PyTLidar: A Python Package for Tree QSM Modeling from Terrestrial Lidar Data

John McDonald Hagood , Fan Yang, Shruti Motiwale, Breanna Shi, Jeffery Cannon, James T. Stroud

PyTLidar is an open-source Python package that reconstructs 3D tree Quantitative Structure Models (QSM) from Terrestrial Lidar Scanning (TLS) data, providing a user-friendly tool that improves and expands upon the MATLAB-based TreeQSM method. QSMs are used to automate detection and calculation of various topological and volumetric measurements that would normally take great effort to gather in the field. PyTLidar provides an accessible, extensible, and GUI-driven workflow for researchers and practitioners in forestry, ecology, and 3D vegetation modeling to create QSMs. The package also integrates interactive visualization tools for inspecting model quality and viewing calculated tree properties. The ease of use and installation of PyTLidar provides ecologists an option to gather forest measurements that does not rely on proprietary tools.

The key features of PyTLidar are a reproduction of the TreeQSM core functionality, and enhancing the experience of setting up experiments and viewing results. It provides functionality for loading and extracting point cloud data from .las and .laz files as well as automatic calculation of a range of initial parameters for the QSM model based on point cloud structure. The QSM creation methods include generation of a Voronoi partition of the point cloud, segment detection, detection of parent-child relationships of branches, and cylinder fitting. PyTLidar also calculates various tree metrics such as branch length and volume and provides these results in text format as well as visual graphics. QSMs are also output in an interactive format which can be viewed directly in the GUI as well as saved and shared as an html file. All of this is packaged within a user-friendly GUI while also providing support for command line and direct Python interfacing.