基于动态随机点立体图的立体视觉深度运动特性分析  

Depth motion characteristics analysis of stereo vision based on dynamic random dot stereograms

作　　者：林丽媛[1] 侯春萍[1] 王凯[1] 周志远[2] 

机构地区：[1]天津大学电子信息工程学院,天津300072 [2]天津津航技术物理研究所,天津300308

出　　处：《红外与激光工程》2015年第4期1365-1369,共5页Infrared and Laser Engineering

基　　金：国家高技术研究发展计划(863计划)(2012AA03A301);教育部博士点基金(20130032110010)

摘　　要：利用动态随机点立体图,排除心理暗示线索对深度运动的影响,对立体视觉深度运动特性进行了研究。利用主观实验,分析了深度运动物体的运动位置、运动方向和运动速度等运动特性对于立体视觉舒适度(Stereo visual comfort,SVC)的影响,并提出深度运动情况下的SVC与运动速度的预测模型。实验结果表明,动态随机点立体图中,运动物体的深度运动诱发了透视效果;在交叉视差(屏幕前)和非交叉视差(屏幕后)情况下,立体视觉深度运动导致的SVC不同。对计算得到的SVC的预测值和实测值进行相关性检测,三种深度运动的皮尔森相关系数(Pearson correlation coefficient,PCC)分别为0.956、0.972、0.977,充分表明该预测模型能够准确地预测SVC。Dynamic random-dot stereograms which can eliminate the effects caused by psychological clues of stereo vision were used to study the depth motion characteristics of stereo vision. The effects on Stereo Visual Comfort（SVC） of the depth motion characteristics including motion positions, motion directions and motion velocities were analyzed by subject experiments and a prediction model of visual comfort and motion velocity during depth motion was built. The results show that depth motion induced perspective and depth motion make different visual comforts under crossed disparity（before the screen） or uncrossed disparity（after the screen） in dynamic random dot stereograms. The Pearson Correlation Coefficient（PCC）of three types of depth motion achieves 0.956, 0.972, 0.977, respectively, which indicates that the prediction model can more accurately predict the SVC.

关 键 词：立体视觉 深度运动 SVC 动态随机点立体图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]