姚天成 王巍
摘要:研究并實現(xiàn)了一種對于實時三維立體圖像渲染的視點相關(guān)透視改進算法,算法使得計算機輸出的三維圖像在傳統(tǒng)顯示器和投影儀設(shè)備等二維平面顯示設(shè)備上,呈現(xiàn)出關(guān)于任意觀察者頭部位置進行透視矯正的立體圖像。使用微軟的Kinect for Windows設(shè)備來捕捉觀察者的視點位置,使用Unity 3D游戲引擎渲染三維場景,使用兩臺投影儀分別在兩個互相垂直的投影面上投影出拼接后的透視矯正的立體圖像。實驗結(jié)果表明本文算法視點相關(guān)的渲染效果具有魯棒性和實時性。
關(guān)鍵詞:立體圖像;頭部相關(guān)透視;虛擬現(xiàn)實; Unity 3D; Kinect for Windows
中圖分類號:TP41 文獻標(biāo)識碼:A 文章編號:1009-3044(2017)01-0175-04
Research and Implementation of Head-coupled Perspective Algorithm for Real-time Stereo Rendering
YAO Tian-cheng, WANG Wei
(Department of Computer Science and Engineering, Dalian Nationalities University, Dalian 116600, China)
Abstract: This article presents an implementation of head-coupled perspective technique which improves traditional computer generated real-time 3D stereo image rendering. This technique enables traditional planar display device such as monitors and projectors to have the ability to present perspective correct stereo image for the observer in arbitrary view positions. In the demonstration, a Microsoft Kinect for Windows is used to capture the observers head position. A PC with the game engine Unity 3D installed is used to take the observers head position data from Kinect as input and render out perspective correct stereo image. Finally, two projector are utilized to project two sub-images output from the rendering PC onto two perpendicular surface respectively to expand display surface area. Experimental results shows real-time and robustness of the proposed algorithm.
Key words:stereo rendering; head-coupled perspective; virtual reality; Unity 3D; Kinect for Windows
1 介紹
1.1 背景
立體顯示技術(shù)[1]已經(jīng)被廣泛運用于生活中的各個領(lǐng)域。這種技術(shù),一些算法中觀察者需要佩戴特制的眼鏡,如分色鏡片,偏振鏡片等,另一些算法則不需要。不過算法本質(zhì)都是令每只眼睛分別看到由顯示屏輸出的帶有瞳距偏移的圖像,從而在大腦中重組獲得帶有深度的立體圖像。
屏幕上的三維圖像可以是事先由計算機生成或者使用立體攝像機拍攝好的,比如說3D電影;也可以是由計算機實時生成,比如說電子游戲、虛擬現(xiàn)實。然而,這些傳統(tǒng)的算法中,無論立體圖像是事先制作,還是計算機實時生成,都假設(shè)了用戶的觀察位置固定于屏幕中心前方的一定距離。當(dāng)觀察者的觀察位置處于這個假設(shè)的觀察點附近時,這種圖像獲得較好的立體效果,一旦觀察者遠離這個假設(shè)的觀察位置,立體效果就會有損失,對于觀察者而言,畫面會出現(xiàn)不期望的變形。……