Help understanding if fabricant provides OpenCV ready device.
Hi,
I'm having a conversation with a supplier for one of their devices, I'm basically asking if the model I'm requesting is OpenCV ready, which would imply it is also "OpenNI ready". I simply asked if I could use OpenCV and their classes, like vc2.VideoCapture to access the feed of the 3D ToF camera. This is what the supplier said:
"The use of OpenCV's image algorithm is no problem, but OpenCV's framework for reading Camera supports UVC Camera traditional lenses. M5 tof lenses generally use the OpenNI framework. You can use our SDK to get the image and do algorithm processing. There will be a fee to modify the code, and the SDK (non-face) used to access this device will replace the original."
Would this be enough? Am I asking the wrong questions? Or they are not being fully honest with me?
Thanks for any help!
not an answer, just an opinion:
you could try that idea, but it means building cv2 from src, with openni sdk onboard. (then you might be able to use cv2.VideoCapture from python)
imho, totally correct, that's what i would do.
unfortunately (i haven't looked !) their SDK is probably in c++, and you (seem to) are a python person
what ? try to ask again for clarification.
do you have to pay for the SDK (bad) or for doing a "special mod" (reasonable) ?
Nono, I have the SDK's and actually a few little examples on .cpp on Linux, and also SDK available for Windows (I plan to use it).
I suppose that what they are giving me is enough to run whatever I might need from the OpenCV library, right? I just need to "wrap" whatever I might need from the OpenNI library into a .cpp file and then call it from python? Is that what you intend to suggest? I like that approach.
Do you know where I can find some examples like this one @berak?
Thanks for your guidance! It has been helpful.