Tree segmentation from aerial point clouds
Forest inventory is a crucial component in ensuring sustainable development. Therefore, significant resources are placed in acquiring data to be used in creation of a comprehensive description of the forest. One of the most novel way of describing the forest is using point clouds, which are produced by aerial laser scanning. Many algorithms have been proposed for extraction of the forest relevant information from point clouds, such as tree height or identification of individual trees. One of the most performant algorithms is based on a weighted oriented graph, which uses as nodes a hierarchical structure derived from point clouds (Strimbu and Strimbu, 2015). The main task of this project is to implement the weighted oriented graph algorithm that segment individual tree crowns such that it can be used by forestry actors.
Objectives
creation of an app that implements the hierarchical oriented graph algorithm for individual tree crown segmentation. The app can be written in Python or R.
Motivations
identification of individual trees from point clouds is an important step in execution of an accurate forest inventory
Qualifications
Minimum Qualifications:
proficiency in Python or R
Preferred Qualifications:knowledge of terrestrial coordinate system / GIS
Details
Project Partner:
Bogdan Strimbu
NDA/IPA:No Agreement Required
Number Groups:1
Project Status:Accepting Applicants
Website:mars.forestry.oregonstate.edu
Keywords:
graph theorysustain
