因目前3D繪圖軟體的成圖方法，都源自1986年Kajiya提出的成圖方程式，對“不符合人類真實環境光照原理”因素的問題都沒解決，根本無法準確的運用在色彩配色工作上。而諸如室內空間配色、產品造形配色、色彩研究配色、甚至包裝設計配色，都需藉助3D電腦成圖模型達到配色目的。所以本文研究目的，是試圖找出3D電腦成圖與真實環境間的色彩誤差關係，然後提出簡易的修正模式。未來只須將該色彩修正模式，導入各領域相關的配色研究中，就能簡易的藉助3D電腦繪圖模型，模仿真實環境，進行正確的配色工作。研究方法分三階段進行，第一階段，建立符合研究目的的3D虛擬實驗模型；第二階段，檢測3D電腦成圖色彩與真實環境色彩是否發生誤差，經檢測結果，發現3D電腦成圖色彩與真實環境色彩間，明度存著明顯的誤差；於是進行第三階段研究，試圖找出3D電腦成圖色彩的明度誤差修正方法。經研究後，計有下列發現：．明度太暗(B≦15)的3D成圖色彩，會呈現錯誤的色相H數位值或錯誤的彩度S數位值。．除了明度太暗(B≦15)的色彩，其餘3D成圖色彩與真實環境色彩的色相、彩度都幾乎相同。．3D電腦成圖色彩與真實環境色彩間的明度關係，發生等差比例的誤差。．結論建議再根據明度誤差的研究結果，提出3D電腦成圖色彩的明度修正模式： y=x(1+tanφ)-d
The rendering method of 3D painting software in these days primarily origins from rendering equation proposed by Kajiya in 1986. The 3D rendering model is required for matching color in the following design, i.e., interior space, products, package, and matching color studies etc. The color brightness from the rendering method is unable to match with original lights in real circumstances, nor can it be used in matching color accurately. In this respect, this study tries to locate the color difference as occurred between 3D rendering method and real circumstances. The study proposes a model to eliminate the difference. It argues that in the future we can get the exactly color by 3D painting model with applying correct model to all kinds of matching color work. The process of this study involves three stages. In the first stage we establish 3D virtual models for experiment. The second stage we discover the color difference between 3D rendering method and real circumstances, and then propose the correct model. The third stage we try to find out how to eliminate the difference of 3D render color brightness. About all, it concludes with the following findings: . When brightness is too weak (B≦15) 3D rendering method will occur color difference. . 3D render color hue and saturation are the same as real circumstances except the brightness is too weak (B≦15). . There is a symmetrical difference in color brightness between 3D render color and real circumstances. . This study proposes the correct model as following: y=x(1+tanφ)-d
