Using a checkered grid or input calibration paremeters in stereo vision

[RogerioBReis]

I would like to use a checkerboard grid, like a chessboard, to obtain the calibration parameters for the stereo view. Can anybody help me? There is the possibility to input the calibration data performed outside Labview?

[Bob_Schor]

I'm unsure what it is you are attempting to do.  Do you have a reference for the algorithm you want to perform?  Is the "checkered grid" a physical object whose image you acquire with one or two cameras?  Is it something you want to generate as an IMAQ image and use in some algorithm (what algorithm)?

Bob Schor

[RogerioBReis]

Bob_Schor. I appreciate the attention. I have several images taken with two cameras that I would like to analyze. The calibration of the stereo system was performed, using MATLAB, with a chessboard. So, I have two options at the moment to generate the depth map of these scenes: use some VI that learns the camera model from a chess board or find some way to input the intrinsic and extrinsic parameters, that I know, in the stereo session of my VI. I don't have the cameras now, but I would need to have the VI that performs the depth map when joining the devices.

[wiebe@CARYA]

OpenCV has functions for this... Not sure if that helps. You can use OCV from Python... And Python from LabVIEW 😁.

https://docs.opencv.org/2.4/modules/calib3d/doc/camera_calibration_and_3d_reconstruction.html

[RogerioBReis]

It is a possibility. I had'n thought this way. I knew I could use Matlab from Labview but not Python. I will analyze that approach. Thank you.

[wiebe@CARYA]

I am working on an open\free OpenCV library for LabVIEW.

Those functions are advanced though, and where simple functions can be used directly from the OCV dll, advanced functions will need a wrapper. I'm not that far in the process. A few more months is optimistic 😒.

[RogerioBReis]

Really? Who am I to try? I couldn't even install OCV.

[wiebe@CARYA]

From Python it doesn't seem too hard: py_setup_in_windows. Of course if you're not familiar with Python there's a learning curve.

There's also OpenCVSharp, one of the OCV .NET wrappers, but you still have to figure out the .net<>LV interface. And, IIRC, it's only has some of the advanced stuff. Checkerboard calibration is there, but 3D I don't know. It doesn't support OCV GPU features well, AFAIK.