基于棱柱镜技术的自由立体显示图像合成  

Image Synthesizing for Autostereoscopic Display Based on Lenticular Technology

作　　者：杨蕾[1] 宋晓炜[1] 侯春萍[1] 戴居丰[1] 

机构地区：[1]天津大学电子信息工程学院,天津300072

出　　处：《天津大学学报》2007年第9期1105-1110,共6页Journal of Tianjin University(Science and Technology)

基　　金：国家自然科学基金资助项目(60672077)

摘　　要：根据棱柱镜的光学和物理结构以及LCD子像素的排列特征,提出了一种用于棱柱镜自由立体LCD显示的多视点合成新方法.该方法属于数字合成算法,由子像素判断准则、视点视图下采样以及多视点子像素的排列3部分组成.将该方法与传统模拟光筛合成方法和常规数字合成方法进行比较,实验结果证明新方法在合成速度和观看效果上优于后两种方法.同时,由于该方法对各视点视图进行了下采样,大幅减少了原始数据量,十分有利于压缩.Based on the optical characteristic and physical structure of lenticular as well as the alignment characteristic of LCD sub-pixels, a multiview synthesizing method for lenticular autostereoscopic LCD display was proposed. The method belongs to digital synthesizing method and is composed of three parts： sub-pixel judgment rule, view down-sampling and alignment of multiview sub-pixel. By comparison, the proposed method is more advantageous than the traditional analog mask synthesizing method and the general digital synthesizing method both in synthesizing efficiency and viewing effect, Furthermore, since down-sampling is made for each view, the proposed method reduces the original amount of data, which is beneficial to compression.

关 键 词：棱柱镜 自由立体 多视点 合成 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]