VI-Sensor Simulator

The simulation of the VI-Sensor. No Guarantees.

This work is described in the letter "Aerial Single-View Depth Completion with Image-Guided Uncertainty Estimation", by Lucas Teixeira, Martin R. Oswald, Marc Pollefeys, Margarita Chli, published in the IEEE Robotics and Automation Letters (RA-L) IEEE link.

Video:

Citations:

If you use this code for research, please cite the following publication:

@article{Teixeira:etal:RAL2020,
    title   = {{Aerial Single-View Depth Completion with Image-Guided Uncertainty Estimation}},
    author  = {Lucas Teixeira and Martin R. Oswald and Marc Pollefeys and Margarita Chli},
    journal = {{IEEE} Robotics and Automation Letters ({RA-L})},
    doi     = {10.1109/LRA.2020.2967296},
    year    = {2020}
}

License

There is copyright code from other libraries. Mostly commented in the source code. For our original code BSD license applies.

Installation

• Requirements:

  • ROS Melodic
  • Blender 2.80 or newer- (https://askubuntu.com/questions/110821/how-to-install-blender-from-the-official-website)

• Initialize catkin workspace (using catkin tools):

  $ mkdir -p ~/catkin_ws/src
  $ cd ~/catkin_ws  
  $ source /opt/ros/melodic/setup.bash
  $ catkin init  # initialize your catkin workspace
  $ catkin config --cmake-args -DCMAKE_BUILD_TYPE=Release
  $ catkin config --merge-devel

• Get the simulator and dependencies

  $ cd ~/catkin_ws/src
  $ sudo apt-get install liblapacke-dev python-wstool python-catkin-tools protobuf-compiler libgoogle-glog-dev libopenexr-dev libatlas-base-dev libeigen3-dev libsuitesparse-dev
  $ sudo apt-get install ros-melodic-joy ros-melodic-octomap-ros
  $ git clone [email protected]:catkin/catkin_simple
  $ git clone [email protected]:ethz-asl/rotors_simulator
  $ git clone [email protected]:ethz-asl/mav_comm
  $ git clone [email protected]:ethz-asl/eigen_catkin
  $ git clone [email protected]:ethz-asl/glog_catkin
  $ git clone [email protected]:ethz-asl/mav_control_rw
  $ pip install OpenEXR
  
  $ git clone [email protected]:VIS4ROB-lab/visensor_simulator.git -b blender_2.8  

• Build the workspace

  $ catkin build

• Open your Blender, navigate to Edit > Preferences > Add-ons, and import the script "visensor_simulator/scripts/blender/visensor_sim_blender_addon.py" as an add-on by clicking on "Install..." (https://docs.blender.org/manual/en/latest/editors/preferences/addons.html)

Step-by-step

1. Create a project - this is a folder with any name. Configure the cameras and the waypoints In the folder resources there is a example "project_test.tar"

2. starts gazebo

  $ roslaunch visensor_simulator uav_vi_blender.launch

3. run the back_end:

  $ roslaunch visensor_simulator ros_backend.launch project_folder:="/home/lucas/data/test/project_testA"

4. open your scene on blender, select the camera, file->import->VISensor Simulator Project(*json), choose the file visim_project.json on your project.

5. Render. Quick render is faster, but it is less realistic.

6. run the bagcreator(namespace is optional):

  $ rosrun visensor_simulator visensor_sim_bagcreator.py --output_bag your_output.bag --project_folder "/home/lucas/data/test/project_testA" --namespace "firefly"

Roadmap

• write a camera path exporter compatible with our waypoint planner, better if we introduce this on our dataset format
• change from firefly to neo11
• develop a software to build a simplified version of the world to allow collision on the simulation. BVH and Octomap are options
• better error msg on the bagcreator when it is not possible to create the bagfile
• expose the simple_planner waypoint tolerance as ros parameters
• write my owm spawn with noise and vi_sensor pose as parameters
• add imu name on the json file
• autoselect a camera from the json
• add option to disable the simple planner
• detect incomplete render sequence and jump to the latest one (support in case of shutdown)
• make everything relative to the project file instead of the project folder
• add test for topic names on the bagcreator