# Data and Code for Hyperspectral Scots Pine Seedlings Image Segmentation with Machine Learning

Data and Code for Hyperspectral Scots Pine Seedlings Image Segmentation with Machine Learning Pine_segmentation is the supplementary information and codes for a scientific publication "Overcoming Difficulties in the Segmentation of Hyperspectral Plant Images with Small Projection Areas Using Machine Learning", Neuwirthová et al., Scientific Reports, (2025; submitted) that involves the segmentation of hyperspectral images of pine seedlings. The data for this project comes from the Photon Systems Instruments (PSI) Research Center (Drásov, Czech Republic). This repository contains a set of R scripts and data files used in the analysis and segmentation of hyperspectral images of pine trees. These images are part of a study conducted at the Drásov Phenotyping Unit, operated by Photon Systems Instruments (PSI), which specializes in advanced phenotyping systems and high-end instrumentation for plant science research. Contents mask_kmeans_script.R: R script implementing VI masking and K-means clustering WLs.txt: Wavelength bands ID_data.txt: Pixel ID information 85centroids.txt: Centroid selected for described classes, used for K-means clustering ...Data_dc_ref.dat and ...Data_dc_ref.hdr: Example data in ENVI format (.dat; ~ 350MB) with header file (.hdr) Before running the scripts, ensure you have the following software and libraries installed: R (tested on R version 4.1.2 Bird Hippie) Required R packages: 'caTools', 'stringr' The codes and detailed technical implementations are available and explained in the accompanying GitHub repository: https://github.com/JCepl/Pine-hyperspectral-image-segmentaion To use the scripts, follow these steps: Download test data at Ensure all data files are placed in the appropriate directories as specified in the scripts. Open R and set the working directory to the repository folder. Related Introductory Study For additional context, readers are referred to the related article “Drought Response and Genetic Variation in Scots Pine Seedlings’ Provenances: Insights from High-Throughput Phenotyping for Climate-Resilient Forestry”. This work provides a detailed description of the experimental design, the source of the plant material, and the full methodology related to the phenotyping system. The supplementary information accompanying that article can be accessed here: https://doi.org/10.6084/m9.figshare.29474057