Hello everyone. I am new to the Poppy project and I find this to be very fascinating. I am currently independently studying computer vision (Mostly OpenCV) and I am wondering if there is a need for giving Poppy sight? I am thinking such simple things as facial recognition and SLAM (simultaneous localization and mapping)
I would love to contribute what little I know on the subject.
Poppy sight is very interesting to study !
For example, Poppy Ergo Jr has a Raspberry Pi camera module (you can see it in this topic in french). We use it to recognize QR codes and it will be useful to go further.
It would be great to have camera at the end of the arm, but the connection cable is a problem (how to pass it along robot without damaging when moving ?).
There is indeed a camera in all Poppy Creatures (Humanoid, Torso, ErgoJr).
Except maybe for the QR Codes, we basically use only standard features from openCV. Yet we did not spend so much time on it, maybe few Friday afternoon.
So if you want to explore deeper the vision, you are very welcome and we will be pleased to help you.
I’ve been playing with Intel F200 camera, it is more precise than kinect.
Unfortunately for the moment, SDK is windows only.
Also it is USB3 only. I tested it with an Odroid XU3, it is recognized by linux, but it is not very stable for the moment but can grab frames.
A guy has blog on it: http://solsticlipse.com/blog.html
I did not know Intel was working on a kinect-like sensor. Thanks for the link.