GPU accelerated Visual SLAM on Jetson Nano

Hi everyone! I have a Jetson Nano 4Gb carrier board and I’m willing to do Visual SLAM (with stereo camera) and autonomous Navigation with it. At first, I wanted to use Isaac ROS Visual SLAM and take advantage of the GPU accelerated code, but then I found out it is not possible to do so on this specific board (correct me if I’m wrong). So now I’m inclined to use RTAB_Map which, unfortunately, does not seem to provide a feature to be launched on the GPU. I’ve read some forums and users suggest building RTAB_Map’s dependencies (such as OpenCV) with GPU enabled and then build RTAB_Map itself towards these motioned dependencies. I wonder how hard it is to do this, considering I’m a newbie, and whether I’ll be able to see a significant performance difference compared to running it on the CPU. I’ll be grateful for any kind of advice.

Hi @juan.r, these ROS containers for Jetson are built with OpenCV that has CUDA enabled:

Or it looks like RTAB-Map might have containers for JetPack:

Thanks for the reply! I’ll take a look.


This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.