WasteMS
收藏资源简介:
WasteMS是由西安交利物浦大学和利物浦大学共同创建的第一个用于湖边废物语义分割的多光谱数据集。该数据集包含117张九通道多光谱图像,分辨率为682×682像素,涵盖多种湖边环境下的废物类型。数据收集自苏州的湖边草地,通过严格标注过程确保准确性。WasteMS旨在通过计算机视觉技术提高湖边环境监测的效率和准确性,特别是在复杂光照和背景条件下区分废物与自然环境。
WasteMS is the first multispectral dataset for lakeside waste semantic segmentation, jointly developed by Xi'an Jiaotong-Liverpool University and The University of Liverpool. This dataset contains 117 nine-channel multispectral images with a resolution of 682×682 pixels, covering various waste types in diverse lakeside environments. The data was collected from lakeside grasslands in Suzhou, with rigorous annotation procedures to ensure annotation accuracy. WasteMS aims to improve the efficiency and accuracy of lakeside environmental monitoring via computer vision technologies, particularly in distinguishing waste from natural environments under complex lighting and background conditions.
WasteMS 数据集
概述
WasteMS 数据集是首个用于湖边垃圾语义分割的多光谱数据集。该数据集包含了在不同光照条件下拍摄的草坪环境中多种类型的垃圾。通过严格标注过程,对图像中的垃圾进行了标记。使用代表性的语义分割框架对 WasteMS 数据集进行了分割准确性评估,并讨论了在湖边草坪上分割垃圾时遇到的问题。
下载链接
数据集可通过 Google Drive 下载。
数据集结构
数据集格式基于 mmsegmentation。
- ann_dir
- train
- .png
- val
- .png
- test
- .png
- train
- img_dir
- train
- .tif
- val
- .tif
- test
- .tif
- train
引用
如果该数据集对您的研究有用,请考虑引用:
@misc{zhu2024enhancingenvironmentalmonitoringmultispectral, title={Enhancing Environmental Monitoring through Multispectral Imaging: The WasteMS Dataset for Semantic Segmentation of Lakeside Waste}, author={Qinfeng Zhu and Ningxin Weng and Lei Fan and Yuanzhi Cai}, year={2024}, eprint={2407.17028}, archivePrefix={arXiv}, primaryClass={cs.CV}, url={https://arxiv.org/abs/2407.17028}, }

- 1Enhancing Environmental Monitoring through Multispectral Imaging: The WasteMS Dataset for Semantic Segmentation of Lakeside Waste西安交通利物浦大学 · 2024年



