SEISMOLOGY AND GEOLOGY ›› 2014, Vol. 36 ›› Issue (1): 137-147.DOI: 10.3969/j.issn.0253-4967.2014.02.011

Previous Articles     Next Articles

REMOTE SENSING DETECTION OF VOLCANIC ASH CLOUD USING INDEPENDENT COMPONENT ANALYSIS

LI Cheng-fan, DAI Yang-yang, ZHAO Jun-juan, YIN Jing-yuan, ZHOU Shi-qiang   

  1. School of Computer Engineering and Science, Shanghai University, Shanghai 200444, China
  • Received:2013-05-12 Revised:2013-07-31 Online:2014-03-30 Published:2014-04-08

利用独立分量分析进行火山灰云遥感检测

李成范, 戴羊羊, 赵俊娟, 尹京苑, 周时强   

  1. 上海大学计算机工程与科学学院, 上海 200444
  • 通讯作者: 尹京苑,教授,E-mail:jyyin@staff.shu.cn
  • 基金资助:
    国家自然科学基金(41172303)资助

Abstract: The volcanic ash cloud is mainly composed of volcanic ash debris and gases. The adequate mixture of the two can form acidic aerosols. It not only causes the major global climate and environmental changes, but also seriously threatens the aviation safety. Remote sensing can quickly and accurately obtain the information of the surface's and the atmosphere's changes; therefore it is playing an important role in the monitoring of volcanic activity. In recent years, with the advancement of sensor technology, the thermal infrared remote sensing technology has become an important means of detecting the volcanic ash cloud. However, due to the large amount of spectral bands and data, the remote sensing data have pretty strong band correlation and obvious information redundancy problem, all of which have decreased to a certain degree the detecting accuracy of volcanic ash cloud. Therefore, it is necessary to introduce new data processing methods into the volcanic ash cloud remote sensing detection field. Principal component analysis(PCA)can compress a large number of complex information effectively into a few principal components; as a result, it is widely applied in the data compression and hyperspectral remote sensing field. Independent component analysis(ICA)is a recently developed new data processing method which can linearly decompose the observed data into mutually dependent components, and achieve the decorrelation and redundancy elimination of remote sensing data; so it has certain potential in volcanic ash cloud detection. A remote sensing detecting algorithm of volcanic ash cloud, which uses ICA method, is proposed after the exploration of the physics and chemical properties of volcanic ash cloud. This paper takes the MODIS remote sensing image of Iceland's Eyjafjallajokull volcanic ash cloud on April 19, 2010 as data source. It uses ICA in volcanic ash cloud detection on the basis of the principal component analysis(PCA)processing of MODIS image, and gives comparison among these following parties: the detected results, the relevant research results, United States Geological Survey(USGS)standard spectral database and SO2 concentration distribution. The results show that: ICA can successfully obtain the information of the volcanic ash cloud from MODIS image; the detected volcanic ash cloud has a good consistency with the USGS standard spectral database and the SO2 concentration distribution, thus, it can obtain pretty good detection results.

Key words: thermal infrared remote sensing, principal component analysis(PCA), independent component analysis(ICA), volcanic ash cloud, aviation safety

摘要: 火山灰云不但引起全球气候和环境系统的重大变化,而且还会威胁航空安全。热红外遥感技术为检测火山灰云提供了新手段,但是遥感数据自身的冗余和波段相关性大大降低了火山灰云的检测精度。独立分量分析(Independent Component Analysis,ICA)能够实现遥感数据的去相关和消除冗余,在火山灰云检测中具有一定的潜力。通过探索火山灰云的物理、化学性质,文中以2010年4月19日冰岛艾雅法拉(Eyjafjallajokull)火山灰云MODIS图像为数据源,在对MODIS数据进行主成分分析处理的基础上,利用ICA进行火山灰云检测。结果表明:ICA能够较好地从MODIS图像中获取火山灰云信息,所得结果与美国地质调查局标准光谱数据库和火山灰云SO2浓度分布具有较好的一致性,取得了较好的检测效果。

关键词: 热红外遥感, 主成分分析, 独立分量分析, 火山灰云, 艾雅法拉火山

CLC Number: