Abstract:

Taking the forest fire in Sumatra, Indonesia on October 15, 2015 as an example, method to extract forest fire hotspots using the new-generation geostationary satellite―Himawari-8 data was explored via clear sky analysis, potential fire pixel identification, sunglint test and background analysis. Within the temporal and spatial range of this study, 10109 Himawari-8 hotspots were extracted in all (142 significant instants), and the maximum number of hotspots for a single moment was 267 (03:40LT). Through quantitative comparison of time-space distributions, it was found that compared with MODIS hotspots, hotspots derived from Himawari-8 data were of smaller amount and more dispersed spatial distribution. However, two kinds of hotspots were very similar in overall spatial distribution. The characteristics of spatial distribution and quantity of Himawari-8 hotspots were consistent with fire information reflected by the true-color image of Himawari-8. In general, extracted Himawari-8 hotspots could well reflect the information of forest fire and provide useful information in forest fire mornitoring.

Key words: Himawari-8, forest fire, hotspot detection

摘要：

以2015年10月15日印度尼西亚苏门答腊岛森林火灾为例, 探索通过云水检测、潜在火点提取、太阳耀斑剔除、背景窗口确定等步骤, 利用新一代地球同步气象卫星——Himawari-8卫星数据提取森林火灾火点的方法。在所研究的时空范围内, 共提取Himawari-8火点10109个(有效时刻数142个), 单一时刻火点数最高为267个(当地时间03:40)。通过对时空分布的定量化比较发现, 与相应时刻的MODIS火点相比, 所提取的Himawari-8火点数量较少, 空间分布更分散, 但二者的空间分布在总体上非常相似, 且所提取的Himawari-8火点空间分布及数量特征与Himawari-8真彩色图反映的森林火情较为吻合, 说明利用该方法提取的Himawari-8火点可以较好地反映森林火情, 能够为森林火灾监测提供有效信息。

关键词: Himawari-8, 森林火灾, 火点探测