Knowledge Management System Of Guangzhou Institute of Geochemistry,CAS
Subpixel-based target detection in hyperspectral imagery with pairwise coupling Support Vector Machines | |
Li H; Wang YP; Li Y; Cao Y | |
Source Publication | International Conference on Informational Technology and Environmental System Science |
2008-05-15 | |
Conference Date | 2008-5-15 |
Conference Place | Jiaozuo, PEOPLES R CHINA |
Indexed By | SCI |
Language | 英语 |
Abstract | Interpretation of mixed pixels has become a key factor in the analysis of hyper spectral imagery. In this paper, an extended Support Vector Machines (SVM) method is designed to estimate abundance fractions of materials present in an image pixel. Tradition Support vector classification predicts only class labels. For getting accurate amounts of abundance material, SVM method is combined with pairwise coupling. The method estimates probabilities which are used as the proportions of materials. There are two constraints imposed on probabilities: the class probabilities sum to one, and class probabilities are non negativity. To demonstrate the usefulness and acceptability of the method, experiments are conducted on simulated hyperspectral data. The results show that the application of extended SVM Method to subpixel-based target detection is effective and feasible. |
Document Type | 会议论文 |
Identifier | http://ir.gig.ac.cn/handle/344008/9384 |
Collection | 中科院广州地化所(-2008) |
Recommended Citation GB/T 7714 | Li H,Wang YP,Li Y,et al. Subpixel-based target detection in hyperspectral imagery with pairwise coupling Support Vector Machines[C],2008. |
Files in This Item: | There are no files associated with this item. |
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment