Performance evaluation with RealSense D435 in ROS2 humble (Intel® Robotics SDK)
Hi,
Platform: Raptor Lake S with RealSense D435
OS: Ubuntu 22.04 with ROS2 humble (Intel® Robotics SDK)
I know realsense-viewer is the tool to test or evaluate D435, but not in ROS2.
How to make performance evaluation with D435 in ROS2?
-
Hi Chic Lee If you install the RealSense ROS2 wrapper at the link below then you can use the D435 with ROS2 Humble.
-
Hi MartyG,
I know the RealSense ROS2 wapper and D435 works good with ROS2 Humble.
I want to know if D435 can hit specs with my RPL-S in ROS2 Humble.
Can you provide more tips for this?
-
I would recommend first finding out the best performance that you can achieve with the default settings on your Raptor Lake S computer. If performance is less than expected - for example, there is a drop in frames per second (FPS) speed or warning messages generating continuously in the terminal log - then you can then start trying various configuration settings to see whether performance and stability improves.
For example, if you are experiencing lower than desired FPS then you could set a low resolution and FPS to determine whether higher resolution and FPS settings are reducing performance.
ros2 launch realsense2_camera rs_launch.py depth_module.depth_profile:=640x480x15 rgb_camera.color_profile:=640x480x15
Another way to enhance performance is to build the ROS2 wrapper from source code with 'colcon build' and add the flag -DBUILD_ACCELERATE_GPU_WITH_GLSL to the colcon build command. Doing so provides GPU acceleration of pointclouds and depth colorization.
colcon build --cmake-args '-DBUILD_ACCELERATE_GPU_WITH_GLSL=ON'
Please sign in to leave a comment.
Comments
3 comments