RGB+ Depth, Video + Kinect = Video depth

Hi again!
Today I’ve been in a workshop of RGB+ Depth trying to add depth into a video! Specially thanks to James George who made it possible!

What we did is add a normal video stream depth. How? Everybody knows how, using Kinect, and some software to get both mixed. The result, quite impresive. You can go throw the video, change the angle, and get an impression of what shape does it take the things you can see. Here is an example.




I hope I will get a real example made by my team with something cool to show. This will need to wait until everything works tomorrow.

But how we can do it?

What we need is a DSLR camera with video function and a Kinect. Once we get both, we need to join with some kind of mounting provided by organization (you could find the free models into website of the project, here). Once we get something like this, we can continue.

DSLR + Kinect

The most important think you need to know is to get completely fixed camera and Kinect in the same position and angle. Why? Because you need to calibrate both to make Kinect know where the camera to get the depth properly. And trust me, once you calibrate your “camerakinect” you will not want to do it again… at least, the same day, so charge your battery at MAX or use mounting camera dock, this will help a lot.

Calibration is tedious, boring, but necessary. So get the time you need the get a good calibration, if not, you will need to repeated… and trust me again, you will not want to do it again…

After that, you will just need to start recording everything you want! Using tripod or handle with your hands, but always remembering that you could not move Kinect+camera. The rest is just your imagination.

Of course, to record with Kinect you will need to get your computer attached and use a power cord to supply power to Kinect. This can make camera stay not in a very big movement, so I prefer to use tripod and get camera fixed. Once start recording data from Kinect is the time to record video, Clap! and start your video depth recording!

Why clap? Why did you see in every film take? To adjust later and get video stream and Kinect data stream in sync. So use great technology called Clap! and start recording, nice, easy, and cheap!

Once you get your scene recorded, you will just need to get everything in folders, the calibration files and Kinect data. Later start RGBDVisualize and you will just need to fine tuning video stream with Kinect data, and get your amazing video out of the oven!

You could see the video in depth moving the camera just like Quake and getting details you need think they are there. It’s really amazing!

Next steps? Improving software, and get more and more supports! And of course, getting better and better videodepth!

If you have one in youtube, video, or just in .mov, please send us and we will post it here! I hope I could show mine too soon!


More information: RGBDToolkit

Uso de cookies

Este sitio web utiliza cookies para que usted tenga la mejor experiencia de usuario. Si continúa navegando está dando su consentimiento para la aceptación de las mencionadas cookies y la aceptación de nuestra política de cookies, pinche el enlace para mayor información.plugin cookies