Estimation of Spectral Reflectances from Multi-Band Images by Multiple Regression Analysis

Norimichi TSUMURA, Hideaki HANEISHI, and Yoichi MIYAKE
 
Department of Information and Image Sciences, Chiba University, 1-33, Yayoi-cho, Inage-ku 263-8522
 

Japanese Journal of Optics Vol. 27 No. 7 pp. 384-391(1998).

Abstract
 In this paper, we applied the multiple regression analysis to the estimation of spectral reflectance from multi-band image. The estimation of spectral reflectance is required in digital museum, telemedicine, internet shopping and so on, because the information of spectral reflectance is not dependent on the spectral characteristics of taking illuminant and imaging systems. It is necessary to estimate it from smaller bands of images for the limitation of capacity and transmission speed. By using the higher order terms of pixel values, the proposed method could estimate the spectral reflectance more accurately than the conventional estimation methods such as low-dimensional linear approximation method and Wiener method.

Text and Figures
[PDF] [PDF figures](in Japanese)

Related menuscripts
[PDF] ( in Japanese ) ’Ѻ“¿“¹ChƒXƒyƒNƒgƒ‰ƒ‹‰æ‘œŒv‘ª–@‚Ìi“WChŒõ‹Zp“®Œü’²¸•ñ‘XV@pp.407-414@iŒõŽY‹Æ‹ZpU‹»‹¦‰ïC‚P‚X‚X‚X”N‚RŒŽj

Back