04:25 PM - edited 05-01-2012 04:45 PM
I've got two OmniVision camera and a dual image sensor FMC Module from Avnet:
AES-FMC-IMAGEOV-G Dual Image Sensor FMC Module
AES-ACC-OV9517-G OmniVision 1 MP Camera Kit with Ribbon Cable
I want to connect the camera to my Xilinx ML605 FPGA board (with ISE 13.1) as the image input, but have no idea how to do it...
I am wondering how to capture pixels using Avnet FMC module and send the pixels to my FPGA board.
It seems that the Avnet FMC module can be directly connected to the FMC LPC of ML605 FPGA board.
Here are some questions:
1. Do I need something like a driver for the Dual Image Sensor FMC Module to work?
2. The camera can be connect to the Dual Image Sensor FMC Module via Ribbon cable, do I need to develop any interface for this connection?
3. If the answers are NO for the previous two questions, I am wondering if the only thing I am supposed to do is develop a interface which received pixels from camera.
I didn't find the datasheet for OV9517, but I referred to other camera OV7620 from Omnivision.
It seems that signal vsyn is HIGH for a few clocks in the beginning of each frame.
And there is another signal HREF, which indicates pixels are valid during HREF is HIGH. And a row of pixels will arrive for each period.
Do I need to develop an interface which only receives pixels after vsyn is HIGH while HREF is HIGH?
If yes, I have another question about the pin connection.
As mentioned, the camera is connected to the FMC module using Ribbon cable. How do I know which pin in the FMC module corresponds to the pin of camera(input/output)?
For the connection between FMC LPC of ML605 FPGA board and the FMC module from Avnet, they are directly connected.
So, with the pin connection known, the only thing I need to do is do the UCF file and the output of self-developed interface will be connected to the camera pin via the Avnet FMC module. Does this sound right?
By the way, is there a way to control the camera, for example, use a signal to tell the camera when to start scan the input image?