Live data from Hacker News

Recent advances in 3D content understanding

ai.facebook.com

11–17 of 17 posts

Re: Recent advances in 3D content understanding

#11

It really seems like we should be close to having robots that can navigate a room.

You can get pretty far building one yourself using the out of the box ROS (https://www.ros.org/) navigation stack with a kinect, off the shelf motor controller, and something like an nvidia jetson

Re: Recent advances in 3D content understanding

#12

I do photogrammetry as a hobby, and would love to see more RGBD cameras on the market. I've even considered hacking together my own... anyone have pointers for cost-conscious options? A few of the models I've shrunk down and posted: https://sketchfab.com/darkphibre

The new Azure Kinect is really impressive. I've used the og Kinect, Kinect v2, and intel realsense d435 and it was much more accurate then all of those.

Sorry for derailing the thread, but how was the new Azure Kinect's performance outdoors? I've had trouble finding this information anywhere.

Re: Recent advances in 3D content understanding

#13
post #7
post #3

Earlier quoted context omitted.

Phones with multiple camera are starting to become common. Time of flight sensors that use projected infrared patterns aren't trivial to out in a phone so the demand needs to be there. Still, there have been tablets that have integrated Intel's depth cameras.

Nit: Projected patterns are structured light, not time-of-flight. As far as I’m aware (would love to be wrong!) you can’t do ToF with a traditional CCD or CMOS sensor and resolution is invariably woeful.

What kind of sensors do they use in time-of-flight? Because I think they do use gated CCD or CMOS...?

Re: Recent advances in 3D content understanding

#15
post #7

Earlier quoted context omitted.

Nit: Projected patterns are structured light, not time-of-flight. As far as I’m aware (would love to be wrong!) you can’t do ToF with a traditional CCD or CMOS sensor and resolution is invariably woeful.

What kind of sensors do they use in time-of-flight? Because I think they do use gated CCD or CMOS...?

Hm, last time I checked I thought the sensors were individual diodes but it looks like you're right, and 'flash LIDAR' cameras do use some funky kind of CCD. Still relatively low resolution but good to know, thanks!

Re: Recent advances in 3D content understanding

#16
post #2

Kinda boggles my mind that there hasn't been a stronger push for hardware support of 3D imaging in phones. It clearly provides more useful information for analysis. Even when we look at an image in a photo we map it to a 3D projection in our mind.

Not sure if this qualifies, but an interesting development which was new to me not so long ago was discovering the iPhone X has the ability to measure things with the camera. I.e. the measure app allows you to designate a point in the scene, draw a line and add another point and it will measure the distance. Also has nice snapping features and the ability to take a pic of the scene with the measurements superimposed on top.

Re: Recent advances in 3D content understanding

#17
post #12

Earlier quoted context omitted.

The new Azure Kinect is really impressive. I've used the og Kinect, Kinect v2, and intel realsense d435 and it was much more accurate then all of those.

Sorry for derailing the thread, but how was the new Azure Kinect's performance outdoors? I've had trouble finding this information anywhere.

np, I have yet to test it outdoors.
Post reply on HN