search

luxonis / depthai-ros

Official ROS Driver for DepthAI Sensors.
MIT License
239 stars 173 forks source link

[BUG] OAK D PRO POE can only get 720P. v2.8.2-humble #463

Open Victorsoeby opened 7 months ago

Victorsoeby commented 7 months ago

Expected result: Able to change resolution to "12MP", "4K", "1080P" Result: Can only get 720P Branch: v2.8.2-humble-12-g48fd970. Launchfile: camera.launch.py (depthai_ros_driver)

I am using this with a OAK D Pro POE camera with the IMX378 sensor, and can for the life of me not get it to output the RGB image in 1080P. I have cloned the V2.8.2, changed the camera.yaml file to have the following:

/oak:
  ros__parameters:
    camera:
      i_enable_imu: true
      i_enable_ir: true
      i_pipeline_type: RGBD
      i_nn_type: ''
    rgb: 
      i_resolution: '1080P'

When launching the camera.launch.py: 

root@agx-orin2:/workspaces# ros2 launch depthai_ros_driver camera.launch.py 
[INFO] [launch]: All log files can be found below /root/.ros/log/2023-11-23-14-02-21-228962-agx-orin2-46251
[INFO] [launch]: Default logging verbosity is set to INFO
[INFO] [component_container-1]: process started with pid [46265]
[component_container-1] [INFO] [1700748142.526463430] [oak_container]: Load Library: /opt/ros/humble/lib/librobot_state_publisher_node.so
[component_container-1] [INFO] [1700748142.532228582] [oak_container]: Found class: rclcpp_components::NodeFactoryTemplate<robot_state_publisher::RobotStatePublisher>
[component_container-1] [INFO] [1700748142.532365673] [oak_container]: Instantiate class: rclcpp_components::NodeFactoryTemplate<robot_state_publisher::RobotStatePublisher>
[component_container-1] [INFO] [1700748142.557751327] [oak_state_publisher]: got segment ctis_camera_link
[component_container-1] [INFO] [1700748142.557883842] [oak_state_publisher]: got segment oak
[component_container-1] [INFO] [1700748142.557931811] [oak_state_publisher]: got segment oak_imu_frame
[component_container-1] [INFO] [1700748142.557974340] [oak_state_publisher]: got segment oak_left_camera_frame
[component_container-1] [INFO] [1700748142.558015269] [oak_state_publisher]: got segment oak_left_camera_optical_frame
[component_container-1] [INFO] [1700748142.558053254] [oak_state_publisher]: got segment oak_model_origin
[component_container-1] [INFO] [1700748142.558089990] [oak_state_publisher]: got segment oak_rgb_camera_frame
[component_container-1] [INFO] [1700748142.558125543] [oak_state_publisher]: got segment oak_rgb_camera_optical_frame
[component_container-1] [INFO] [1700748142.558162888] [oak_state_publisher]: got segment oak_right_camera_frame
[component_container-1] [INFO] [1700748142.558199529] [oak_state_publisher]: got segment oak_right_camera_optical_frame
[INFO] [launch_ros.actions.load_composable_nodes]: Loaded node '/oak_state_publisher' in container 'oak_container'
[component_container-1] [INFO] [1700748142.562600043] [oak_container]: Load Library: /workspaces/install/depthai_ros_driver/lib/libdepthai_ros_driver.so
[component_container-1] [INFO] [1700748142.664661165] [oak_container]: Found class: rclcpp_components::NodeFactoryTemplate<depthai_ros_driver::Camera>
[component_container-1] [INFO] [1700748142.664867602] [oak_container]: Instantiate class: rclcpp_components::NodeFactoryTemplate<depthai_ros_driver::Camera>
[component_container-1] [INFO] [1700748142.680244008] [oak]: No ip/mxid specified, connecting to the next available device.
[component_container-1] [INFO] [1700748149.563710927] [oak]: Camera with MXID: 1844301071F88E0F00 and Name: 10.46.28.215 connected!
[component_container-1] [INFO] [1700748149.564009910] [oak]: PoE camera detected. Consider enabling low bandwidth for specific image topics (see readme).
[component_container-1] [INFO] [1700748149.588918081] [oak]: Device type: OAK-D-PRO-POE-FF
[component_container-1] [INFO] [1700748149.597741125] [oak]: Pipeline type: RGBD
[component_container-1] [INFO] [1700748149.883668521] [oak]: Finished setting up pipeline.
[component_container-1] [INFO] [1700748150.994838049] [oak]: Camera ready!
[INFO] [launch_ros.actions.load_composable_nodes]: Loaded node '/oak' in container '/oak_container'

