Sentinel-2 영상과 클러스터링 기법을 이용한 산불피해강도 분류 - 2020년 안동 산불을 사례로 -

Abstract

The increased frequency and intensity of wildfires can cause damages to the ecosystem and the atmospheric environment. Rapid identification of the wildfire damages is also important for establishing forest restoration, budget planning, and human resources allocation. Because the wildfires need to be examined for vast areas, satellite remote sensing has been adopted as an effective method. Many studies for the detection of wildfires and the analysis of burn severity have been conducted using mid- and high-resolution images. However, they had difficulties in the sensitivity problem of NBR (Normalized Burn Ratio) for multi-temporal images. This paper describes the feasibility of the detection and classification of wildfire burn severity using Sentinel-2 images with K-means and ISODATA (Iterative Self-Organizing Data Analysis Techniques Algorithm) methods for a case of the Andong fire in April 2020. The result can be a reference to the appropriate classification of large-scale wildfire severity and decision-making for forest restoration planning.