干旱区生态环境知识资源中心(Arid): Semi-Supervised Data Summarization: Using Spectral Libraries to Improve Hyperspectral Clustering

Arid

报告编号	IPN-PR-42-163
来源ID	NTRS_Document_ID: 20060008090
	Semi-Supervised Data Summarization: Using Spectral Libraries to Improve Hyperspectral Clustering
	Wagstaff, K. L.; Shu, H. P.; Mazzoni, D.; Castano, R.
英文摘要	Hyperspectral imagers produce very large images, with each pixel recorded at hundreds or thousands of different wavelengths. The ability to automatically generate summaries of these data sets enables several important applications, such as quickly browsing through a large image repository or determining the best use of a limited bandwidth link (e.g., determining which images are most critical for full transmission). Clustering algorithms can be used to generate these summaries, but traditional clustering methods make decisions based only on the information contained in the data set. In contrast, we present a new method that additionally leverages existing spectral libraries to identify materials that are likely to be present in the image target area. We find that this approach simultaneously reduces runtime and produces summaries that are more relevant to science goals.
英文关键词	CLUSTER ANALYSIS BANDWIDTH IMAGERY PIXELS SPECTRA LIBRARIES
出版年	2005
报告类型	Technical Report
语种	英语
国家	美国
URL	http://hdl.handle.net/2060/20060008090
资源类型	科技报告
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/259324
推荐引用方式 GB/T 7714	Wagstaff, K. L.,Shu, H. P.,Mazzoni, D.,et al. Semi-Supervised Data Summarization: Using Spectral Libraries to Improve Hyperspectral Clustering,2005.