Color Reproduction Research Of Multispectral Images And The Application In Imaging Spectrometer

The Information in Multispectral images is more abundant than in ordinary images. Except the two-dimensional image, there is multi-dimensional spectral information. The spatial image dimensional describes two-dimensional space characteristics of target. The spectrum dimensional shows spectral curves feature of pixel of images. Therefore, multispectral images have been used in various fields widely, but multispectral images are gray scale images. Due to lack of the color information, we cannot observe them intuitively. So we need to reproduce the true color of multispectral images and used it in imaging spectrometer.The pixel level fusion of the false color was used to visualize the multispectral images. However, false color just is qualitative description, lacking of quantitative description. The reproduction of image true color in this thesis was based on CMFs of 1964 CIE was proposed, and also used other methods. We also discussed the evaluation methods of multispectral images, and choose the suitable evaluation criteria such as cross entropy,RMS error,Peak Signal to Noise Ratio(PSNR) . A comparative analysis revealed CMFs is the most effective processing method, and analysis the Influence of different factors. Then the reproduction of near infrared band was proposed, got the reproduction image which can distinguish the color characteristics of objectsThe software components in this thesis were based on VC 6.0++.The system not only need to open,save,deal with,close the images, also processes the image date,reproduce the multispectral images and display the results. Then, the software was developed based on VC 6.0++. The software supply friendly man-machine interface can draw the pixel curve of multispectral images and output the reproduction image.Finally, we draw the pixel curve of multispectral images and got the reproduction image. by means of comparing the results of different reduction method, we can find CMFs is the most effective method. Then, we analysis the influence of different spectral situation for reproduction.