Segmentation

Tree and sub-tree segmentation algorithms

Background

Technological advancements allows forest inventory using individual trees as measurement units. However, the errors associated with ground delineation, tree identification, or estimation  of height to the base of the crown hinder the widespread application of spatial explicit tree-level forest inventories.

Objective

The projects are focused on individual tree segmentation or identification and measurement of parts of the tree, such as branches, knots, or leaves, use as input point clouds obtained from airborn sensors, terrestrial scanning, or photogrammetric techniques. The projects produced two major algorithms: one that identifies trees from ALS (called TrEx),and one that measure the stem from photogrammetric point clouds or lidar.

Products

TrEx is implemented as a stand alone software implemented in Java that can be download as a freeware. DOWNLOAD TrEx.

A Phuython implementation of TrEx, more functional and faster can be found on Github.

Both implmentations can be used under the  GNU General Public License

Publications