Point cloud transformation with coherent nearest neighbours.

• Install the openni-git package from the AUR.
• Install the pcl with all dependencies from the AUR. Make sure that openni-git is found.
• Install packages primesense-nite2 and sensorkinect-git.

In order to execute the examples, the kernel module gspca_kinect must be removed. The module can be disabled for one session with:

$ sudo modprobe -r gspca_kinect

To disable the module permanently, add an entry to modprobe.d:

$ sudo vim /etc/modprobe.d/disable_gspca_kinect

with the content

blacklist gspca_kinect

This setup has been tested on a Ubuntu 14.04 AMD64 VM. Unfortunately, the installation of the libpcl-all package according to the instructions on the PCL homepage resulted in unresolved problems. Our workaround was to compile the PCL sources ourself: https://github.com/PointCloudLibrary/pcl/archive/pcl-1.7.1.tar.gz

• Install the packages:

cmake g++ git libboost-all-dev libflann-dev libeigen3-dev libopenni-dev  libusb-1.0 libvtk5-dev libvtk-java libvtk5-qt4-dev tcl-vtk libvtk-java python-vtk libqhull-dev libopencv-dev libopenni-sensor-primesense0

• Compile the sources with mkdir build, cd build and

cmake ../ \
-DBUILD_visualization=ON \
-DBUILD_app_cloud_composer=ON \
-DBUILD_app_in_hand_scanner=ON \
-DBUILD_app_modeler=ON \
-DBUILD_app_point_cloud_editor=ON \

• The actual building takes some while and needs a lot of memory: make -j 5 if more than 5 GB RAM available. make otherwise
• Install with make install

Our project currently consists of several small sample demos which can be used to test various PCL components. Running $ cmake . in the top-level directory will create a Makefile for every subproject. You can compile a subproject by running $ make in the corresponding subdirectory. Note: Compile times are horrible as of right now, because of the excessive C++ templating. We're currently looking for ways to resolve this issue.

The RGB-D Object Dataset. A collection of various household object models. Available for non-commercial reseach/educational use: http://rgbd-dataset.cs.washington.edu/dataset.html