All seems to go well, but the image that is published is 1280x720. Here is me doing a ros2 topic echo /oak/rgb/image_raw image

When i try to change the resolution to 4MP in the camera.yaml config file i get this: 

root@agx-orin2:/workspaces# ros2 launch depthai_ros_driver camera.launch.py 
[INFO] [launch]: All log files can be found below /root/.ros/log/2023-11-23-14-14-20-489431-agx-orin2-50833
[INFO] [launch]: Default logging verbosity is set to INFO
[INFO] [component_container-1]: process started with pid [50847]
[component_container-1] [INFO] [1700748861.767367651] [oak_container]: Load Library: /opt/ros/humble/lib/librobot_state_publisher_node.so
[component_container-1] [INFO] [1700748861.771652002] [oak_container]: Found class: rclcpp_components::NodeFactoryTemplate<robot_state_publisher::RobotStatePublisher>
[component_container-1] [INFO] [1700748861.771843334] [oak_container]: Instantiate class: rclcpp_components::NodeFactoryTemplate<robot_state_publisher::RobotStatePublisher>
[component_container-1] [INFO] [1700748861.788308182] [oak_state_publisher]: got segment oak
[component_container-1] [INFO] [1700748861.788446137] [oak_state_publisher]: got segment oak-d-base-frame
[component_container-1] [INFO] [1700748861.788497114] [oak_state_publisher]: got segment oak_imu_frame
[component_container-1] [INFO] [1700748861.788539579] [oak_state_publisher]: got segment oak_left_camera_frame
[component_container-1] [INFO] [1700748861.788579388] [oak_state_publisher]: got segment oak_left_camera_optical_frame
[component_container-1] [INFO] [1700748861.788617853] [oak_state_publisher]: got segment oak_model_origin
[component_container-1] [INFO] [1700748861.788655197] [oak_state_publisher]: got segment oak_rgb_camera_frame
[component_container-1] [INFO] [1700748861.788693214] [oak_state_publisher]: got segment oak_rgb_camera_optical_frame
[component_container-1] [INFO] [1700748861.788731935] [oak_state_publisher]: got segment oak_right_camera_frame
[INFO] [launch_ros.actions.load_composable_nodes]: Loaded node '/oak_state_publisher' in container 'oak_container'
[component_container-1] [INFO] [1700748861.788771744] [oak_state_publisher]: got segment oak_right_camera_optical_frame
[component_container-1] [INFO] [1700748861.792288142] [oak_container]: Load Library: /workspaces/install/depthai_ros_driver/lib/libdepthai_ros_driver.so
[component_container-1] [INFO] [1700748861.887910275] [oak_container]: Found class: rclcpp_components::NodeFactoryTemplate<depthai_ros_driver::Camera>
[component_container-1] [INFO] [1700748861.888111528] [oak_container]: Instantiate class: rclcpp_components::NodeFactoryTemplate<depthai_ros_driver::Camera>
[component_container-1] [INFO] [1700748861.902331045] [oak]: No ip/mxid specified, connecting to the next available device.
[component_container-1] [INFO] [1700748868.791367718] [oak]: Camera with MXID: 1844301071F88E0F00 and Name: 10.46.28.215 connected!
[component_container-1] [INFO] [1700748868.791592395] [oak]: PoE camera detected. Consider enabling low bandwidth for specific image topics (see readme).
[component_container-1] [INFO] [1700748868.816391796] [oak]: Device type: OAK-D-PRO-POE-FF
[component_container-1] [INFO] [1700748868.825218361] [oak]: Pipeline type: RGBD
[component_container-1] [WARN] [1700748868.838168090] [oak]: Resolution 4MP not supported by sensor IMX378. Using default resolution 1080P
[component_container-1] [INFO] [1700748869.108435399] [oak]: Finished setting up pipeline.
[component_container-1] [INFO] [1700748870.097634423] [oak]: Camera ready!
[INFO] [launch_ros.actions.load_composable_nodes]: Loaded node '/oak' in container '/oak_container'

But still only 720P when i check the images.

