Multispectral Image Acquisition Method Based On The Three Color Camera

In recent years, with the rapid development and wide application of Colorduplication technology and digital imaging devices, standard of color reproductionaccuracy become very high, and the spectral image has become a research trend. Colorreproduction technology based on the spectrum is spectral matching, which requiresspectral characteristics of reproduction color same as target color, that mean they canmaintain the color color reappearance performance in any conditions. Spectral colorreproduction technology can meet the requirements of spectral color reproduction, it uses6-31spectral channels of the visible light range to capture color information, rather thanuses red, green and blue channel. Spectral reflectance or transmittance of the object as thesole criterion for calibration of color, and through the reconstruction of the spectrum andforecast ink trapping to achieve precise control of color or multi-spectral imagereconstruction.The main purpose of this paper is to build a multi-channel multi-spectral imagereconstruction system based on three-color camera. And complete the conversion of thedigital response value obtained by digital cameras to the spectral reflectance of the color.Truly achieve the color copy effect of "what you see is what you get" in color copyingwhole process, make the color reproduction independent of equipment, media, standardillumination and observer.This paper has fitted the conversion equation of the digital signal value and the lightradiation energy suffered CCD. The accuracy of the spectral reconstruction of the systemhas been largely improved by linear correction processing on the digital value.System uses the calibration date to do two reconstruction experiments which respectivelybased on the pseudo-inverse method and basis function. And by the comparison methodof calculation of the basis functions determined using singular value decomposition (SVD)method to solve. Through the analysis the results of two spectral reconstruction found thatthe average root mean square error of24blocks of color reconstructed by the spectralreconstruction experiment based on pseudo-inverse is0.044, and the minimum colordifference is0.94, the average color difference is5.273. the average root mean squareerror of24blocks of color reconstructed by the spectral reconstruction experiment basedon basis function is0.044,and the minimum color difference is1.03,the average colordifference is5.3.The experimental results show that this spectral reconstruction systemhas a higher feasibility and accuracy, can also prove that multi-channel spectral reconstruction system based on three-color camera implemented in this article have betterperformance and higher application value.Finally，the main conclusions of the multi-channel spectral reconstruction systemproposed in this dissertation were summarized, and the application trend of multi-spectraltechnology in high-fidelity printing, cross-media publishing and color management in thefuture were also forecasted.