/PPFMap

Point Pair Features are used for rigid object detection in point clouds

Primary LanguageC++

PPFMap

PPFMap is a parallel implementation of the Point Pair Feature matching algorithm from Drost, B.. The parallel implementation is as described in the SLAM++ project.

Requirements

To build this project you will need:

  • Point Cloud Library (PCL 1.7) : built with CUDA support.
  • Eigen library 3.0
  • CUDA 5.0
  • cmake 2.8

Compile and run

Before compiling the project, check first the cuda capability of your device. You can set the specific capability on the CMakeLists.txt file, in the CUDA_NVCC_FLAGS.

    mkdir build
    cd build
    cmake ..
    make
    ./demo

Demo preview

Documentation

Doxygen is used for the documentation. To generate it, simply execute the following in the build directory:

    make doc