slam and R200

[pcroc]

can we do slam and object recognition/tracking with an R200?

[MartyG]

Yes, you can use the R200 for robotic SLAM. The article linked to below may be of use to you.

The '2016 R2' RealSense SDK for Windows also come packaged with Object Recognition and Scene Perception pre-made sample programs that work with R200. These samples can be accessed via the 'Sample Browser' application. This is inside a folder on the desktop called Intel RealSense SDK Gold that is put on the desktop when the R2 SDK is installed.

[pcroc]

ok thanks for your swift reply

can we use the same sdk under linux?

[MartyG]

For using the R200 with Linux, I would recommend using the open-source Librealsense SDK, and combining it with OpenCV and / or ROS for object tracking capability via software modules.

.

[pcroc]

uhmm, not very useful

the library you pointed out does not work apparently for R200. And even if it was working it does only provide point cloud and 3d depth support.

What is the point of using it with openCV, as it won't run on a GPU nor an ASIC and thus would not be efficient enough to be real time.

With windows, can we do real time slam ?

why can't we simply take the windows code and adapt it to linux?

[MartyG]

Can you make clear for me please what you mean when you say that the library does not work for R200? Are you looking at the RealSense SDK 2.0? Yes, that does not have R200 support. But I was referring to the earlier Librealsense SDK, which is R200 compatible.

[pcroc]

when searching for librealsense, here is the link I found https://github.com/IntelRealSense/librealsense GitHub - IntelRealSense/librealsense: Intel® RealSense™ SDK

And here are the first lines of the README of this library

Intel® RealSense™ SDK 2.0 is a cross-platform library for Intel® RealSense™ depth cameras (D400 series and the SR300).

the R200 series is not mentioned. but yes, in fact it is the SDK2

Thanks for giving me the former version link

Can we do more than just depth point cloud with the R200?

Otherwise a regular camera and a raspberry pi with opencv will do just exactly the same thing at a fraction of the cost.