色溫調整影像合成技術之研究

Abstract

影像合成技術在數位內容中扮演著重要的角色，為了呈現出高品質的畫面，本研究從影像色溫調整技術切入影像合成的部分，首先偵測影像拍攝環境之光源色溫，其中包含影像亮部光源偵測方法、色域重疊度光源偵測方法、色相分割多色色溫軌跡光源偵測方法及無色相分割多色色溫軌跡光源偵測方法等四種色溫偵測方法，以改變光源色溫的概念進行影像色彩處理，其後將欲合成之影像色溫調為一致再進行合成，排除影像合成後光影不自然的問題，實驗證明本研究發展的影像合成技術可以合成出自然的影像畫面。

Image synthesis technology plays an important role in digital content creation. In order to present the image with much higher quality, the developed technology is based on color temperature adjustment of the foreground and background images. This research includes two parts. First of all, the color temperature of the illuminant in the shooting environment is estimated. Several methods of color temperature estimation are derived, including the highlight estimation, gamut estimation, hue-segment color temperature loci estimation, and hue-non-segment color temperature loci estimation. Secondly, by the concept of changing illuminant color temperature of images, adjusting the color temperature of the images to be the same and synthesize them. The proposed method can avoid the problem of unnatural image synthesis. Our method is proved by the research to create natural image synthesis results.