docker

ifm3d - Docker

NOTE: This work with Docker is currently experimental. The process will be refined over time and is subject to change without notice. Thanks for your patience.

As of ifm3d version 0.12.0, we are using Docker as a means to build binary packages of the software, for Linux, to help ease installation for our users. We use Docker in two ways:

1. To build the software packages (deb files)
2. To provide a quick and easy way to test those debs on a fresh OS installation.

Whether you plan to use Docker for building, running, or both, you will need to install Docker. There are two steps to getting Docker setup correctly:

1. Install Docker using these instructions
2. Carry out the post installation setup using these instructions

Building

This section is only for ifm3d developers. Users can safely skip this part.

To use Docker to build the ifm3d software packages, you should utilize the build-debs.py script (requires Python 3). The help output for the script is as follows:

$ ./build-debs.py --help
usage: build-debs.py [-h] [--docker_file DOCKER_FILE [DOCKER_FILE ...]]
                     [--image IMAGE] [--vtag VTAG] [--run_tests]

Build ifm3d debs via Docker

optional arguments:
  -h, --help            show this help message and exit
  --docker_file DOCKER_FILE [DOCKER_FILE ...]
                        DockerFile list to process. If not specified, all
                        `DockerFile' specifications encountered in all
                        subirectories (with respect to your current working
                        directory), recursively, will be processed.
  --image IMAGE         Docker image name.
  --vtag VTAG           ifm3d git version tag to build. If omitted, `master'
                        will be checked out and built.
  --run_tests           Run unit tests during build. Only specify this flag if
                        testing against live hardware (e.g., O3D or O3X).

Depending upon how you parameterize build-debs.py, the built packages will be placed next to the Dockerfile in a subdirectory called debs. These debs can then be uploaded to ifm's Nexus server for easy install via apt-get.

Running

If you just want to quickly evaluate an ifm 3D camera via ifm3d, you can bring up a Docker container appropriate to your situation. We maintain a few to make this process easy.

Non-ROS users

Dockerfiles to build can be found here.

Ubuntu 16.04

Be sure you are in the 16.04 directory.

Build the Docker container:

$ docker build -t ifm3d1604 .

Once the container is done being built, if you simply want to inspect your connected camera via the ifm3d command line tool, you can:

$ docker run -u ifm -ti --rm ifm3d1604:latest /bin/bash
ifm@c8c9865efb14:/$ ifm3d ls
[
  {
    "Active": true,
    "Description": "",
    "Id": 351437496,
    "Index": 1,
    "Name": "Sample Application"
  }
]
ifm@c8c9865efb14:/$ ifm3d hz
FrameGrabber running at: 4.95815 Hz
10 frames captured, over 1 runs
ifm@c8c9865efb14:/$ ^D

If you want to visualize the data using the ifm3d-pcl-viewer:

NOTE: The xhost command below is considered insecure. For exemplary purposes, we use this approach in these instructions. However, we urge you to read the man page for xhost and decide if this is a security concern for you before running the commands below. You have been warned.

$ xhost +local:root
non-network local connections being added to access control list

$ docker run -u ifm -ti --rm -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix:rw ifm3d1604:latest /bin/bash
ifm@b79de0545832:/$ ifm3d-pcl-viewer

If your camera is connected at the default ip address of 192.168.0.69, you should now see a window rendering the streaming pointcloud.

Ubuntu 18.04

Be sure you are in the 18.04 directory.

Build the Docker container:

$ docker build -t ifm3d1804 .

Use the command line tools:

$ docker run -u ifm -ti --rm ifm3d1804:latest /bin/bash
ifm@330fc18969c4:/$ ifm3d ls
[
  {
    "Active": true,
    "Description": "",
    "Id": 351437496,
    "Index": 1,
    "Name": "Sample Application"
  }
]
ifm@330fc18969c4:/$ ifm3d hz
FrameGrabber running at: 5.33508 Hz
10 frames captured, over 1 runs
ifm@330fc18969c4:/$ ^D

Use the visualizer:

$ xhost +local:root
non-network local connections being added to access control list

$ docker run -u ifm -ti --rm -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix:rw ifm3d1804:latest /bin/bash
ifm@f98b659d1e71:/$ ifm3d-pcl-viewer

ROS users

Ubuntu 16.04 - Kinetic

Be sure you are in the 16.04 ROS directory.

Build the Docker container:

$ docker build -t ifm3d-1604-kinetic .

Assuming you have an ifm 3D camera plugged in (at the default IP -- modify as necessary), you can use the container like a standard ROS machine. To view the point cloud in rviz:

$ xhost +local:root
non-network local connections being added to access control list

$ docker run -u ifm -ti --rm -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix:rw ifm3d-1604-kinetic:latest /bin/bash
ifm@4aa58228052f:~$ roslaunch ifm3d camera.launch >/dev/null 2>&1 &
[1] 112
ifm@4aa58228052f:~$ roslaunch ifm3d rviz.launch

If your camera is plugged in, you should now see an rviz window streaming data from it. If your camera is at a different ip address than the default, you should pass that as the ip:=... argument to the camera.launch file.

Ubuntu 18.04 - Melodic

Be sure you are in the 18.04 ROS directory.

Build the Docker container:

$ docker build -t ifm3d-1804-melodic .

Assuming you have an ifm 3D camera plugged in (at the default IP -- modify as necessary), you can use the container like a standard ROS machine. To view the point cloud in rviz:

$ xhost +local:root
non-network local connections being added to access control list

$ docker run -u ifm -ti --rm -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix:rw ifm3d-1804-melodic:latest /bin/bash
ifm@4aa58228052f:~$ roslaunch ifm3d camera.launch >/dev/null 2>&1 &
[1] 112
ifm@4aa58228052f:~$ roslaunch ifm3d rviz.launch