Burned area detection using convolutional neural network based on spatial information of synthetic aperture radar data in Indonesia

Q2 Agricultural and Biological Sciences Geography, Environment, Sustainability Pub Date : 2024-07-05 DOI:10.24057/2071-9388-2024-3109
A. I. Lestari, D. Kushardono, A. A. Bayanuddin
{"title":"Burned area detection using convolutional neural network based on spatial information of synthetic aperture radar data in Indonesia","authors":"A. I. Lestari, D. Kushardono, A. A. Bayanuddin","doi":"10.24057/2071-9388-2024-3109","DOIUrl":null,"url":null,"abstract":"Forest and land fires are disasters that often occur in Indonesia which affects neighbouring countries. The burned area can be observed using remote sensing. Synthetic aperture radar (SAR) sensor data is advantageous since it can penetrate clouds and smoke. However, image analysis of SAR data differs from optical data, which is based on properties such as intensity, texture, and polarimetric feature. This research aims to propose a method to detect burned areas from the extracted feature of Sentinel-1 data. The features were classified using the Convolutional Neural Network (CNN) classifier. To find the best input features, several classification schemes were tested, including intensity and polarimetric features by applying the Boxcar speckle filter and the Gray Level Co-occurrence Matrix (GLCM) texture feature without using the Boxcar speckle filter. Additionally, this research investigates the significance of a window size parameter for each scheme. The results show the highest overall accuracy achieved 84% using CNN classification utilizing the GLCM texture features and without conducting the Boxcar speckle filter on the window size of 17×17 pixels when tested on the part region of Pulang Pisau Regency and Kapuas Regency, Central Kalimantan in 2019. The total burned area was 76,098.6 ha. The use of GLCM texture features without conducting the Boxcar speckle filter as input classification performs better than using intensity and polarimetric features that undergo the Boxcar speckle filter. Combining intensity and polarimetric features with performing the Boxcar speckle filter improves better classification performance over utilizing them separately. Furthermore, the selection of window size also contributes to improve the model performance.","PeriodicalId":37517,"journal":{"name":"Geography, Environment, Sustainability","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Geography, Environment, Sustainability","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.24057/2071-9388-2024-3109","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"Agricultural and Biological Sciences","Score":null,"Total":0}
引用次数: 0

Abstract

Forest and land fires are disasters that often occur in Indonesia which affects neighbouring countries. The burned area can be observed using remote sensing. Synthetic aperture radar (SAR) sensor data is advantageous since it can penetrate clouds and smoke. However, image analysis of SAR data differs from optical data, which is based on properties such as intensity, texture, and polarimetric feature. This research aims to propose a method to detect burned areas from the extracted feature of Sentinel-1 data. The features were classified using the Convolutional Neural Network (CNN) classifier. To find the best input features, several classification schemes were tested, including intensity and polarimetric features by applying the Boxcar speckle filter and the Gray Level Co-occurrence Matrix (GLCM) texture feature without using the Boxcar speckle filter. Additionally, this research investigates the significance of a window size parameter for each scheme. The results show the highest overall accuracy achieved 84% using CNN classification utilizing the GLCM texture features and without conducting the Boxcar speckle filter on the window size of 17×17 pixels when tested on the part region of Pulang Pisau Regency and Kapuas Regency, Central Kalimantan in 2019. The total burned area was 76,098.6 ha. The use of GLCM texture features without conducting the Boxcar speckle filter as input classification performs better than using intensity and polarimetric features that undergo the Boxcar speckle filter. Combining intensity and polarimetric features with performing the Boxcar speckle filter improves better classification performance over utilizing them separately. Furthermore, the selection of window size also contributes to improve the model performance.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
利用基于合成孔径雷达数据空间信息的卷积神经网络检测印度尼西亚的烧毁区域
森林和土地火灾是印度尼西亚经常发生的灾害,也影响到邻国。可以利用遥感技术观测被烧毁的区域。合成孔径雷达 (SAR) 传感器数据具有穿透云层和烟雾的优势。然而,合成孔径雷达数据的图像分析不同于光学数据,后者基于强度、纹理和偏振特征等属性。本研究旨在提出一种从哨兵-1 数据提取的特征中检测烧毁区域的方法。使用卷积神经网络(CNN)分类器对特征进行分类。为了找到最佳输入特征,测试了几种分类方案,包括应用 Boxcar斑点滤波器的强度和偏振特征,以及不使用 Boxcar斑点滤波器的灰度共现矩阵(GLCM)纹理特征。此外,本研究还探讨了窗口大小参数对每种方案的重要性。结果表明,2019 年在加里曼丹中部普朗皮绍地区和卡普阿斯地区的部分区域进行测试时,利用 GLCM 纹理特征的 CNN 分类法在不使用 Boxcar斑点滤波器的情况下,窗口大小为 17×17 像素,总体准确率达到 84%。烧毁总面积为 76,098.6 公顷。与使用经过 Boxcar斑点滤波器的强度和偏振特征相比,使用未经过 Boxcar斑点滤波器的 GLCM 纹理特征作为输入分类的效果更好。将强度和偏振特征与箱形斑点滤波器相结合,比单独使用这两种特征更能提高分类性能。此外,窗口大小的选择也有助于提高模型的性能。
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 去求助
来源期刊
Geography, Environment, Sustainability
Geography, Environment, Sustainability Social Sciences-Geography, Planning and Development
CiteScore
2.50
自引率
0.00%
发文量
37
审稿时长
12 weeks
期刊介绍: Journal “GEOGRAPHY, ENVIRONMENT, SUSTAINABILITY” is founded by the Faculty of Geography of Lomonosov Moscow State University, The Russian Geographical Society and by the Institute of Geography of RAS. It is the official journal of Russian Geographical Society, and a fully open access journal. Journal “GEOGRAPHY, ENVIRONMENT, SUSTAINABILITY” publishes original, innovative, interdisciplinary and timely research letter articles and concise reviews on studies of the Earth and its environment scientific field. This goal covers a broad spectrum of scientific research areas (physical-, social-, economic-, cultural geography, environmental sciences and sustainable development) and also considers contemporary and widely used research methods, such as geoinformatics, cartography, remote sensing (including from space), geophysics, geochemistry, etc. “GEOGRAPHY, ENVIRONMENT, SUSTAINABILITY” is the only original English-language journal in the field of geography and environmental sciences published in Russia. It is supposed to be an outlet from the Russian-speaking countries to Europe and an inlet from Europe to the Russian-speaking countries regarding environmental and Earth sciences, geography and sustainability. The main sections of the journal are the theory of geography and ecology, the theory of sustainable development, use of natural resources, natural resources assessment, global and regional changes of environment and climate, social-economical geography, ecological regional planning, sustainable regional development, applied aspects of geography and ecology, geoinformatics and ecological cartography, ecological problems of oil and gas sector, nature conservations, health and environment, and education for sustainable development. Articles are freely available to both subscribers and the wider public with permitted reuse.
期刊最新文献
Modeling land use change of mid-sized cities in the process of metropolization. Case study La Serena-Coquimbo conurbation, Chile Land suitability of coffee cultivation under climate change influence in the Ecuadorian Amazon The 3Ps (profits, problems & planning) of dams as inevitable developmental source: a review GIS mapping of the soil cover of an urbanized territory: drainage basin of the Setun river in the west of Moscow (Russian Federation) Unraveling the spatial dynamics: exploring the urban form characteristics and COVID-19 cases in Yogyakarta city, Indonesia
×
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
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 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 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1