An Improved Bald Eagle Search Algorithm with Deep Learning Model for Forest Fire Detection Using Hyperspectral Remote Sensing Images

IF 2 4区地球科学 Q3 REMOTE SENSING

Canadian Journal of Remote Sensing Pub Date : 2022-06-16 DOI:10.1080/07038992.2022.2077709

A. Algarni, Nazik Alturki, N. Soliman, S. Abdel-Khalek, A. Mousa

{"title":"An Improved Bald Eagle Search Algorithm with Deep Learning Model for Forest Fire Detection Using Hyperspectral Remote Sensing Images","authors":"A. Algarni, Nazik Alturki, N. Soliman, S. Abdel-Khalek, A. Mousa","doi":"10.1080/07038992.2022.2077709","DOIUrl":null,"url":null,"abstract":"Abstract This paper presents an improved Bald Eagle Search Algorithm with Deep Learning model for forest fire detection (IBESDL-FFD) technique using hyperspectral images (HSRS). The major intention of the IBESDL-FFD technique is to identify the presence of forest fire in the HSRS images. To achieve this, the IBESDL-FFD technique involves data pre-processing in two stages namely data augmentation and noise removal. Besides, IBES algorithm with NASNetLarge method was utilized as a feature extractor to determine feature vectors. Finally, Firefly algorithm (FFA) with denoising autoencoder (DAE) is applied for the classification of forest fire. The design of IBES and FFA techniques helps to adjust optimally the parameters contained in the NSANetLarge and DAE models respectively. For demonstrating the better outcomes of the IBESDL-FFD approach, a wide-ranging simulation was implemented and the outcomes are examined. The results reported the better outcomes of the IBESDL-FFD technique over the existing techniques with maximum average accuracy of 93.75%.","PeriodicalId":48843,"journal":{"name":"Canadian Journal of Remote Sensing","volume":"48 1","pages":"609 - 620"},"PeriodicalIF":2.0000,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Canadian Journal of Remote Sensing","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1080/07038992.2022.2077709","RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"REMOTE SENSING","Score":null,"Total":0}

引用次数: 3

Abstract

Abstract This paper presents an improved Bald Eagle Search Algorithm with Deep Learning model for forest fire detection (IBESDL-FFD) technique using hyperspectral images (HSRS). The major intention of the IBESDL-FFD technique is to identify the presence of forest fire in the HSRS images. To achieve this, the IBESDL-FFD technique involves data pre-processing in two stages namely data augmentation and noise removal. Besides, IBES algorithm with NASNetLarge method was utilized as a feature extractor to determine feature vectors. Finally, Firefly algorithm (FFA) with denoising autoencoder (DAE) is applied for the classification of forest fire. The design of IBES and FFA techniques helps to adjust optimally the parameters contained in the NSANetLarge and DAE models respectively. For demonstrating the better outcomes of the IBESDL-FFD approach, a wide-ranging simulation was implemented and the outcomes are examined. The results reported the better outcomes of the IBESDL-FFD technique over the existing techniques with maximum average accuracy of 93.75%.

查看原文本刊更多论文

基于深度学习模型的改进秃鹰搜索算法在高光谱遥感图像森林火灾探测中的应用

摘要提出了一种基于高光谱图像(HSRS)的基于深度学习模型的改进秃鹰搜索算法(IBESDL-FFD)。IBESDL-FFD技术的主要目的是识别HSRS图像中是否存在森林火灾。为了实现这一点，IBESDL-FFD技术包括两个阶段的数据预处理，即数据增强和去噪。此外，利用IBES算法结合NASNetLarge方法作为特征提取器，确定特征向量。最后，将萤火虫算法(FFA)与去噪自编码器(DAE)相结合，对森林火灾进行分类。IBES和FFA技术的设计分别有助于优化调整NSANetLarge和DAE模型中包含的参数。为了证明IBESDL-FFD方法的更好结果，实施了广泛的模拟并检查了结果。结果表明，IBESDL-FFD技术优于现有技术，最高平均准确率为93.75%。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Canadian Journal of Remote Sensing REMOTE SENSING-

自引率

3.80%

发文量

期刊介绍： Canadian Journal of Remote Sensing / Journal canadien de télédétection is a publication of the Canadian Aeronautics and Space Institute (CASI) and the official journal of the Canadian Remote Sensing Society (CRSS-SCT). Canadian Journal of Remote Sensing provides a forum for the publication of scientific research and review articles. The journal publishes topics including sensor and algorithm development, image processing techniques and advances focused on a wide range of remote sensing applications including, but not restricted to; forestry and agriculture, ecology, hydrology and water resources, oceans and ice, geology, urban, atmosphere, and environmental science. Articles can cover local to global scales and can be directly relevant to the Canadian, or equally important, the international community. The international editorial board provides expertise in a wide range of remote sensing theory and applications.