Sentinel-2 dataset with Dynamic World labels
收藏资源简介:
本研究贡献了一个定制的Sentinel-2数据集,包含了Dynamic World标签,专门为湿地分类任务设计,并公开供研究人员使用。数据集使用了Sentinel-2卫星图像,覆盖了荷兰六个湿地区域,并通过深度学习方法进行了预处理。数据集提供了丰富的植被、水体和土壤特征信息,对于湿地生态系统的监测和管理具有重要意义。数据集的构建旨在解决湿地分类任务中标注数据稀缺的问题,通过引入监督学习和自监督学习方法,提高了模型在湿地区域的分割和分类准确性。
This study presents a custom Sentinel-2 dataset equipped with Dynamic World labels, which is specifically designed for wetland classification tasks and made publicly available to researchers. The dataset employs Sentinel-2 satellite imagery covering six wetland areas across the Netherlands, and has been preprocessed via deep learning methods. It provides abundant feature information related to vegetation, water and soil, which holds critical importance for the monitoring and management of wetland ecosystems. The construction of this dataset aims to address the problem of scarce labeled data in wetland classification tasks, and by introducing supervised and self-supervised learning approaches, it enhances the accuracy of model segmentation and classification within wetland regions.




