基于Contourlet变换和主成分分析的高光谱数据噪声消除方法

常威威; 郭雷; 刘坤; 付朝阳

doi:10.3724/SP.J.1146.2008.01675

基于Contourlet变换和主成分分析的高光谱数据噪声消除方法

doi: 10.3724/SP.J.1146.2008.01675

基金项目:

国家自然科学基金(60802084)资助课题

计量
- 文章访问数: 3473
- HTML全文浏览量: 85
- PDF下载量: 1348
- 被引次数: 0
出版历程
- 收稿日期: 2008-12-10
- 修回日期: 2009-05-15
- 刊出日期: 2009-12-19

Denoising of Hyperspectral Data Based on Contourlet Transform and Principal Component Analysis

摘要

摘要: 该文提出了一种适合于高光谱超维数据处理的基于Contourlet变换和主成分分析的噪声消除方法。该方法首先利用Contourlet变换实现图像的稀疏表示，再利用主成分分析对Contourlet系数进行适当地消噪处理。通过对OMIS图像的实验结果表明该方法能够同时消除高光谱多个波段图像中的噪声，从整体上改善高光谱图像质量，且性能上要优于PCA和Contourlet变换方法。
- 图像处理; 高光谱遥感; 去噪; Contourlet变换; 主成分分析
Abstract: This paper proposes a denoising method of hyperspectral super-dimensional data based on Contourlet transform and principal component analysis. At first the sparse representation of images is accomplished with Contourlet transform. Then the Contourlet coefficients are processed with principal component analysis. The experimental results based on OMIS images show that the proposed method can simultaneously eliminate noises in multi-band hyperspectral images, improve the quality of the whole hyperspectral data and outperforms methods based on PCA and Contourlet transform respectively.

HTML全文

参考文献(1)

浦瑞良, 宫鹏. 高光谱遥感及其应用. 北京: 高等教育出版社,2000, 第1 章.Pu Rui-liang and Gong Peng. Hyperspectral Remote Sensingand Its Applications[M]. Beijing: Higher Education Press,2000, Chapter 1.[2]Vaiphasa C. Consideration of smoothing techniques ofhyperspectral remote sensing[J].ISPRS Journal ofPhotogrammetry and Remote Sensing.2006, 60(2):91-99[3]Torrecilla E, Aymerich I F, and Pons S, et al.. Effect ofspectral resolution in hyperpspectral data analysis[J].Geoscience and Remote Sensing Symposium, 2007, 23(28):910-913.[4]陈志刚, 束炯. 高光谱图像光谱域噪声去除的经验模态分解方法[J]. 红外与毫米波学报, 2008, 27(5): 378-382.Chen Zhi-gang and Shu Jiong. Empirical mode decompositionon removing spectral noise in hyperspectral image[J]. Journalof Infrared and Millimeter-wave, 2008, 27(5): 378-382.[5]Backer S D, Piurica A, and Huysmans B, et al.. Denoising ofmulticomponent images using wavelet least-squaresestimators[J].Image and Vision Computing.2008, 26(7):1038-1051[6]Rakwatin P, Takeuchi W, and Yasuoka Y. Stripe noisereduction in MODIS data by combining histogram matchingwith facet filter[J].IEEE Transactions on Geoscience andRemote Sensing.2007, 45(6):1844-1856[7]Stephan K, Hibbitts C A, and Hoffmann H, et al.. Reductionof instrument-dependent noise in hyperspectral image datausing the principal component analysis: Applications toGalileo NIMS data[J].Planetary and Space Science.2008,56(3):406-419[8]McCabe G P. Principal variables[J].Technometrics.1984,26(2):137-144[9]Ramakrishna B, Wang Jing, and Chang C I, et al..Spectral/spatial hyperspectral image compression inconjunction with virtual dimensionality. Algorithms andtechnologies for multispectral, hyperspectral, andultraspectral imagery XI, Proc. 5806, SPIE, 2005: 772-781.[10]Do Mi N and Vetterli Ma. The Contourlet transform: Anefficient directional multiresolution image representation[J].IEEE Transactions on Image Processing.2005, 14(2):2091-2106[11]张瑾, 方勇. 基于分块Contourlet 变换的图像独立分量分析方法[J].电子与信息学报.2007, 29(8):1813-1816浏览[12]张晶, 方勇华. 基于Contourlet 变换的遥感图像去噪新算法[J]. 光学学报, 2008, 28(3): 462-466.Zhang Jing and Fang Yong-hua. Novel denoising method forremoter sensing image based on Contourlet transform. ActaOptica Sinica, 2008, 28(3): 462-466.[13]Li Shu-tao, Kwok J T, and Wang Yao-nan. Multi-focus imagefusion using artificial neural networks[J].Pattern RecognitionLetters.2002, 23(6):985-997

施引文献

资源附件(0)

访问统计