Dataset used for the Master's Thesis on Semantic Segmentation of Dead Picea abies Trees Using Deep Learning

This dataset contains the orthomosaic imagery and associated shapefiles used in the master’s thesis titled “Semantic Segmentation of Dead Trees (Picea abies) Using Deep Learning”

Link to the thesis

The shapefiles include tree annotations categorized into the following health classes:

0 – Healthy

1 – Slight Damage

2 – Moderate Damage

4 – Dead

This dataset is intended to support research in the fields of remote sensing, semantic segmentation, forest health monitoring, and deep learning-based image analysis. It can be used for various scientific studies focused on vegetation condition assessment, object detection, or geospatial analysis.Researchers can generate custom training and testing datasets using these shapefiles. A supporting R script is available in the GitHub repository.

https://github.com/muhammedsinan1997/Semantic_Segmentation_of_dead_trees

If you use this dataset, please cite both the dataset and the following thesis:

Sinan, M. (2022). Semantic Segmentation of Dead Trees (Picea abies) Using Deep Learning (Master's thesis, Hochschule für nachhaltige Entwicklung Eberswalde).

Muhammed Sinan. (2025). Dataset used for the Master's Thesis on Semantic Segmentation of Dead Picea abies Trees Using Deep Learning [Data set]. Zenodo. https://doi.org/10.5281/zenodo.15168163