烟雾
植被(病理学)
环境科学
远东
遥感
卫星
卫星图像
地理
气象学
医学
考古
病理
航空航天工程
工程类
作者
Koichi Kawano,Naoyuki Komatsu
标识
DOI:10.1109/igarss.2012.6352256
摘要
In the Russian Far East, forest fires appear every year. Russia is the greatest forest possession country. It is difficult to estimate the damaged area by the forest fire because the Russian Far East is very vast. The image analysis that uses the satellite imagery is suitable for understanding the actual situation of forest fires. In the image analysis on a forest fire that uses the satellite image, each of scar, smoke, and hot spot is generally used for the analysis method. The scar is suitable for estimating the burned area after extinguishing forest fires. However, scars are not detected when they covered with the cloud. Moreover, scars are not detected when vegetation recovers. To detect smoke is effective to early alert of forest fires. However, the smoke is not detected when a fire is extinguished. Moreover, even if smoke is detected, the total burnt area is not estimated. In this paper, we proposed smoke and scar detection methods, and analyzed seven-year data of the Russian Far East by using NOAA AVHRR images. As a result of the analysis, scars have been detected in April and October, and smoke has been detected in around May every year. In addition, occurrence of a fire is affected by the scar which appears in October of the previous year. The spot of the scar burns again next spring because the vegetation of the scar recovers in May. Moreover, a forest fire occurs every year in the territories of Khabarovsk and the Zeya River.
科研通智能强力驱动
Strongly Powered by AbleSci AI