I am using the complete default version of the branch, here is my camera.launch.py file: 

import os

from ament_index_python.packages import get_package_share_directory
from launch import LaunchDescription
from launch.actions import DeclareLaunchArgument, IncludeLaunchDescription, OpaqueFunction
from launch.launch_description_sources import PythonLaunchDescriptionSource
from launch.substitutions import LaunchConfiguration
from launch.conditions import IfCondition
from launch_ros.actions import ComposableNodeContainer, Node
from launch_ros.descriptions import ComposableNode

def launch_setup(context, *args, **kwargs):
    log_level = 'info'
    if(context.environment.get('DEPTHAI_DEBUG')=='1'):
        log_level='debug'

    urdf_launch_dir = os.path.join(get_package_share_directory('depthai_descriptions'), 'launch')

    params_file = LaunchConfiguration("params_file")
    camera_model = LaunchConfiguration('camera_model',  default = 'OAK-D-PRO')

    name = LaunchConfiguration('name').perform(context)

    parent_frame = LaunchConfiguration('parent_frame',  default = 'oak-d-base-frame')
    cam_pos_x    = LaunchConfiguration('cam_pos_x',     default = '0.0')
    cam_pos_y    = LaunchConfiguration('cam_pos_y',     default = '0.0')
    cam_pos_z    = LaunchConfiguration('cam_pos_z',     default = '0.0')
    cam_roll     = LaunchConfiguration('cam_roll',      default = '0.0')
    cam_pitch    = LaunchConfiguration('cam_pitch',     default = '0.0')
    cam_yaw      = LaunchConfiguration('cam_yaw',       default = '0.0')
    use_composition = LaunchConfiguration('rsp_use_composition', default='true')
    imu_from_descr = LaunchConfiguration('imu_from_descr', default='false')
    pass_tf_args_as_params = LaunchConfiguration('pass_tf_args_as_params', default='false')
    override_cam_model = LaunchConfiguration('override_cam_model', default='false')

    tf_params = {}
    if(pass_tf_args_as_params.perform(context) == 'true'):
        cam_model = ''
        if override_cam_model.perform(context) == 'true':
            cam_model = camera_model.perform(context)
        tf_params = {'camera': {
            'i_publish_tf_from_calibration': True,
            'i_tf_tf_prefix': name,
            'i_tf_camera_model': cam_model,
            'i_tf_base_frame': name,
            'i_tf_parent_frame': parent_frame.perform(context),
            'i_tf_cam_pos_x': cam_pos_x.perform(context),
            'i_tf_cam_pos_y': cam_pos_y.perform(context),
            'i_tf_cam_pos_z': cam_pos_z.perform(context),
            'i_tf_cam_roll': cam_roll.perform(context),
            'i_tf_cam_pitch': cam_pitch.perform(context),
            'i_tf_cam_yaw': cam_yaw.perform(context),
            'i_tf_imu_from_descr': imu_from_descr.perform(context),
        }
        }

    use_gdb      = LaunchConfiguration('use_gdb',       default = 'false')
    use_valgrind = LaunchConfiguration('use_valgrind',  default = 'false')
    use_perf     = LaunchConfiguration('use_perf',      default = 'false')

    launch_prefix = ''

