It would be great if objects in the surroundings could be recognized during visual localization. That way, when I use voice commands to tell the robot to go 'next to the sofa,' the robot could understand what I said and autonomously navigate to the sofa's location.
@@rikumurai no way... that's impressive! Is there a comparison between results with depth cameras? Even if it's worse it'd still be regardlessly impressive
It would be great if objects in the surroundings could be recognized during visual localization. That way, when I use voice commands to tell the robot to go 'next to the sofa,' the robot could understand what I said and autonomously navigate to the sofa's location.
Great work! When will the code be released?
I am kneeling in front of the mighty video.
Can this be done with a 360 camera?
If 3090 can reproduce it? How much video memory is needed?
what is the GPU requirements for a real-time application of this method?
We used a single RTX 4090 for all the results shown in the video
Which robotic system is that?
It's Diablo from Direct Drive Tech
Is it a normal RGB camera or a depth camera like Intel Realsense ones?
It uses a normal RGB camera!
@@rikumurai no way... that's impressive!
Is there a comparison between results with depth cameras? Even if it's worse it'd still be regardlessly impressive
the robot only found six ring wraiths; where the other three are; we do not know.
Wow
怎么使用