`isaac_ros_data_replayer`

Source code on GitHub.

Overview

The isaac_ros_data_replayer package enables replaying sensor data recorded using the isaac_ros_data_recorder package. This package also enables reliable extraction of stereo cameras in different formats, including depth using AI-based perception. Supported formats include MP4, PNG, and PFM. Stereo camera frames are extracted in synchronized pairs while depth frames are aligned to the left stereo imager.

Quickstart

Set Up Development Environment

Set up your development environment by following the instructions in getting started.

Clone isaac_ros_common under ${ISAAC_ROS_WS}/src.

cd ${ISAAC_ROS_WS}/src && \
   git clone -b release-3.2 https://github.com/NVIDIA-ISAAC-ROS/isaac_ros_common.git isaac_ros_common

(Optional) Install dependencies for any sensors you want to use by following the sensor-specific guides.

Note

We strongly recommend installing all sensor dependencies before starting any quickstarts. Some sensor dependencies require restarting the Isaac ROS Dev container during installation, which will interrupt the quickstart process.

Download Quickstart Assets

Download quickstart data from NGC:

Make sure required libraries are installed.

sudo apt-get install -y curl jq tar

Then, run these commands to download the asset from NGC:

NGC_ORG="nvidia"
NGC_TEAM="isaac"
PACKAGE_NAME="isaac_ros_data_replayer"
NGC_RESOURCE="isaac_ros_data_replayer_assets"
NGC_FILENAME="quickstart.tar.gz"
MAJOR_VERSION=3
MINOR_VERSION=2
VERSION_REQ_URL="https://catalog.ngc.nvidia.com/api/resources/versions?orgName=$NGC_ORG&teamName=$NGC_TEAM&name=$NGC_RESOURCE&isPublic=true&pageNumber=0&pageSize=100&sortOrder=CREATED_DATE_DESC"
AVAILABLE_VERSIONS=$(curl -s \
    -H "Accept: application/json" "$VERSION_REQ_URL")
LATEST_VERSION_ID=$(echo $AVAILABLE_VERSIONS | jq -r "
    .recipeVersions[]
    | .versionId as \$v
    | \$v | select(test(\"^\\\\d+\\\\.\\\\d+\\\\.\\\\d+$\"))
    | split(\".\") | {major: .[0]|tonumber, minor: .[1]|tonumber, patch: .[2]|tonumber}
    | select(.major == $MAJOR_VERSION and .minor <= $MINOR_VERSION)
    | \$v
    " | sort -V | tail -n 1
)
if [ -z "$LATEST_VERSION_ID" ]; then
    echo "No corresponding version found for Isaac ROS $MAJOR_VERSION.$MINOR_VERSION"
    echo "Found versions:"
    echo $AVAILABLE_VERSIONS | jq -r '.recipeVersions[].versionId'
else
    mkdir -p ${ISAAC_ROS_WS}/isaac_ros_assets && \
    FILE_REQ_URL="https://api.ngc.nvidia.com/v2/resources/$NGC_ORG/$NGC_TEAM/$NGC_RESOURCE/\
versions/$LATEST_VERSION_ID/files/$NGC_FILENAME" && \
    curl -LO --request GET "${FILE_REQ_URL}" && \
    tar -xf ${NGC_FILENAME} -C ${ISAAC_ROS_WS}/isaac_ros_assets && \
    rm ${NGC_FILENAME}
fi

Build `isaac_ros_data_replayer`

Launch the Docker container using the run_dev.sh script:

cd ${ISAAC_ROS_WS}/src/isaac_ros_common && \
  ./scripts/run_dev.sh

Install the prebuilt Debian package:

sudo apt-get update

sudo apt-get install -y ros-humble-isaac-ros-data-replayer
source /opt/ros/humble/setup.bash

Install the required assets:

sudo apt-get install -y ros-humble-isaac-ros-ess-models-install
source /opt/ros/humble/setup.bash

ros2 run isaac_ros_ess_models_install install_ess_models.sh --eula

Clone this repository under ${ISAAC_ROS_WS}/src:

cd ${ISAAC_ROS_WS}/src && \
   git clone -b release-3.2 https://github.com/NVIDIA-ISAAC-ROS/isaac_ros_nova.git isaac_ros_nova

Launch the Docker container using the run_dev.sh script:

cd ${ISAAC_ROS_WS}/src/isaac_ros_common && \
   ./scripts/run_dev.sh

Use rosdep to install the package’s dependencies:

sudo apt-get update

rosdep update && rosdep install -i -r --from-paths ${ISAAC_ROS_WS}/src/isaac_ros_nova/isaac_ros_data_replayer/ --rosdistro humble -y

