Datasets of SH's AI4ER MRes Project

This dataset contains the data used a generated during the course of SH's AI4ER MRes project. The dataset consists of LiDAR and RGB data over Sepilok Forest Reserve, in Sabah, Malaysia, collected and processed by NERC and NEODAAS, along with 901 manually delineated tree crowns in the area, and tree crown delineations predicted by two models: a Mask R-CNN model developed by SH, and the ITCfast algorithm, developed by Tom Swinfield and optimised by SH. LiDAR data is given for two years, 2014 and 2020, which was used to calculate changes in the carbon stock of the forest. The LiDAR and RGB data are provided as tiffs, while tree crowns are provided as shapefiles.