학술논문
농림위성 활용을 위한 산불 피해지 분류 딥러닝 알고리즘 평가
Document Type
Article
Author
차성은 / Sungeun Cha; 원명수 / Myoungsoo Won; 장근창 / Keunchang Jang; 김경민 / Kyoungmin Kim; 김원국 / Wonkook Kim; 백승일 / Seungil Baek; 임중빈 / Joongbin Lim
Source
대한원격탐사학회지 / Korean Journal of Remote Sensing. Dec 31, 2022 38(6):1273
Subject
Language
Korean
English
English
ISSN
1225-6161
Abstract
최근 기후변화로 인해 중대형 산불이 빈번하게 발생하여 매년 인명 및 재산피해로 이어지고 있다. 원격탐사를 활용한 산불 피해지 모니터링 기법은 신속한 정보와 대규모 피해지의 객관적인 결과를 취득할 수 있다. 본 연구에서는 산불 피해지를 분류하기 위해 Sentinel-2의 분광대역, 정규식생지수(normalized difference vegetation index, NDVI), 정규수역지수(normalized difference water index, NDWI)를 활용하여 2022년 3월 발생한 강릉·동해 산불 피해지를 대상으로 U-net 기반 convolutional neural networks (CNNs) 딥러닝 모형을 모의하였다. 산불 피해지 분류 결과 강릉·동해 산불 피해지의 경우 97.3% (ƒ1=0.486, IoU= 0.946)로 분류 정확도가 높았으나, 과적합(overfitting)의 가능성을 배제하기 어려워 울진·삼척 지역으로 동일한 모형을 적용하였다. 그 결과, 국립산림과학원에서 보고한 산불 피해 면적과의 중첩도가 74.4%로 확인되어 모형의 불확도를 고려하더라도 높은 수준의 정확도를 확인하였다. 본 연구는 농림위성과 유사한 분광대역을 선택적으로 사용하였으며, Sentinel-2 영상을 활용한 산불 피해지 분류가 정량적으로 가능함을 시사한다.
Recently, forest fires have frequently occurred due to climate change, leading to human and property damage every year. The forest fire monitoring technique using remote sensing can obtain quick and large-scale information of fire-damaged areas. In this study, the Gangneung and Donghae forest fires that occurred in March 2022 were analyzed using the spectral band of Sentinel-2, the normalized difference vegetation index (NDVI), and the normalized difference water index (NDWI) to classify the affected areas of forest fires. The U-net based convolutional neural networks (CNNs) model was simulated for the fire-damaged areas. The accuracy of forest fire classification in Donghae and Gangneung classification was high at 97.3% (ƒ1=0.486, IoU= 0.946). The same model used in Donghae and Gangneung was applied to Uljin and Samcheok areas to get rid of the possibility of overfitting often happen in machine learning. As a result, the portion of overlap with the forest fire damage area reported by the National Institute of Forest Science (NIFoS) was 74.4%, confirming a high level of accuracy even considering the uncertainty of the model. This study suggests that it is possible to quantitatively evaluate the classification of forest fire-damaged area using a spectral band and indices similar to that of the Compact Advanced Satellite 500 (CAS500-4) in the Sentinel-2.
Recently, forest fires have frequently occurred due to climate change, leading to human and property damage every year. The forest fire monitoring technique using remote sensing can obtain quick and large-scale information of fire-damaged areas. In this study, the Gangneung and Donghae forest fires that occurred in March 2022 were analyzed using the spectral band of Sentinel-2, the normalized difference vegetation index (NDVI), and the normalized difference water index (NDWI) to classify the affected areas of forest fires. The U-net based convolutional neural networks (CNNs) model was simulated for the fire-damaged areas. The accuracy of forest fire classification in Donghae and Gangneung classification was high at 97.3% (ƒ1=0.486, IoU= 0.946). The same model used in Donghae and Gangneung was applied to Uljin and Samcheok areas to get rid of the possibility of overfitting often happen in machine learning. As a result, the portion of overlap with the forest fire damage area reported by the National Institute of Forest Science (NIFoS) was 74.4%, confirming a high level of accuracy even considering the uncertainty of the model. This study suggests that it is possible to quantitatively evaluate the classification of forest fire-damaged area using a spectral band and indices similar to that of the Compact Advanced Satellite 500 (CAS500-4) in the Sentinel-2.