Camera Configuration

Camera Configuration

The pipeline supports two camera sources, selected at launch time via the input_source (or input_mode) argument.

Value	Camera
realsense (default)	Intel RealSense D-series
robot	Kinova arm-mounted vision camera

---

Intel RealSense

Launch the RealSense camera with aligned RGB-D and ordered point cloud:

ros2 launch realsense2_camera rs_launch.py \
  enable_rgbd:=true \
  enable_sync:=true \
  align_depth.enable:=true \
  enable_color:=true \
  enable_depth:=true \
  pointcloud.enable:=true \
  rgb_camera.color_profile:=640x480x30 \
  depth_module.depth_profile:=640x480x30 \
  pointcloud.ordered_pc:=true

The pointcloud.ordered_pc:=true flag is required for correct 3D pose extraction.

---

Kinova Arm Camera

Arm IP Addresses

Arm	IP
Left Arm	192.168.1.10
Right Arm	192.168.1.12

Terminal 1 — Start Zenoh Router

ros2 run rmw_zenoh_cpp rmw_zenohd

Terminal 2 — Launch Kinova Vision

export RMW_IMPLEMENTATION=rmw_zenoh_cpp
 
ros2 launch kinova_vision kinova_vision.launch.py \
  device:=192.168.1.12 \
  depth_registration:=true \
  color_camera_info_url:=file:///home/mohsin/r4s/src/ros2_kortex_vision/launch/calibration/default_color_calib_1280x720.ini \
  depth_camera_info_url:=file:///home/mohsin/r4s/src/ros2_kortex_vision/launch/calibration/default_depth_calib_480x270.ini

Resolution errors? Log into the arm's admin portal at 192.168.1.1X → Camera Settings → set calibration to 1280x720.

---

Topic Mapping by Camera Source

Each node automatically selects the correct point cloud topic based on input_mode:

input_mode	Point Cloud Topic
robot	/camera/depth/color/points
realsense	/camera/camera/depth/color/points