sosswald / gpu-coverage

GPU-accelerated next-best-view coverage of articulated scenes

GPU-Accelerated Next-Best-View Coverage of Articulated Scenes

This repository contains the code accompanying our paper on GPU-accelerated coverage of articulated scenes.

Disclaimer: The code in this repository is research code created for a feasibility study. The code is not ready for production and the authors will not provide support. Use at your own risk.

For details on the approach see our paper:

Stefan Oßwald and Maren Bennewitz: GPU-Accelerated Next-Best-View Coverage of Articulated Scenes. Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018.

License

The code is released under the 3-clause BSD license. See LICENSE for details.

Requirements

• An OpenGL implementation and GPU hardware supporting at least OpenGL ES 3.2 or OpenGL Core 3.3 profiles.
• An EGL implementation for headless access to the GPU.
• Assimp, the Open Asset Importer Library.
• GLFW3 for handling OpenGL windows, contexts, and surfaces.
• GLM, the OpenGL mathematics library.
• CMake is used as the build system.
• Optional: Catkin for ROS package building support.
• Optional: OpenCV for loading image textures, both versions 2.4 and 3.0 are supported.
• Optional: NVLM, the NVIDIA management library for handling GPUs on multi-GPU servers

Compiling

Compile the code using either CMake or Catkin:

mkdir build
cd build
cmake .. [-D FLAG=VALUE ...]
make

Compile flags:

Flag	Default value	Description
OPENGL_API	OPENGL_API	OpenGL API to use. Set this either to OPENGL_API for the core API or to OPENGL_ES_API for the embedded systems API.
OPENGL_MAJOR	4	Major version of the OpenGL or OpenGL ES API
OPENGL_MINOR	4	Minor version of the OpenGL or OpenGL ES API
USE_XFB	ON	Boolean flag, use transfer buffer (XFB) variant of the Bellman-Ford implementation

API Documentation

See the Doxygen documentation for an overview of the source code and API.