UnderWater RGB&Sonar (UW-RS)
收藏资源简介:
UnderWater RGB&Sonar(简称UW-RS)数据集由南开大学人工智能学院创建,专注于复杂海底场景下的伪装物体检测。该数据集包含1972张图像,分为水下光学数据部分(UW-R,1472张图像)和水下声纳数据部分(UW-S,502张图像)。UW-R数据集整理了大量包含水下伪装物体的图像,而UW-S数据集则包含大量侧扫声纳图像。数据集的创建过程涉及从多个现有数据集中选择图像和标签,以及通过专业人员手动标记声纳图像。UW-RS数据集主要用于水下环境的伪装物体检测研究,旨在解决水下目标识别和分类的问题。
The UnderWater RGB&Sonar (abbreviated as UW-RS) dataset was developed by the College of Artificial Intelligence, Nankai University, focusing on camouflage object detection in complex underwater seabed scenarios. This dataset contains a total of 1972 images, which are divided into two subsets: the underwater optical data subset (UW-R, consisting of 1472 images) and the underwater sonar data subset (UW-S, consisting of 502 images). The UW-R dataset curates a large volume of images featuring underwater camouflage objects, while the UW-S dataset includes numerous side-scan sonar images. The construction of the UW-RS dataset involves selecting images and labels from multiple existing datasets, as well as manually annotating the sonar images by professionals. The UW-RS dataset is primarily intended for research on camouflage object detection in underwater environments, with the goal of addressing challenges in underwater target recognition and classification.



