Hardware Setup

RealSense camera not found

Symptom

No RealSense cameras detected when running realsense-viewer or ros2 launch realsense2_camera rs_launch.py

Solution

Ensure that you follow the cameras to install the correct versions for both librealsense and realsense2_camera as mentioned in the doc.

To confirm this, run ros2 launch realsense2_camera rs_launch.py inside the docker container. You should see the following:

[realsense2_camera_node-1] [INFO] [1715878272.636553721] [camera.camera]: RealSense ROS v4.54.1
[realsense2_camera_node-1] [INFO] [1715878272.636742684] [camera.camera]: Built with LibRealSense v2.55.1
[realsense2_camera_node-1] [INFO] [1715878272.636772605] [camera.camera]: Running with LibRealSense v2.55.1

RealSense Driver Does Not Work with ROS 2 Humble

Symptom

As of Sep 13 2022, the Realsense ROS 2 wrapper package’s ros2 branch does not support ROS 2 Humble. Building the package on Humble produces the below error:

CMake Error at CMakeLists.txt:142 (message):   Unsupported ROS Distribution: humble

Solution

Use the RealSense ROS 2 driver ros2-development branch

Input Images Must Have an Even Height and Width

Symptom

Isaac ROS nodes terminate when given images that have an odd width or height:

[component_container_mt-1] [INFO] [1655745951.986574909] [NitrosContext]: [NitrosContext] Loading application: '/workspaces/isaac_ros-dev/ros_ws/install/isaac_ros_nitros/share/isaac_ros_nitros/config/type_adapter_nitros_context_graph.yaml'
[component_container_mt-1] [ERROR] [1655747175.384136779] [NitrosImage]: [convert_to_custom] Image width/height must be even for creation of gxf::VideoBuffer
[component_container_mt-1] terminate called after throwing an instance of 'std::runtime_error'
[component_container_mt-1]   what():  [convert_to_custom] Odd Image width or height.
[ERROR] [component_container_mt-1]: process has died [pid 28538, exit code -6, cmd '/opt/ros/humble/install/lib/rclcpp_components/component_container_mt --ros-args -r __node:=apriltag_container -r __ns:=/'].

Solution

Replace the input image source with one that produces images that have even width and height.

RealSense ros2 launch realsense2_camera rs_launch.py Error

Symptoms

Launching the RealSense node using ros2 launch realsense2_camera rs_launch.py, produces the following error message:

[realsense2_camera_node-1]  29/06 01:10:27,431 WARNING [140061797918464] (rs.cpp:310) null pointer passed for argument "device"
[realsense2_camera_node-1] [WARN] [1656465033.464454660] [camera.camera]: Device 1/1 failed with exception: failed to set power state
[realsense2_camera_node-1] [ERROR] [1656465033.464505994] [camera.camera]: The requested device with  is NOT found. Will Try again.
[realsense2_camera_node-1]  29/06 01:10:33,463 ERROR [140061781133056] (handle-libusb.h:51) failed to open usb interface: 0, error: RS2_USB_STATUS_NO_DEVICE
[realsense2_camera_node-1]  29/06 01:10:33,463 ERROR [140061797918464] (sensor.cpp:572) acquire_power failed: failed to set power state

Solution

Before starting the isaac_ros-dev Docker container using run_dev.sh, run setup_udev_rules.sh from librealsense in a terminal outside Docker.

Solution 2

  1. Check if the 99-realsense-libusb.rules file exists in /etc/udev/rules.d/.

  2. If not, disconnect the camera, copy this file to /etc/udev/rules.d/, then reconnect the camera.

Failed to Get Valid Output from Isaac ROS Rectify or AprilTag Node

Symptom

If there is no available calibration data for an Argus camera, you see warning messages similar to:

WARN  extensions/hawk/argus_camera.cpp@677: Failed to get calibration data from Argus!

Solution

Most camera modules require calibration for lens distortion. Without camera calibration, rectification outputs invalid results and impacts the accuracy of downstream Nodes in the graph, such as AprilTag detection.

See the camera calibration guide to calibrate the camera. If you use Argus cameras, the Isaac ROS Argus Node provides options to get the calibration parameters from either the device driver or a .ini file. See this section for additional details.

RealSense Does Not Stream IR Stereo Images

Symptom

Within the Docker container: Cannot see any IR images in realsense-viewer, but depth images can be streamed.

Outside the Docker container: No metadata containing projector status attached to the IR frames.

Solution

  1. Download and install the dkms for Kernel 5.15 from GitHub

    1. Download the .deb file named librealsense2-dkms-dkms_1.3.14_amd64.deb.

    2. In the directory where you downloaded the file run: sudo apt install ./librealsense2-dkms-dkms_1.3.14_amd64.deb

  2. Build librealsense within the container without CUDA. Run the following commands within the Docker container:

  • git clone https://github.com/JetsonHacksNano/installLibrealsense

  • cd installLibrealsense

  • ./installLibrealsense.sh

  • ./buildLibrealsense.sh --no_cuda

Note

For more details on this issue, refer to the nvblox troubleshooting section

RealSense Error Failed to resolve the request

Symptom

When any RealSense tutorial is launched, no images are streamed and the following error is found in the terminal logs:

