自定义珊瑚礁底栖结构数据集
收藏资源简介:
自定义珊瑚礁底栖结构数据集由巴西东北部的海洋生态实验室/UFRN收集和标注,用于自动化珊瑚礁底栖图像注释。该数据集包含1549张图像,每张图像标注了25个随机点,总计38,725个注释点,涵盖54个不同的标签。数据集的创建过程涉及与机器学习专家和海洋生态学家的合作,确保标签的层次结构符合生物分类原则。该数据集主要用于提高珊瑚礁监测的自动化水平,解决气候变化对珊瑚礁生态系统的威胁。
The Customized Coral Reef Benthic Structure Dataset was collected and annotated by the Marine Ecology Laboratory/UFRN in Northeast Brazil for automated coral reef benthic image annotation. This dataset comprises 1549 images, with 25 randomly selected points annotated on each image, resulting in a total of 38,725 annotated points spanning 54 distinct labels. The development of this dataset involved collaboration between machine learning experts and marine ecologists to ensure that the hierarchical structure of the labels adheres to biological classification principles. This dataset is primarily designed to advance the automation of coral reef monitoring and mitigate the threats posed by climate change to coral reef ecosystems.

- 1Hierarchical Classification for Automated Image Annotation of Coral Reef Benthic Structures法国蒙彼利埃大学 · 2024年



