I want to depict a Model in 3d which is basically a rectanglular piece of Weld in Ros Humble Gazebo and Rviz. Could you give some suggestions on which camera do i need to us? Do i need a Depth Camera or a combination of Depth Camera and RGB camera?
1 Answer
$\begingroup$
$\endgroup$
3
Based on the scope of your application, certain cameras can provide depth data tailored to the specific range you're interested in. Not sure about all the details of your project, but maybe check out the Intel RealSense D435? I've found it's pretty solid and a lot of folks use it. It's kinda the go-to for both virtual and real-world stuff.
-
$\begingroup$ ok Thanks a lot. I will look into it. $\endgroup$– KedarCommented Nov 2, 2023 at 7:46
-
$\begingroup$ The Goal of the Project is bring a 3d Object from the Real world and perform a Image detection with the Help of a Deep Learning Model, which i plan on training. The Robot should move to certain location after it identifies some set of features depending on the image to move there. $\endgroup$– KedarCommented Nov 3, 2023 at 7:06
-
$\begingroup$ Well it is hard answer without a specific question. In my experience, if you are going to run a real-time inference with a deep learning model with camera it really depends on the model that you are using. It seems you are in the preliminary phase of the project, so my recommendation is to get familiar with the relasense camera, compile it with GPU support, and play around the ros2 topics it provides(check github link). There is one topic that gives you the depth aligned with the RGB image which you can subscribe to and use the bounding box coordinate from the deep learning model. $\endgroup$ Commented Nov 3, 2023 at 14:31