SWINySEG

该数据集名为SWINySEG，它是由SWIMSEG和SWINSEG数据集增强合并而成的复合数据集，旨在平衡白天和夜晚的图像数据，为云分割提供图像资料。该数据集包含了来自SWIMSEG的1013张白天图像和来自SWINSEG的115张夜晚图像，所有图像都被调整为300×300像素以便进行处理。这是一个涵盖白天和夜晚图像的复合数据集，其任务是利用地面图像进行云分割。

This dataset is named SWINySEG. It is a composite dataset constructed by augmenting and merging the SWIMSEG and SWINSEG datasets, aiming to balance daytime and nighttime image data and provide image resources for cloud segmentation. This dataset contains 1013 daytime images sourced from SWIMSEG and 115 nighttime images from SWINSEG, and all images have been resized to 300×300 pixels for processing. This is a composite dataset covering both daytime and nighttime images, whose task is to perform cloud segmentation using ground-based images.

搜集汇总

背景与挑战

背景概述

SWINySEG数据集包含6768张白天和夜晚的天空/云块图像及其对应的二进制地面实况图，图像来源于SWIMSEG和SWINSEG两个早期数据集。所有图像于2016年在新加坡使用WAHRSIS设备采集，并得到新加坡气象局专家的标注支持，数据集采用知识共享许可协议发布。