ExtractAllSingleTrees

Rough Extraction of All individual Trees from a forest/orchard pointclouds.

Build

The current implementation depends on the Point Cloud Library (PCL) and its dependencies (e.g., Eigen and the Visualization Toolkit (VTK)). Please install PCL (and its dependencies) first.

To build this project, you need CMake (>= 3.12) and a compiler that supports >= C++17. With CMake, this project can be built on almost all platforms, although so far we have only tested it on Windows.

• On Windows with Microsoft Visual Studio, use the x64 Native Tools Command Prompt for VS XXXX (don't use the x86 one), then

  $ cd path-to-root-dir-of-this project
  $ mkdir Release
  $ cd Release
  $ cmake -G Ninja -DCMAKE_BUILD_TYPE=Release ..
  $ ninja
  

Usage

The main.cpp file demonstrates how to use it. To process raw forest or orchard point cloud data and extract individual trees, run the built executable like this:

./MakeDataset.exe <raw_data_folder> <rough_data_folder>

• <raw_data_folder> specifies the directory containing the original PCD files (e.g., LiDAR scans of forests).
• <rough_data_folder> specifies the output directory where the processed tree-extracted PCD files will be saved.
• This will process all .pcd files in the raw_data_folder directory, extract individual trees, and save the results in rough_data_folder.

References & Acknowledgments

This project is based on the original open-source project GlobalMatch. Many thanks to the original author for their valuable contributions!