NEON NIST data science evaluation challenge: Methods and results of team Shawn
- Published
- Accepted
- Subject Areas
- Ecology, Forestry, Spatial and Geographic Information Science
- Keywords
- lidar, NEON, hyperspectral
- Copyright
- © 2018 Taylor
- Licence
- This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ Preprints) and either DOI or URL of the article must be cited.
- Cite this article
- 2018. NEON NIST data science evaluation challenge: Methods and results of team Shawn. PeerJ Preprints 6:e26967v1 https://doi.org/10.7287/peerj.preprints.26967v1
Abstract
This paper describes the methods used in the submission for team Shawn for the data science competition “Airborne Remote Sensing to Ecological Information”. I used canopy height rasters as well as NDVI rasters of the study area. I first filtered out pixels using a minimum NDVI threshold, then derived individual tree crowns using a watershed algorithm. I imposed limits on tree crown size and number using a minimum distance between two crowns and a maximum crown radius. All parameters were derived by minimizing the Jaccard coefficient. The final Jaccard coefficient on the training data was 0.117. All methods were implemented in Python are are available in code repositories.
Author Comment
This short methods paper describes the details of submission to the data science competition "Airborne Remote Sensing to Ecological Information".