[component_container_mt-1] [ERROR] [1665684669.343408502] [camera]: /workspaces/isaac_ros-dev/ros_ws/src/third_party/realsense-ros/realsense2_camera/src/rs_node_setup.cpp:344:An exception has been thrown:
[component_container_mt-1] Failed to resolve the request:
[component_container_mt-1]  Format: Y8, width: 640, height: 480
[component_container_mt-1]  Format: Y8, width: 640, height: 480
[component_container_mt-1]
[component_container_mt-1] Into:
[component_container_mt-1]  Formats:
[component_container_mt-1]   Y8I
[component_container_mt-1]
[component_container_mt-1]
[component_container_mt-1] [ERROR] [1665684669.343506391] [camera]: Error starting device:
[component_container_mt-1] Failed to resolve the request:
[component_container_mt-1]  Format: Y8, width: 640, height: 480
[component_container_mt-1]  Format: Y8, width: 640, height: 480
[component_container_mt-1]
[component_container_mt-1] Into:
[component_container_mt-1]  Formats:
[component_container_mt-1]   Y8I
[component_container_mt-1]

Solution

Update the RealSense driver to the latest version using the following steps:

  1. Install librealsense on your local system using instructions from IntelRealSense/librealsense,

  2. Download the latest firmware binary(for example: Signed_Image_UVC_5_13_0_55.bin) from Firmware releases D400.

  3. Extract the binary from the .zip file.

  4. Run the following command, being sure to replace <binary_filename> with the name of the binary you downloaded:

rs-fw-update -f <binary_filename>

Note

For more information refer to the RealSense update tool.

RealSense incompatible QoS Policy Error

Symptom

When any RealSense tutorial is launched, the output topic from RealSense is not subscribed and the following warning is found in the terminal logs:

[component_container_mt-2] [WARN] [1666117836.386565040] [left_encoder_node]: New publisher discovered on topic '/color/image_raw', offering incompatible QoS. No messages will be sent to it. Last incompatible policy: RELIABILITY_QOS_POLICY
[component_container_mt-2] [WARN] [1666117836.392021093] [camera]: New subscription discovered on topic '/color/image_raw', requesting incompatible QoS. No messages will be sent to it. Last incompatible policy: RELIABILITY_QOS_POLICY

Solution

Change the QoS policy in realsense.yaml file from the launched package. Take isaac_ros_h264_encoder as an example, change policy from SENSOR_DATA to SYSTEM_DEFAULT could make the RealSense rgb_camera publisher compatible with the subscriber of isaac_ros_h264_encoder.

Note

For more information on Quality of Service compatibilities see Quality of Service settings

Linker Error foo.so: file format not recognized; treating as linker script

Symptom

When attempting to build an Isaac ROS package by running colcon build, the linker complains with an error that resembles the following:

--- stderr: isaac_ros_nitros
/usr/bin/ld:/workspaces/isaac_ros-dev/src/isaac_ros_nitros/isaac_ros_nitros/gxf/lib/gxf_x86_64_cuda_11_7/core/libgxf_core.so: file format not recognized; treating as linker script
/usr/bin/ld:/workspaces/isaac_ros-dev/src/isaac_ros_nitros/isaac_ros_nitros/gxf/lib/gxf_x86_64_cuda_11_7/core/libgxf_core.so:1: syntax error
collect2: error: ld returned 1 exit status
make[2]: *** [CMakeFiles/isaac_ros_nitros.dir/build.make:311: libisaac_ros_nitros.so] Error 1
make[1]: *** [CMakeFiles/Makefile2:139: CMakeFiles/isaac_ros_nitros.dir/all] Error 2
make: *** [Makefile:146: all] Error 2
---
Failed   <<< isaac_ros_nitros [0.52s, exited with code 2]

Solution

This error typically arises when the local copy of a .so binary on your machine contains only the Git LFS pointer, instead of the actual binary contents.

Note

If you’re encountering this error, you may have forgotten to install and perform one-time setup for Git LFS on your machine. See the instructions for development environment setup.

To correct this, navigate to the Isaac ROS package that is failing to build and re-pull the binary files using Git LFS.

For example, if isaac_ros_nitros is failing to build as shown above:

  1. Navigate to the package:

    cd ${ISAAC_ROS_WS}/src/isaac_ros_nitros
    
  2. Pull from Git LFS:

    git lfs pull
    
  3. Return to the workspace and run the build command again:

    cd ${ISAAC_ROS_WS} && colcon build --symlink-install
    

Intel RealSense Camera Accidentally Enables Laser Emitter

Symptom

Despite emitter_enabled set to 0, the laser emitter on Intel RealSense comes on.

Solution

Run the following command in the run time:

ros2 param set /camera/camera depth_module.emitter_enabled 0

Intel RealSense D455’s Infra Camera Capped at 15fps

Symptom

Even though it’s specified to run at 90fps, Intel RealSense D455’s infra camera capped at 15fps.

Solution

Due to a known issue, the initial framerate might be capped around 15fps.

If that is the case, you need to issue this command:

ros2 param set /camera/camera depth_module.enable_auto_exposure true

Cannot Access Recovery Mode Button On Jetson AGX Orin

The Jetson AGX Orin needs to be put into recovery mode but the recovery mode button is inaccessible.