光源感測元件對色彩頻譜之響應及穩定性研究

Abstract

色域空間的色彩校正,可以直接以矩陣運算來達成。線性的矩陣轉換可 以看成是 對掃瞄器內部濾波器的重新組合。一般的掃瞄器取像系統均 以三個濾波器取出代 表性的色彩。因此,其所需的線性校正函數為一 個3*3的矩陣。假使,使用高次項 的訊號作為修正,則所需要的矩陣即 為3*P,P代表所使用的訊號項數。在使用高次 項修正時,內部濾波器的 重新組合,就會跟測試樣本成級數的相關。也就是測試樣 本,不只決定 濾波器的組合比例,也影響到濾波器的組合方式。本論文,即在此比 較 線性組合與高次項組合之優劣。 The color correction of color gamut can be accomplished directly by matrix operations. Linear matrix transforms may be regarded as recom- binations of the spectral responses of the scanners. Common scanners usually process color signals via three filters corresponding to the three tristimulus values. Thereby, a 3 by 3 matrix representing the color correction function is needed for the scanners. If higher order polynomials are used, a 3 by P matrix is needed, where P designates the number of the terms of the polynomials. When we use the higher or- der polynomials, we would find that recombinations of the spectral re- sponses of the filters of the scanners are highly correlated with the samples. In other words, the samples affect not only the proportions of the recombinations of the spectral responses of the filters, but also how these spectral responses combine. This thesis compares linear combination with higher order polynomials.