Treelogdefectsegmentation

#Introduction Source code of the paper ICPR 2016 "Segmentation of defects on log surface from terrestrial Lidar data" #Online demo You can test the program [online] (http://ipol-geometry.loria.fr/~kerautre/ipol_demo/TDD_IPOLDemo/) #Installation This guide is to install the program on linux operation system. This may work for unix-like operation system like MacOs or BSD, but we did not test. It could be compiled and run in windows operating system, however some additional software should be installed like mingw or MS C++ compiler.

Dependencies

The program uses some C++ 11 feature, so we recommend the use of gcc 4.7 or later to compile. The program requires these libraries to be installed:

Boost 1.46 or later

DGtal version 0.9 or later. To install DGtal see [DGtal installation] (https://github.com/DGtal-team/DGtal/blob/master/README.md)

PCL version 1.7 or later

Eigen3

GNU GSL

CMake 2.6 or later

Install

mkdir build
cd build
cmake ..  -DDGtal_DIR=/path/to/DGtal
make

#User guide If you want you can also directly test the programm online:

======= User can find the example in the Main.cpp that include:

• Computation of centerline:

Centerline acc(scaledMesh, accRadius, trackStep, invertNormal);
std::vector<Z3i::RealPoint> fiber = acc.compute();

• Interpolation by spline

std::vector<Z3i::RealPoint> centerline = CenterlineHelper::getSmoothCenterlineBSplines(domain, fiber);

• Initialize the DefectSegmentation class:

DefectSegmentation sa(pointCloud, centerline, arcLength, patchHeight);
sa.init();

• get the point ids of defect

std::vector<unsigned int> defects = sa.getDefect();

• Export the results: Now you can use the helper class IOHelper to write the results to files. For example, to write the point ids of defect:

IOHelper::export2Text(defects, outputPrefix + "-defect.id");

• Run the program

  ./segmentation -i mesh.off --voxelSize 5 --accRadius \
  100 --trackStep 25 --patchWidth 25 --patchHeight 100 \
  --binWidth 5 --invertNormal true --output prefix