kinect2_bridge

Kinect2 Bridge

Maintainer

• Thiemo Wiedemeyer <[email protected]>, Institute for Artificial Intelligence, University of Bremen

Note: Please use the GitHub issues for questions and problems regarding the iai_kinect2 package and its components. Do not write emails.

Description

This is a bridge between libfreenect2 and ROS.

Highlights

• delivers up to 30 frames per second on non high end hardware
• delivers up to 30 frames per second over gigabit ethernet
• support for compressed image transport
• utilizes multiple cores and uses special OpenCL based implementation of the depth registration

Dependencies

• ROS Hydro/Indigo
• OpenCV
• libfreenect2

for the ROS packages look at the package.xml

First steps

For the depth registration the camera intrinsic and extrinsic parameters need to be known. The program reads in the values from the data/<serialnumber> folder. For each new sensor you need to add a sub-folder with the serial number of the device as the folder name. In this folder you need to provide 4 yaml files with the intrinsic and extrinsic parameters. These files can be created by the kinect2_calibration tool (or you can copy the files provided in one of the other folders, but results can be sub optimal). The device serial number is shown when kinect2_bridge or Protonect from libfreenect2 is started, it also appears in dmesg when you connect the sensor. More information on calibration.

When kinect2_bridge is running you can use the kinect2_viewer to display the images or point cloud: rosrun kinect2_viewer kinect2_viewer sd image or rosrun kinect2_viewer kinect2_viewer sd cloud.

Topics

HD Topics

The images in this topics have a FullHD resolution (1920x1080).

Note: For correct registration of the depth image to the color image it is needed to perform a calibration.

/kinect2/hd/camera_info
/kinect2/hd/image_color
/kinect2/hd/image_color/compressed
/kinect2/hd/image_color_rect
/kinect2/hd/image_color_rect/compressed
/kinect2/hd/image_depth_rect
/kinect2/hd/image_depth_rect/compressed
/kinect2/hd/image_mono
/kinect2/hd/image_mono/compressed
/kinect2/hd/image_mono_rect
/kinect2/hd/image_mono_rect/compressed
/kinect2/hd/points

Quater HD Topics

The images in this topics have a quarter FullHD resolution (960x540).

Note: For correct registration of the depth image to the color image it is needed to perform a calibration.

/kinect2/qhd/camera_info
/kinect2/qhd/image_color
/kinect2/qhd/image_color/compressed
/kinect2/qhd/image_color_rect
/kinect2/qhd/image_color_rect/compressed
/kinect2/qhd/image_depth_rect
/kinect2/qhd/image_depth_rect/compressed
/kinect2/qhd/image_mono
/kinect2/qhd/image_mono/compressed
/kinect2/qhd/image_mono_rect
/kinect2/qhd/image_mono_rect/compressed
/kinect2/qhd/points

IR/Depth Topics

This are the raw IR and depth images from the sensor (512x424).

Note: The registration of the color image is available without a calibration. Parameters for the registration are provided by the sensor itself.

/kinect2/sd/camera_info
/kinect2/sd/image_color_rect
/kinect2/sd/image_color_rect/compressed
/kinect2/sd/image_depth
/kinect2/sd/image_depth/compressed
/kinect2/sd/image_depth_rect
/kinect2/sd/image_depth_rect/compressed
/kinect2/sd/image_ir
/kinect2/sd/image_ir/compressed
/kinect2/sd/image_ir_rect
/kinect2/sd/image_ir_rect/compressed
/kinect2/sd/points

Notes

• Point clouds are only published when the launch file is used. Run roslaunch kinect2_bridge kinect2_bridge.launch.
• Images from the same frame have the same timestamp. Using the message_filters::sync_policies::ExactTime policy is recommended.

Usage

roslaunch kinect2_bridge kinect2_bridge.launch [options:=value]
base_name:=<string>
    default: kinect2
    info:    set base name for all topics
sensor:=<string>
    default:
    info:    serial of the sensor to use
fps_limit:=<double>
    default: -1.0
    info:    limit the frames per second
calib_path:=<string>
    default: /home/wiedemeyer/work/src/iai_kinect2/kinect2_bridge/data/
    info:    path to the calibration files
use_png:=<bool>
    default: false
    info:    Use PNG compression instead of TIFF
jpeg_quality:=<int>
    default: 90
    info:    JPEG quality level from 0 to 100
png_level:=<int>
    default: 1
    info:    PNG compression level from 0 to 9
depth_method:=<string>
    default: cuda
    info:    Use specific depth processing: default, cpu, opengl, opencl, cuda, clkde, cudakde
depth_device:=<int>
    default: -1
    info:    openCL device to use for depth processing
reg_method:=<string>
    default: opencl
    info:    Use specific depth registration: default, cpu, opencl
reg_device:=<int>
    default: -1
    info:    openCL device to use for depth registration
max_depth:=<double>
    default: 12.0
    info:    max depth value
min_depth:=<double>
    default: 0.1
    info:    min depth value
queue_size:=<int>
    default: 2
    info:    queue size of publisher
bilateral_filter:=<bool>
    default: true
    info:    enable bilateral filtering of depth images
edge_aware_filter:=<bool>
    default: true
    info:    enable edge aware filtering of depth images
publish_tf:=<bool>
    default: false
    info:    publish static tf transforms for camera
base_name_tf:=<string>
    default: as base_name
    info:    base name for the tf frames
worker_threads:=<int>
    default: 4
    info:    number of threads used for processing the images

Key bindings

Terminal:

• CRTL+c: Quit