中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 地理科学与资源研究所
  3. 中国科学院地理科学与资源研究所
  4. 资源与环境信息系统国家重点实验室
  5. 外文论文
Improving semantic segmentation accuracy in thin cloud interference scenarios by mixing simulated cloud-covered samples

文献类型：期刊论文

作者Wang, Haoyu6,7,8,9,10,11; Li, Junli8,11; Shen, Zhanfeng5,10; Zhang, Zihan4; Bai, Linze3; Li, Ruifeng2; Zhou, Chenghu1; De Maeyer, Philippe6,7,9,11; van de Voorde, Tim6,7,9,11
刊名INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
出版日期2024-09-01
卷号133页码:13
关键词Semantic segmentation Thin cloud interference Data augmentation Atmospheric scattering model Cloud simulation
ISSN号1569-8432
DOI10.1016/j.jag.2024.104087
产权排序11
英文摘要Thin cloud interference presents a significant challenge for the semantic segmentation of optical satellite imagery, which directly degrades the model accuracy and causes difficulties in sample selection. This paper generated a dataset named Populus euphratica and Tamarix chinensis discrimination (PTD), containing both cloudless and thin cloud scenarios. Based on this PTD dataset, an enhanced Atmospheric Scattering Model with Nonlinear Optimization (ASM_NL) was proposed to simulate high-fidelity thin clouds by incorporating two vital nonlinear terms: the point spread function and the Perlin noise. Additionally, we adopt a strategy of mixing simulated thin cloud-covered images (STCI) into the training set at a certain proportion to improve the semantic segmentation accuracy in thin cloud-covered scenarios. The conclusions are as follows: 1) ASM_NL can simulate high-fidelity clouds at an average Jensen-Shannon distance of 0.0699. 2) When dealing with medium- and highcloud density datasets, mixing STCI proved to be more effective than cloud removal in mitigating thin cloud interference, resulting in average macro F1 score improvements of 0.164 and 0.094, respectively. 3) The semantic segmentation accuracy improved significantly by mixing STCI with a minimal proportion of 1/60, demonstrating the activation of model transfer capabilities. This study provides a concise and efficient methodology for effectively mitigating thin cloud interference in deep learning-based optical satellite imagery analysis.
WOS关键词REMOVAL ; HAZE
资助项目National Natural Sciences Foundation of China[U2003201] ; Tianshan Talent-Science and Technology Innovation Team[2022TSYCTD0006] ; Third Integrated Scientific Expedition Project in Xinjiang[2021xjkk1403] ; Chinese Academy of Sciences President's International Fellowship Initiative[2024PVB0064]
WOS研究方向Remote Sensing
语种英语
WOS记录号WOS:001295713800001
出版者ELSEVIER
资助机构National Natural Sciences Foundation of China ; Tianshan Talent-Science and Technology Innovation Team ; Third Integrated Scientific Expedition Project in Xinjiang ; Chinese Academy of Sciences President's International Fellowship Initiative
源URL[http://ir.igsnrr.ac.cn/handle/311030/209214]  
专题资源与环境信息系统国家重点实验室_外文论文
通讯作者Li, Junli; Shen, Zhanfeng
作者单位1.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, State Key Lab Resources & Environm Informat Syst, Beijing 100101, Peoples R China
2.Univ Ghent, Dept Informat Technol, B-9000 Ghent, Belgium
3.Zhejiang Univ, Sch Earth Sci, Hangzhou 310058, Peoples R China
4.Peking Univ, Inst Remote Sensing & Geog Informat Syst, Sch Earth & Space Sci, Beijing Key Lab Spatial Informat Integrat & 3S App, Beijing 100871, Peoples R China
5.Chinese Acad Sci, Aerosp Informat Res Inst, Natl Engn Res Ctr Geomat, Beijing 100101, Peoples R China
6.Sino Belgian Joint Lab Geo Informat, Urumqi 830011, Peoples R China
7.Sino Belgian Joint Lab Geo Informat, B-9000 Ghent, Belgium
8.Key Lab GIS & RS Applicat Xinjiang Uygur Autonomou, Urumqi 830011, Peoples R China
9.Univ Ghent, Dept Geog, B-9000 Ghent, Belgium
10.Univ Chinese Acad Sci, Beijing 100049, Peoples R China
推荐引用方式
GB/T 7714		 Wang, Haoyu,Li, Junli,Shen, Zhanfeng,et al. Improving semantic segmentation accuracy in thin cloud interference scenarios by mixing simulated cloud-covered samples[J]. INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION,2024,133:13.
APA Wang, Haoyu.,Li, Junli.,Shen, Zhanfeng.,Zhang, Zihan.,Bai, Linze.,...&van de Voorde, Tim.(2024).Improving semantic segmentation accuracy in thin cloud interference scenarios by mixing simulated cloud-covered samples.INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION,133,13.
MLA Wang, Haoyu,et al."Improving semantic segmentation accuracy in thin cloud interference scenarios by mixing simulated cloud-covered samples".INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 133(2024):13.

入库方式： OAI收割

来源：地理科学与资源研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。