Install the required assets:

sudo apt-get install -y ros-humble-isaac-ros-ess-models-install
source /opt/ros/humble/setup.bash

ros2 run isaac_ros_ess_models_install install_ess_models.sh --eula

Build the ROS package in the Docker container:

colcon build --symlink-install --packages-up-to isaac_ros_data_replayer
source install/setup.bash

Run Launch File

ros2 launch isaac_ros_data_replayer data_replayer.launch.py \
   rosbag:=${ISAAC_ROS_WS}/isaac_ros_assets/isaac_ros_data_replayer/quickstart

Replay a single camera

ros2 launch isaac_ros_data_replayer data_replayer.launch.py \
   rosbag:=${ISAAC_ROS_WS}/isaac_ros_assets/isaac_ros_data_replayer/quickstart \
   enabled_stereo_cameras:=front_stereo_camera enabled_fisheye_cameras:=None enable_3d_lidar:=False

Visualize

Open a new terminal inside the Docker container:

cd ${ISAAC_ROS_WS}/src/isaac_ros_common && \
   ./scripts/run_dev.sh

Run RViz to visualize data:
```
rviz2
```

Data Extraction

To extract synchronized camera frames from a rosbag, run:

ros2 run isaac_ros_data_replayer data_extraction.py \
   --rosbag ${ISAAC_ROS_WS}/isaac_ros_assets/isaac_ros_data_replayer/quickstart \
   --camera front_stereo_camera --output ${ISAAC_ROS_WS} \
   --engine_file_path ${ISAAC_ROS_WS}/isaac_ros_assets/models/dnn_stereo_disparity/dnn_stereo_disparity_v4.1.0_onnx/ess.engine

ros2 run isaac_ros_data_replayer data_extraction.py \
   --rosbag ${ISAAC_ROS_WS}/isaac_ros_assets/isaac_ros_data_replayer/quickstart \
   --camera front_stereo_camera --output ${ISAAC_ROS_WS} \
   --engine_file_path ${ISAAC_ROS_WS}/isaac_ros_assets/models/dnn_stereo_disparity/dnn_stereo_disparity_v4.1.0_onnx/light_ess.engine

By default, rectified camera frames will be extracted alongside disparity and depth images generated by ESS. To extract rectified images without depth or disparity, run:

ros2 run isaac_ros_data_replayer data_extraction.py \
   --rosbag ${ISAAC_ROS_WS}/isaac_ros_assets/isaac_ros_data_replayer/quickstart \
   --camera front_stereo_camera --output ${ISAAC_ROS_WS} \
   --mode rectify

To extract raw images, run:

ros2 run isaac_ros_data_replayer data_extraction.py \
   --rosbag ${ISAAC_ROS_WS}/isaac_ros_assets/isaac_ros_data_replayer/quickstart \
   --camera front_stereo_camera --output ${ISAAC_ROS_WS} \
   --mode raw

Rosbag Conversion

The isaac_ros_data_recorder records camera streams as sensor_msgs/msg/CompressedImage messages encoded with H.264 encoding, and 3D LIDAR data as hesai_ros_driver/msg/UdpFrame.

The rosbag_converter tool converts rosbags from isaac_ros_data_recorder so that camera streams are stored uncompressed as sensor_msgs/msg/Image messages and 3D LIDAR data as sensor_msgs/msg/PointCloud2 messages (with fields x, y, z, intensity, ring, timestamp). For additional details refer to isaac_ros_hesai documentation.

ros2 run isaac_ros_data_replayer rosbag_converter.py \
   -i ${ISAAC_ROS_WS}/isaac_ros_assets/isaac_ros_data_replayer/quickstart \
   -o ${ISAAC_ROS_WS}/uncompressed_rosbag/ \
   --all

Note

Disk space requirements: The rosbag_converter tool requires free disk space of at least 100 times the size of the original rosbag. Specific requirements may vary and depend on the number of sensors recorded in the original rosbag, among other things.

Note

The rosbag_converter tool is expected to be used on x86_64, with sufficient disk space under the /tmp directory. The original rosbag needs to be copied to the x86_64 machine where the conversion will be performed.

Troubleshooting

Isaac ROS Troubleshooting

For solutions to problems with Isaac ROS, see troubleshooting.

Not all camera frames extracted from recording

Data extraction extracts synchronized camera frames from a recording. If the number of frames from the left and right camera are not equal, then some frames will be ignored to keep the left and right cameras in sync.

