Noise estimating and filtering of hyperspectral infrared radiance

Yanni Qu; Mitchell D. Goldberg

doi:10.1117/12.467645

16 June 2003 Noise estimating and filtering of hyperspectral infrared radiance

Yanni Qu, Mitchell D. Goldberg

Proceedings Volume 4897, Multispectral and Hyperspectral Remote Sensing Instruments and Applications; (2003) https://doi.org/10.1117/12.467645
Event: Third International Asia-Pacific Environmental Remote Sensing Remote Sensing of the Atmosphere, Ocean, Environment, and Space, 2002, Hangzhou, China

Abstract

Atmospheric InfraRed Sounder (AIRS) is a hyper-spectral infrared instrument on the EOS-Aqua satellite. Principal Component Analysis (PCA) of realistic simulations of AIRS radiances, which includes the effects of variable clouds and surface parameters, can be used to estimate and filter AIRS instrumental noise. The PCA uses noise scaled radiance, i.e. radiance divided by noise (R/N). Since the square root of the eigenvalues is equivalent to the standard deviation of the principal component score of the dependent ensemble, the square root of the eigenvalues in the R/N domain can be interpreted as a signal to the noise ratio for the new principal components (PC) or say "abstract channels". New PCs are arranged from the largest to the smallest eigenvalues. Once the R/N ratio is below unity, the signal has less contribution than noise for that PC and all the remaining PCs. This physical meaningful fact can be a criterion for radiance noise filtering. Using the linearity of PCA to the pure signal S plus ideal noise N in the R/N domain, PCA(N/N) becomes PCA(I). I is the identity matrix and hence eigenvalues of I are all equal. If reconstructing I, only k/m information can be recovered by k eigenvectors, where k = 1 .... m eigenvectors. Thus, it is easy to know how much noise merged into the reconstructed radiance and how many eigenvectors are needed for filtering noise when performing PCA in R/N domain. The number of eigenvectors to be used for noise estimating is around 60 in all sky conditions. 60 eigenvectors can filter over 97% of noise. Only 3% noise remains in the reconstructed data. Signal information can be recovered with accuracy in noise level. In addition, there is less than 5% error using PCA to estimate noise. Preliminary results from the real AIRS observation are discussed briefly.

Citation Download Citation

Yanni Qu and Mitchell D. Goldberg "Noise estimating and filtering of hyperspectral infrared radiance", Proc. SPIE 4897, Multispectral and Hyperspectral Remote Sensing Instruments and Applications, (16 June 2003); https://doi.org/10.1117/12.467645

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
12 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Interference (communication)

Signal to noise ratio

Principal component analysis

Infrared radiation

Computer simulations

Error analysis

Calibration

Show All Keywords

Keywords/Phrases

Search In:

Publication Years