[upl-image-preview url=https://discuss.luxonis.com/assets/files/2023-05-10/1683744441-418096-screenshot-2023-05-10-at-113831-am-1.png] Ran [OpenCV's Intrinsic Calibration](https://docs.opencv.org/4.x/dc/dbb/tutorial_py_calibration.html) to improve calibration for our OakD W Pro. Can someone please explain why these intrinsics are so different compared to the default intrinsics from OakD? Is it possible the units are mismatched?

OpenCV Intrinsics not the same as Default OakD W Pro Intrinsics

saiborg

Ran OpenCV's Intrinsic Calibration to improve calibration for our OakD W Pro. Can someone please explain why these intrinsics are so different compared to the default intrinsics from OakD? Is it possible the units are mismatched?

jakaskerl

Hi saiborg
I don' think they are mismatched, just scaled differently. Notice the 1e-3 in the first matrix z coordinate as opposed to 1 in the second. They should both return a similar result (wide FOV does introduce more distortion though).

Hope this helps,
Jaka

HiepNguyen

Hi @jakaskerl ,

I'm facing the same issue, when using then OAK-D camera with Stella SLAM. With the default OAK-D camera intrinsics, the output of Stella SLAM is wrong. Then I used OpenCV calibration tools to calibrate the OAK-D camera and use that output for Stella SLAM, the output is pretty correct.

With Intel Realsense camera intrinsics, the Stella SLAM works normally.

Therefore, I don't know what is different between default OAK-D camera intrinsics and OpenCV calibration tools, because the OAK-D calibration tools use the same functions as OpenCV calibration tools.

Can you explain the scale and how can I re-scale these OAK-D camera intrinsics, to get the same result as OpenCV calibration tools ?

Many thanks in advance !

jakaskerl

Hi HiepNguyen
Could you post an example of intrinsics you are receiving? Afaik the common practice is to keep bottom right value of a homogenous matrix 1.

Thanks,
Jaka