Not enough free disk space to run the `rosbag_converter` tool

Symptom

Execution of the rosbag_converter tool crashes with a python error, including but not limited to:

RuntimeError: Exception on parsing info file: invalid node; first invalid key: "version"
[ros2run]: Process exited with failure 1

Solution

Free additional disk space before executing the rosbag_converter tool.

Alternatively, consider reducing the size of the input rosbag, or consider converting only the subset of topics needed for your application. Note that the argument -a or --all converts all the possible topics in the input rosbag, but the arguments --topics_decode_h264 and --topics_udp_to_pc2 allow for the conversion of specific topics.

Make sure to use the tool on a x86_64 machine.

API

Usage

ros2 launch isaac_ros_data_replayer data_replayer.launch.py rosbag:=<rosbag>

ROS Launch Arguments

ROS Launch Argument	Default Value	Description
`rosbag`	None	Path to recording.
`enabled_stereo_cameras`	`front_stereo_camera` `back_stereo_camera` `left_stereo_camera` `right_stereo_camera`	Stereo cameras enabled for replay.
`enabled_fisheye_cameras`	`front_fisheye_camera` `back_fisheye_camera` `left_fisheye_camera` `right_fisheye_camera`	Fisheye cameras enabled for replay.
`enable_3d_lidar`	`True`	Enable 3D lidar for replay.
`replay_loop`	`False`	Loop replay.
`replay_rate`	`1.0`	Extraction rate; 1.0 for real-time, <1.0 for sub-real-time, >1.0 for super-real-time.
`replay_delay`	None	Delay to start replay.
`replay_additional_args`	`--disable-keyboard-controls`	Additional arguments for replay.

ROS Topics Published

ROS Topic	Interface	Description
`/rosout`	rcl_interfaces/msg/Log	Console logs.
`/tf`	tf2_msgs/msg/TFMessage	Movable transforms on the robot.
`/tf_static`	tf2_msgs/msg/TFMessage	Fixed transforms on the robot.
`/front_stereo_camera/left/image_raw`	sensor_msgs/msg/Image	Front stereo camera left camera stream.
`/front_stereo_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Front stereo camera left camera intrinsics.
`/front_stereo_camera/right/image_raw`	sensor_msgs/msg/Image	Front stereo camera right camera stream.
`/front_stereo_camera/right/camera_info`	sensor_msgs/msg/CameraInfo	Front stereo camera right camera intrinsics.
`/back_stereo_camera/left/image_raw`	sensor_msgs/msg/Image	Back stereo camera left camera stream.
`/back_stereo_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Back stereo camera left camera intrinsics.
`/back_stereo_camera/right/image_raw`	sensor_msgs/msg/Image	Back stereo camera right camera stream.
`/back_stereo_camera/right/camera_info`	sensor_msgs/msg/CameraInfo	Back stereo camera right camera intrinsics.
`/left_stereo_camera/left/image_raw`	sensor_msgs/msg/Image	Left stereo camera left camera stream.
`/left_stereo_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Left stereo camera left camera intrinsics.
`/left_stereo_camera/right/image_raw`	sensor_msgs/msg/Image	Left stereo camera right camera stream.
`/left_stereo_camera/right/camera_info`	sensor_msgs/msg/CameraInfo	Left stereo camera right camera intrinsics.
`/right_stereo_camera/left/image_raw`	sensor_msgs/msg/Image	Right stereo camera left camera stream.
`/right_stereo_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Right stereo camera left camera intrinsics.
`/right_stereo_camera/right/image_raw`	sensor_msgs/msg/Image	Right stereo camera right camera stream.
`/right_stereo_camera/right/camera_info`	sensor_msgs/msg/CameraInfo	Right stereo camera right camera intrinsics.
`/front_fisheye_camera/left/image_raw`	sensor_msgs/msg/Image	Front fisheye camera stream.
`/front_fisheye_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Front fisheye camera intrinsics.
`/back_fisheye_camera/left/image_raw`	sensor_msgs/msg/Image	Back fisheye camera stream.
`/back_fisheye_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Back fisheye camera intrinsics.
`/left_fisheye_camera/left/image_raw`	sensor_msgs/msg/Image	Left fisheye camera stream.
`/left_fisheye_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Left fisheye camera intrinsics.
`/right_fisheye_camera/left/image_raw`	sensor_msgs/msg/Image	Right fisheye camera stream.
`/right_fisheye_camera/left/camera_info`	sensor_msgs/msg/CameraInfo	Right fisheye camera intrinsics.
`/front_2d_lidar/scan`	sensor_msgs/msg/LaserScan	Front 2D lidar scan.
`/back_2d_lidar/scan`	sensor_msgs/msg/LaserScan	Back 2D lidar scan.
`/front_3d_lidar/lidar_points`	sensor_msgs/msg/PointCloud2	Front 3D lidar point cloud.
`/front_stereo_imu/imu`	sensor_msgs/msg/Imu	Front stereo camera inertial measurement unit.
`/chassis/imu`	sensor_msgs/msg/Imu	Chassis inertial measurement unit.
`/chassis/ticks`	isaac_ros_nova_interfaces/msg/EncoderTicks	Chassis encoder count.
`/chassis/odom`	nav_msgs/msg/Odometry	Chassis odometry.
`/chassis/battery_state`	sensor_msgs/msg/BatteryState	Chassis battery state.

