AI4Shipwrecks
收藏资源简介:
AI4Shipwrecks数据集由密歇根大学创建,包含286张高分辨率标记的侧扫声纳图像,用于推动自主声纳图像理解的前沿技术。数据集通过在密歇根州的Thunder Bay国家海洋保护区使用自主水下航行器(AUV)进行调查收集。该数据集特别丰富,因为该区域拥有大量已知沉船目标,有助于机器学习在海洋探索中的应用。数据集旨在解决沉船检测算法的发展问题,特别是在声纳数据处理方面,以加速新沉船地点的发现。
The AI4Shipwrecks dataset was created by the University of Michigan, comprising 286 high-resolution annotated side-scan sonar images for advancing cutting-edge autonomous sonar image understanding technologies. The dataset was collected via surveys conducted using an Autonomous Underwater Vehicle (AUV) at the Thunder Bay National Marine Sanctuary in Michigan. This dataset is particularly valuable, as the sanctuary hosts a large number of known shipwreck targets, which supports the application of machine learning in marine exploration. The dataset aims to support the development of shipwreck detection algorithms, especially in sonar data processing, to accelerate the discovery of new shipwreck sites.

- 1Machine Learning for Shipwreck Segmentation from Side Scan Sonar Imagery: Dataset and Benchmark密歇根大学 · 2024年



