彩色影像處理流程設計

Abstract

本篇論文提出一個彩色影像處理流程，將CCD或CMOS感測器上擷取到的原始資料處理為正確顏色的照片。儘管有許多個別的彩色影像處理方法已經被發表，但卻沒有優良的彩色影像處理流程被提出。在此，我們不但提出一個彩色影像處理流程，以及色彩及色階重現演算法。它為彩色影像處理系統中的理論影像科學和實際影像處理流程之間的缺口建立起橋樑。而經由實驗的結果可得知，提出的影像處理流程和新的色彩處理演算法在大部份的場景以及光源下，都可以得到不錯的結果。This thesis presents a new color image processing pipeline (IPP) which processes the image raw data captured from CCD/CMOS sensors to the final color and tone corrected picture. Although many individual image processing steps have been well addressed, very few good image pipeline designs proposed to integrate these processing stages. In this thesis, we present a new IPP as well as several color and tone reproduction algorithms. It bridges the gap between theoretic color imaging sciences and the practical IPP implementation issues of digital imaging systems. The demonstrations by processing different scenes of pictures show that the proposed pipeline and several new color processing algorithms runs well in diversified scenes and illuminants.