首页  >  科研动态  >  正文
科研动态
朱邵聪(硕士生)、沈焕锋的论文在IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING刊出
发布时间:2024-03-21 10:26:10     发布者:易真     浏览次数:

标题: Transferring Deep Models for Cloud Detection in Multisensor Images via Weakly Supervised Learning

作者: Zhu, SC (Zhu, Shaocong); Li, ZW (Li, Zhiwei); Shen, HF (Shen, Huanfeng)

来源出版物: IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING : 62 文献号: 5609518 DOI: 10.1109/TGRS.2024.3358824 出版年: 2024

摘要: Recently, deep learning has been widely used for cloud detection in satellite images; however, due to radiometric and spatial resolution differences in images from different sensors and time-consuming process of manually labeling cloud detection datasets, it is difficult to effectively generalize deep learning models for cloud detection in multisensor images. This article propose a weakly supervised learning method for transferring deep models for cloud detection in multisensor images (TransMCD), which leverages the generalization of deep models and the spectral features of clouds to construct pseudo-label dataset to improve the generalization of models. A deep model is first pretrained using a well-annotated cloud detection dataset, which is used to obtain a rough cloud mask of unlabeled target image. The rough mask can be used to determine the spectral threshold adaptively for cloud segmentation of target image. Block-level pseudo labels with high confidence in target image are selected using the rough mask and spectral mask. Unsupervised segmentation technique is used to construct a high-quality pixel-level pseudo-label dataset. Finally, the pseudo-label dataset is used as supervised information for transferring the pretrained model to target image. The TransMCD method was validated by transferring model trained on 16-m Gaofen-1 wide field of view(WFV)images to 8-m Gaofen-1, 4-m Gaofen-2, and 10-m Sentinel-2 images. The F1-score of the transferred models on target images achieves improvements of 1.23%-9.63% over the pretrained models, which is comparable to the fully-supervised models trained with well-annotated target images, suggesting the efficiency of the TransMCD method for cloud detection in multisensor images.

作者关键词: Cloud computing; Clouds; Sensors; Predictive models; Supervised learning; Image segmentation; Deep learning; Cloud detection; deep learning; multisensor images; weakly supervised learning

KeyWords Plus: SHADOW DETECTION; AUTOMATED CLOUD; SNOW DETECTION; LAND-SURFACE; ALGORITHM; VALIDATION; LANDSAT-8; FEATURES; MODIS

地址: [Zhu, Shaocong; Shen, Huanfeng] Wuhan Univ, Sch Resource & Environm Sci, Wuhan 430079, Peoples R China.

[Li, Zhiwei] Hong Kong Polytech Univ, Dept Land Surveying & Geoinformat, Hong Kong, Peoples R China.

通讯作者地址: Shen, HF (通讯作者)Wuhan Univ, Sch Resource & Environm Sci, Wuhan 430079, Peoples R China.

电子邮件地址: zhushaocong@whu.edu.cn; zhiwei.li@polyu.edu.hk; shenhf@whu.edu.cn

影响因子:8.2


信息服务
学院网站教师登录 学院办公电话 学校信息门户登录

版权所有 © 太阳成tyc7111cc(股份)有限公司-搜狗百科
地址:湖北省武汉市珞喻路129号 邮编:430079 
电话:027-68778381,68778284,68778296 传真:027-68778893    邮箱:sres@whu.edu.cn