BioSCape Processed Training Dataset
收藏scholardata.sun.ac.za2024-11-09 更新2025-01-15 收录
下载链接:
https://scholardata.sun.ac.za/articles/dataset/BioSCape_Processed_Training_Dataset/27633564/1
下载链接
链接失效反馈官方服务:
资源简介:
These data are collected for the NASA-funded BioSCape Project: “Impacts of Invasive Alien Species on Biodiversity and Ecosystem Functioning”. This particular dataset is curated for the alien tree classification of the NASA airborne remote sensing data. The dataset includes training data points of invasive alien trees and other classes for the three study sites in Cape Floristic Region (Figure 1). The dataset was prepared according to the following pre-agreed criteria:As many points as possible were collectedThe classes needed to be even (same number of training points) for the machine learning algorithmsPoints didn’t need to be paired (i.e. paired invasive alien tree and fynbos points)It was not necessary to collect training data in all sampling units, though a general effort to avoid bias and to sample across different sampling units was attempted
本数据集系为美国国家航空航天局(NASA)资助的BioSCape项目“入侵性外来物种对生物多样性和生态系统功能的影响”所收集。该特定数据集旨在对NASA航空遥感数据中的外来树种进行分类整理。数据集涵盖了入侵性外来树种及其他类别在开普植物区三个研究地点的训练数据点(见图1)。数据集的整理遵循了以下预定的标准:尽可能多地收集数据点;各类别需保持数量均衡(即相同数量的训练数据点)以适应机器学习算法;数据点无需配对(即不需要将入侵性外来树种与南非针茅的数据点配对);虽然不必在所有采样单元中收集训练数据,但努力避免偏差,并在不同的采样单元中进行采样。
提供机构:
SUNScholarData



