iVision-MRSSD: Multi-Resolution Satellite based SAR Ship Detection Dataset

This is a comprehensive labeled satellite based SAR ship detection dataset aimed for usage with state of the art deep learning based object detectors. The dataset contains 11,590 images along with the same number of label files. This is a diverse dataset containing images from six different space borne sensors and includes a wide variety of in-shore and off-shore scenes under different conditions. The dataset can be used to train, test and optimize object detection models to achieve high ship detection performance for any SAR sensor and under varying conditions. Related Data Article is available at: https://authors.elsevier.com/sd/article/S2352340923006054

本数据集为一款基于星载合成孔径雷达（SAR）的标注型综合船舶检测数据集，旨在适配当前最先进的基于深度学习的目标检测器。本数据集共包含11590幅图像与等量的标注文件，具有高度多样性：涵盖来自6种不同星载传感器的图像，包含不同环境条件下的近岸与离岸船舶场景。该数据集可用于训练、测试与优化目标检测模型，以在任意SAR传感器及多变环境条件下实现高性能船舶检测。相关数据集研究论文可通过以下链接获取：https://authors.elsevier.com/sd/article/S2352340923006054

搜集汇总

背景与挑战

背景概述

iVision-MRSSD是一个用于合成孔径雷达（SAR）船舶检测的综合性数据集，包含11,590张图像及对应YOLO格式标签文件，已按70:20:10的比例划分为训练、验证和测试集。该数据集具有多分辨率特点，图像来源于六种不同星载传感器，涵盖多种近岸和离岸场景，旨在支持深度学习模型在SAR船舶检测任务中的训练和优化。