从摄像头图像中检测野火烟雾的注意力驱动型 YOLOv5

IF 1.3 4区 地球科学 Q3 GEOSCIENCES, MULTIDISCIPLINARY
Himadri Vaidya, Akansha Gupta, Kamal Kumar Ghanshala
{"title":"从摄像头图像中检测野火烟雾的注意力驱动型 YOLOv5","authors":"Himadri Vaidya, Akansha Gupta, Kamal Kumar Ghanshala","doi":"10.1007/s12040-024-02391-3","DOIUrl":null,"url":null,"abstract":"<p>Wildfires are serious hazards for the environment, and WFSD (Wildfire Smoke Detection) is a challenge for ensuring optimal response and mitigation efforts. Hence, this study suggests an attention-based YOLOv5 (You Only Look Once) network for detecting smoke instances within video frames, particularly ECA (Efficient Channel Attention), GAM (Global Attention Module) and CA (Coordinate Attention). Here, an open-source wildfire smoke dataset divided into train, validation and test set is used for experimentation. The comprehensive research and evaluations show that the incorporation of attention mechanisms successfully enhances the accuracy and robustness of the YOLOv5 model for WFSD. In the training among the attention modules, GAM appears as the most effective, attaining an improved 95% F1 score on the dataset. This research provides the impact of attention mechanisms on object detection in the context of wildfire smoke. The findings of the research paper contribute to improving the capabilities of deep learning models for emergency response and environmental monitoring. The proposed methodology not only outperforms regular YOLOv5 but also sets up a benchmark for future research of WFSD.</p>","PeriodicalId":15609,"journal":{"name":"Journal of Earth System Science","volume":"14 1","pages":""},"PeriodicalIF":1.3000,"publicationDate":"2024-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Attention-driven YOLOv5 for wildfire smoke detection from camera images\",\"authors\":\"Himadri Vaidya, Akansha Gupta, Kamal Kumar Ghanshala\",\"doi\":\"10.1007/s12040-024-02391-3\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>Wildfires are serious hazards for the environment, and WFSD (Wildfire Smoke Detection) is a challenge for ensuring optimal response and mitigation efforts. Hence, this study suggests an attention-based YOLOv5 (You Only Look Once) network for detecting smoke instances within video frames, particularly ECA (Efficient Channel Attention), GAM (Global Attention Module) and CA (Coordinate Attention). Here, an open-source wildfire smoke dataset divided into train, validation and test set is used for experimentation. The comprehensive research and evaluations show that the incorporation of attention mechanisms successfully enhances the accuracy and robustness of the YOLOv5 model for WFSD. In the training among the attention modules, GAM appears as the most effective, attaining an improved 95% F1 score on the dataset. This research provides the impact of attention mechanisms on object detection in the context of wildfire smoke. The findings of the research paper contribute to improving the capabilities of deep learning models for emergency response and environmental monitoring. The proposed methodology not only outperforms regular YOLOv5 but also sets up a benchmark for future research of WFSD.</p>\",\"PeriodicalId\":15609,\"journal\":{\"name\":\"Journal of Earth System Science\",\"volume\":\"14 1\",\"pages\":\"\"},\"PeriodicalIF\":1.3000,\"publicationDate\":\"2024-09-02\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Journal of Earth System Science\",\"FirstCategoryId\":\"89\",\"ListUrlMain\":\"https://doi.org/10.1007/s12040-024-02391-3\",\"RegionNum\":4,\"RegionCategory\":\"地球科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"GEOSCIENCES, MULTIDISCIPLINARY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Earth System Science","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1007/s12040-024-02391-3","RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"GEOSCIENCES, MULTIDISCIPLINARY","Score":null,"Total":0}
引用次数: 0

摘要

野火是对环境的严重危害,而 WFSD(野火烟雾检测)是确保最佳响应和减灾工作的一项挑战。因此,本研究提出了一种基于注意力的 YOLOv5(只看一次)网络,用于检测视频帧中的烟雾实例,特别是 ECA(高效通道注意力)、GAM(全局注意力模块)和 CA(协调注意力)。本文使用了一个开源的野火烟雾数据集进行实验,该数据集分为训练集、验证集和测试集。综合研究和评估结果表明,注意力机制的加入成功地提高了 YOLOv5 模型用于 WFSD 的准确性和鲁棒性。在注意力模块的训练中,GAM显得最为有效,在数据集上的F1得分提高了95%。这项研究提供了注意力机制对野火烟雾背景下物体检测的影响。本文的研究成果有助于提高深度学习模型在应急响应和环境监测方面的能力。所提出的方法不仅优于常规的 YOLOv5,还为 WFSD 的未来研究树立了标杆。
本文章由计算机程序翻译,如有差异,请以英文原文为准。

Attention-driven YOLOv5 for wildfire smoke detection from camera images

Attention-driven YOLOv5 for wildfire smoke detection from camera images

Wildfires are serious hazards for the environment, and WFSD (Wildfire Smoke Detection) is a challenge for ensuring optimal response and mitigation efforts. Hence, this study suggests an attention-based YOLOv5 (You Only Look Once) network for detecting smoke instances within video frames, particularly ECA (Efficient Channel Attention), GAM (Global Attention Module) and CA (Coordinate Attention). Here, an open-source wildfire smoke dataset divided into train, validation and test set is used for experimentation. The comprehensive research and evaluations show that the incorporation of attention mechanisms successfully enhances the accuracy and robustness of the YOLOv5 model for WFSD. In the training among the attention modules, GAM appears as the most effective, attaining an improved 95% F1 score on the dataset. This research provides the impact of attention mechanisms on object detection in the context of wildfire smoke. The findings of the research paper contribute to improving the capabilities of deep learning models for emergency response and environmental monitoring. The proposed methodology not only outperforms regular YOLOv5 but also sets up a benchmark for future research of WFSD.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Journal of Earth System Science
Journal of Earth System Science Earth and Planetary Sciences-General Earth and Planetary Sciences
CiteScore
3.20
自引率
5.30%
发文量
226
期刊介绍: The Journal of Earth System Science, an International Journal, was earlier a part of the Proceedings of the Indian Academy of Sciences – Section A begun in 1934, and later split in 1978 into theme journals. This journal was published as Proceedings – Earth and Planetary Sciences since 1978, and in 2005 was renamed ‘Journal of Earth System Science’. The journal is highly inter-disciplinary and publishes scholarly research – new data, ideas, and conceptual advances – in Earth System Science. The focus is on the evolution of the Earth as a system: manuscripts describing changes of anthropogenic origin in a limited region are not considered unless they go beyond describing the changes to include an analysis of earth-system processes. The journal''s scope includes the solid earth (geosphere), the atmosphere, the hydrosphere (including cryosphere), and the biosphere; it also addresses related aspects of planetary and space sciences. Contributions pertaining to the Indian sub- continent and the surrounding Indian-Ocean region are particularly welcome. Given that a large number of manuscripts report either observations or model results for a limited domain, manuscripts intended for publication in JESS are expected to fulfill at least one of the following three criteria. The data should be of relevance and should be of statistically significant size and from a region from where such data are sparse. If the data are from a well-sampled region, the data size should be considerable and advance our knowledge of the region. A model study is carried out to explain observations reported either in the same manuscript or in the literature. The analysis, whether of data or with models, is novel and the inferences advance the current knowledge.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信