    if (use_gdb.perform(context) == 'true'):
        launch_prefix += "gdb -ex run --args "
    if (use_valgrind.perform(context) == 'true'):
        launch_prefix += "valgrind --tool=callgrind"
    if (use_perf.perform(context) == 'true'):
        launch_prefix += "perf record -g --call-graph dwarf --output=perf.out.node_name.data --"
    return [
            Node(
                condition=IfCondition(LaunchConfiguration("use_rviz").perform(context)),
                package="rviz2",
                executable="rviz2",
                name="rviz2",
                output="log",
                arguments=["-d", LaunchConfiguration("rviz_config")],
            ),
        IncludeLaunchDescription(
            PythonLaunchDescriptionSource(
                os.path.join(urdf_launch_dir, 'urdf_launch.py')),
            launch_arguments={'tf_prefix': name,
                              'camera_model': camera_model,
                              'base_frame': name,
                              'parent_frame': parent_frame,
                              'cam_pos_x': cam_pos_x,
                              'cam_pos_y': cam_pos_y,
                              'cam_pos_z': cam_pos_z,
                              'cam_roll': cam_roll,
                              'cam_pitch': cam_pitch,
                              'cam_yaw': cam_yaw,
                              'use_composition': use_composition,
                              'use_base_descr': pass_tf_args_as_params}.items()),

        ComposableNodeContainer(
            name=name+"_container",
            namespace="",
            package="rclcpp_components",
            executable="component_container",
            composable_node_descriptions=[
                    ComposableNode(
                        package="depthai_ros_driver",
                        plugin="depthai_ros_driver::Camera",
                        name=name,
                        parameters=[params_file, tf_params],
                    )
            ],
            arguments=['--ros-args', '--log-level', log_level],
            prefix=[launch_prefix],
            output="both",
        ),

    ]

def generate_launch_description():
    depthai_prefix = get_package_share_directory("depthai_ros_driver")

    declared_arguments = [
        DeclareLaunchArgument("name", default_value="oak"),
        DeclareLaunchArgument("parent_frame", default_value="oak-d-base-frame"),
        DeclareLaunchArgument("camera_model", default_value="OAK-D-Pro"),
        DeclareLaunchArgument("cam_pos_x", default_value="0.0"),
        DeclareLaunchArgument("cam_pos_y", default_value="0.0"),
        DeclareLaunchArgument("cam_pos_z", default_value="0.0"),
        DeclareLaunchArgument("cam_roll", default_value="0.0"),
        DeclareLaunchArgument("cam_pitch", default_value="0.0"),
        DeclareLaunchArgument("cam_yaw", default_value="0.0"),
        DeclareLaunchArgument("params_file", default_value=os.path.join(depthai_prefix, 'config', 'camera.yaml')),
        DeclareLaunchArgument("use_rviz", default_value='false'),
        DeclareLaunchArgument("rviz_config", default_value=os.path.join(depthai_prefix, "config", "rviz", "rgbd.rviz")),
        DeclareLaunchArgument("rsp_use_composition", default_value='true'),
        DeclareLaunchArgument("pass_tf_args_as_params", default_value='false', description='Enables TF publishing from camera calibration file.'),
        DeclareLaunchArgument("imu_from_descr", default_value='false', description='Enables IMU publishing from URDF.'),
        DeclareLaunchArgument("override_cam_model", default_value='false', description='Overrides camera model from calibration file.'),
        DeclareLaunchArgument("use_gdb", default_value='false'),
        DeclareLaunchArgument("use_valgrind", default_value='false'),
        DeclareLaunchArgument("use_perf", default_value='false')
    ]

    return LaunchDescription(
        declared_arguments + [OpaqueFunction(function=launch_setup)]
    )

I know it is not hardware related, as the drivers have been working fine when i used the Humble 2.7.1 branch earlier. Reason i updated was that i found a bug where the timestamps of the images were wrong, which seems to have been fixed in this version.

Serafadam commented 7 months ago

Hi, by default ISP output is scaled when running with default parameters, you can set rgb.i_set_isp_scale: false to turn that off.

RMonica commented 3 weeks ago

Hi,

I ran into this problem yesterday. It was confusing also because the camera info changes resolution according to i_height and i_width, but the actual image doesn't.

The link is broken, by the way.

Serafadam commented 3 weeks ago

Hi, the output image depends on DAI node output type, you can have isp or video, by default it is ISP which will get scaled, setting i_width and i_height sets video outputs size. Updated link here