remap algorithm [closed]

asked 2014-10-15 07:01:07 -0600

vestaproman gravatar image

Hello,

I am working on a stereoscopic system using an FPGA to do the disparity map. I use OpenCv to do the calibration of the cameras. cvInitUndistortRectifyMap give me the mx1, my1, mx2, my2 map. Using remap to rectify the camera with OpenCv give me good results. Now I would like to use the same map on my FPGA, and I would like to implement a remap fonction on my FPGA. Could anyone tell my the algorithm use in remap to do the functions?

Best regards,

VPM

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by sturkmen
close date 2020-09-17 07:14:21.648854

Comments

1

Did you see the docs?

boaz001 gravatar imageboaz001 ( 2014-10-15 07:52:11 -0600 )edit

Hello, Thanks for your comments Yes I read the doc but it is not very helpful. I also read this page http://docs.opencv.org/doc/tutorials/imgproc/imgtrans/remap/remap.html. But I have no explanation how it is done inside

Regards,

Jérôme

vestaproman gravatar imagevestaproman ( 2014-10-15 09:47:00 -0600 )edit

You asked for the algorithm, not the implementation. For the implementation I would say you can study the source here. But since you are going to write VHDL or Verilog I think the approach will be totally different. I suggest you start with understanding the algorithm and do some calculations (on paper) yourself. Then you can start drawing a hardware design with basic blocks and then see if you can make a translation to a HDL. You can let me know if you run into problems there.

boaz001 gravatar imageboaz001 ( 2014-10-16 02:31:09 -0600 )edit

Thanks boaz001, I do it in first step without interpolation, and it seems working

vestaproman gravatar imagevestaproman ( 2014-10-19 04:00:54 -0600 )edit
add a comment