Sparse Scene Flow

This repository contains code for sparse scene flow estimation using stereo cameras, proposed by P. Lenz etal.: Sparse Scene Flow Segmentation for Moving Object Detection in Urban Environments, Intelligent Vehicles Symposium (IV), 2011. This method can be used as a component in your visual object tracking / 3D reconstruction / SLAM applications as an alternative to dense (and typically expensive to compute) scene flow methods.

Note: The repository contains scene flow estimator only, there is no implementation for scene flow clustering or object tracking provided in this repository.

If you want to know what is the difference between scene and optical flow, see this quora thread.

Demo Video

Click here to watch the video.

Prerequisite

In order to run the code, your setup has to meet the following minimum requirements (tested versions in parentheses. Other versions might work, too):

• GCC 4.8.4
  • Eigen (3.x)
  • pybind11

Install

Compiling the code using CMake

0. mkdir build
1. cmake ..
2. make all

Running the sparse flow app

0. Download KITTI
1. See python/python_example.py to see how to use visual odometry estimator

Remarks

• External libraries

  • The tracker ships the following external modules:
    • libviso2 - egomotion estimation, feature matching (http://www.cvlibs.net/software/libviso/)

• For optimal performance, run the sf-estimator in release mode.

UPDATE (Jan'20): I added bindings for python and removed most of the "old" exmaple code in order to shrink the dependencies to the minimum. See the python example.

If you have any issues or questions about the code, please contact me https://www.vision.rwth-aachen.de/person/13/

Citing

If you find this code useful in your research, you should cite:

@inproceedings{Lenz2011IV,
  author = {Philip Lenz and Julius Ziegler and Andreas Geiger and Martin Roser},
  title = {Sparse Scene Flow Segmentation for Moving Object Detection in Urban Environments},
  booktitle = {Intelligent Vehicles Symposium (IV)},
  year = {2011}
}

License

GNU General Public License (http://www.gnu.org/licenses/gpl.html)

Copyright (c) 2017 Aljosa Osep Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.