Great! So glad to see the video (and that things are also working smoothly)! Thanks for your time : )

This video made me order one! I can't wait to put it on my Robot body. Thanks!

Thanks for comparing a range of sensors there :) Have you been able to run this camera on ROS? Also would be keen to see some examples where you use cuda for point cloud/stereo processing.

@JetsonHacks

8 жыл бұрын

Intel has a ROS node for interfacing with the Realsense cameras, which I have successfully run on the Jetson. The next few video and article topics will be compiling a linux kernel and building a module for the R200 on the Jetson TK1, installing librealsense (which is a library for interfacing with the cameras) and building the ROS node and interfacing it with rviz on a remote computer. Stay tuned, and thanks for watching!

Hello sir, in this video, you showed multiple stereo cameras. According to you, which one is the best (including realsense D435) for robotics application (like rover avoiding obstacles autonomously). TIA.

Just bought one of these

@JetsonHacks

5 жыл бұрын

It should be fun to play with. Thanks for watching!

Hey Jim, thanks for doing a size comparison of all the cameras. Do you have an overall opinion of which camera you like best as far as performance goes? Extra bit of information: The kinect 360 (model 1414) out of the box is 548.6 grams. Stripping it down to only usable components makes it 176.5 grams.

@JetsonHacks

8 жыл бұрын

That's a difficult question to answer. It brings up other questions to qualify the answer. Is it an indoor or outdoor environment? If it's indoor only then there are several cameras to consider, if it has to work outdoors then that thins the heard considerably. Does the camera need to provide a color video stream? If so that eliminates cameras like the Structure camera and DUO-3D. Then there's packaging and integration. The Kinects can be a little beastly if you're trying to mount them on a robot, but are great for stationary indoor installations. Then there's dirty, filthy lucre. Inexpensive versus expensive in what I'll call the 3D consumer camera space. Say $100 to $500 USD. Then you have the bit about performance, and what it means in the application. Is frame rate important? Is the depth map accuracy important? How much post processing needs to be done? All those types of things. We can have that discussion on the website, KZread isn't the best place for such things. wp.me/p51nTS-AR

@jacksonkr_

8 жыл бұрын

Let's say both indoor and outdoor, in your opinion.

@JetsonHacks

8 жыл бұрын

That narrows the field to the Sterolabs ZED and the Intel RealSense R200. The combination of the ZED and a TX1 provides a high fidelity performant stereo camera, basically a 1920x1080 Depth stream and RGB stream. The drawback of that system is that it pretty much maxes out the TX1 because the TX1 GPU is used to build the depth map, there aren't a lot of compute cycles left for doing much else. I view the pair together as a nice stereo camera. Running the ZED on a TK1 can be a bit challenging. The R200 does everything on board the camera, so basically the Jetson just consumes the depth stream. The size of the depth map can be as large as 640x480. A depth map of that size is adequate for a lot of applications (such as mobile robotics) especially at the 60fps frame rate. The ZED has a longer range up to 15m, compared to the R200 which is As of this writing, the R200 requires a bit of tuning to get good results in any given environment. It's very early in the Linux driver development stage, so for example a transition from indoors to outdoors may need the camera parameters changed to get a good depth map out of it.

@jacksonkr_

8 жыл бұрын

That's quite detailed information, impressive. I'm going to do more research on the ZED output to see if I can leverage it without suffering a debilitating performance loss. Thanks for the info! Looking forward to more videos.

@patrickwasp

7 жыл бұрын

Have you had a chance to test the camera 4 months on? Have good results become easier to come by now?

Can you please tell which one will be better? Intel Realsense or the stereolabs zed camera for depth sensing

@JetsonHacks

6 жыл бұрын

Both are good cameras. 'Better' depends on the application. Thanks for watching!

I'm getting a green screen when running this tutorial on a TX2. Could you provide any insight?

how you get all that 3d camera??

How does this camera compared to the occipital one for 3d scanning?

BIG! SMALL! I want one! Will this need to be paired with a laptop to make it work with a robot? I saw the intel UP thingmabob, but its expensive. I am guessing an arduino or pi isnt gonna manage?

@JetsonHacks

7 жыл бұрын

The RealSense requires USB 3.0, which in practice means that you need something more powerful than a Raspberry Pi. The UP board is an x86 alternative, a Jetson is an ARM alternative. Both can run Linux and ROS.

Hello Jet , Which one is better this or Zed ?

Hey Jim - I'm upgrading my Turtlebot Waffle from an RPi 4 to a Jetson Nano. I have the R200 from when my Waffle had the Intel Joule, but the Ubuntu/ROS install on the Joule never worked properly. I also never had the R200 working with the RPi 4 and ROS. I'd like to use the R200 with the Nano to save having to buy yet another depth camera as well as for it's size. What is your opinion on trying to get the R200 working on the Nano with Ubuntu 18.04 and ROS Melodic? Im not finding much in the way of anyone doing that, probably due to the age of R200. Thanks

@JetsonHacks

3 жыл бұрын

I do not know the answer. It's been quite a while since I have used a R200. I do know that librealsense (v1) should work on the R200, but there may be a lot of work to get everything working smoothly on the Nano.

@wheagy

3 жыл бұрын

@@JetsonHacks I appreciate the reply. Thanks. Win

Does this have wireless connect ? I want attach it to my handmade helmet to do facial mocap hardware.

@JetsonHacks

7 жыл бұрын

The camera is USB 3.0 and requires a wire. Thanks for watching!

got this working on the raspberry pi 4 on raspbian finally lol

interesting channel subscribed

@JetsonHacks

8 жыл бұрын

Thank you for subscribing, and thank you for watching!