Usage

ros2 run isaac_ros_data_replayer data_extraction.py --rosbag <rosbag> --camera <camera> --output <output>

ROS Run Arguments

ROS Run Argument	Default Value	Description
`--rosbag`	None	Path to recording.
`--camera`	None	Camera to extract i.e. `front_stereo_camera`.
`--output`	None	Directory to store extracted data.
`--min_disparity`	`3.0`	Minimum disparity value for normalization.
`--max_disparity`	`200.0`	Maximum disparity value for normalization.
`--threshold`	`0.0`	ESS threshold.
`--engine_file_path`	`isaac_ros_assets/models/dnn_stereo_disparity/dnn_stereo_disparity_v4.1.0_onnx/ess.engine`	ESS engine file path.
`--mode`	`depth`	Image extraction mode (‘raw’, ‘rectify’, or ‘depth’)

ROS Topics Subscribed

ROS Topic	Interface	Description
`/<camera>/left/image_compressed`	sensor_msgs/msg/CompressedImage	Left camera stream.
`/<camera>/left/camera_info`	sensor_msgs/msg/CameraInfo	Left camera intrinsics.
`/<camera>/right/image_compressed`	sensor_msgs/msg/CompressedImage	Right camera stream.
`/<camera>/right/camera_info`	sensor_msgs/msg/CameraInfo	Right camera intrinsics.

Usage

ros2 run isaac_ros_data_replayer rosbag_converter.py --input_rosbag <input_rosbag> --output_rosbag <output_rosbag> --all

ROS Run Arguments

ROS Run Argument	Default Value	Description
`-i`, `--input_rosbag`	None	Path to the input rosbag (recording from the `isaac_ros_data_recorder`).
`-o`, `--output_rosbag`	None	Path to the output (converted) rosbag.
`-a`, `--all`	None	Whether to convert all possible topics in the input rosbag. If provided, it takes priority over other arguments listing topics.
`--topics_decode_h264`	[]	List of image topics with sensor_msgs/msg/CompressedImage h264 messages to be decoded as sensor_msgs/msg/Image.
`--topics_udp_to_pc2`	[]	List of LIDAR topics with hesai_ros_driver/msg/UdpFrame packets to be converted to sensor_msgs/msg/PointCloud2.

ROS Topics Subscribed

ROS Topic	Interface	Description
`/<camera>/left/image_compressed`	sensor_msgs/msg/CompressedImage	Left camera stream.
`/<camera>/left/camera_info`	sensor_msgs/msg/CameraInfo	Left camera intrinsics.
`/<camera>/right/image_compressed`	sensor_msgs/msg/CompressedImage	Right camera stream.
`/<camera>/right/camera_info`	sensor_msgs/msg/CameraInfo	Right camera intrinsics.
`/<lidar>/lidar_packets`	hesai_ros_driver/msg/UdpFrame	LIDAR UDP packets.

isaac_ros_data_replayer

Overview

Quickstart

Set Up Development Environment

Download Quickstart Assets

Build isaac_ros_data_replayer

Run Launch File

Replay a single camera

Visualize

Data Extraction

Rosbag Conversion

Troubleshooting

Isaac ROS Troubleshooting

Not all camera frames extracted from recording

Not enough free disk space to run the rosbag_converter tool

Symptom

Solution

API

Usage

ROS Launch Arguments

ROS Topics Published

Usage

ROS Run Arguments

ROS Topics Subscribed

Usage

ROS Run Arguments

ROS Topics Subscribed

`isaac_ros_data_replayer`

Build `isaac_ros_data_replayer`

Not enough free disk space to run the `rosbag_converter` tool