Deactivation performed! Your visits to this website will no longer be collected by the web analytics. Please note that the Matomo deactivation cookie of this website will also be deleted if you remove the cookies stored in your browser. In addition, if you use a different computer or a different web browser, you will need to complete the deactivation procedure again.

  Your visit to this website is not currently collected by Matomo web analytics. Check this box to opt-in.
X
Tree-Challenge

AI for Tree Species Classification

17. June 2020.

The Bavarian National Park is Germany`s first national park and has recently had to fight massive tree death caused by native beetles. To be able to determine the remaining trees and the amount of dead wood, a precise and fast technology was sought that would support the employees in the inventory. In his final thesis, a student at MUAS devoted himself to a sub-area of machine learning, called deep learning, and analysed to what extent this could benefit the operators of the national park. The method of deep learning was chosen because it learns hierarchically, like the human brain, and has a high classification accuracy. The challenge is still running as an offer for other theses and the Bavarian Forest National Park would like to cooperate with the Digital Transformation Lab for further challenges in the future.

The student was supported by the team of Amazon Web Services, who helped with the the “Working Backwards” methodology as well as the use of AWS services.

The full report on the challenge can be found here.