摘要
立体印刷图像通过柱透镜光栅阵列进行分像,形成左右视觉图像而产生立体感觉,因而,图像的分像处理就是立体印刷图像的关键。通过摄影得到的平面图像不包含左右视图而失去了深度信息,因此,用单张平面图像不能获得真实感的三维信息,通过图像处理一般只能得到有限层的假立体图像,而不能得到真实立体感的图像,必须增加额外的约束条件。在已知目标物体表面形状函数的特定情况下可以进行三维形状重构处理,得到真实感立体图像。对该处理方法进行了讨论,该方法可以应用于立体印刷图像的处理和三维场景的再现。
The principle of stereo printing is based on the binocular vision and the lenticular is needed to separate left and right view images. The key of getting 3D images is the image processing algorithm to obtain left and right view images. Because 2D image loses depth information, 3D objects can not be reconstructed by a single image, or only discontinuous layers can be gotten. A method to reconstruct 3D information and get realistic stereo vision images is proposed in the paper in case of the surface function of the object is known, and the algorithm is discussed. This image processing method is relative simple and useful for stereo printing.
出处
《北京印刷学院学报》
2009年第4期7-9,共3页
Journal of Beijing Institute of Graphic Communication
基金
北京市信号与信息处理重点建设学科资助项目(01422313010006-77)
关键词
图像处理
三维重构
单图像模型
立体视觉
立体
印刷
image processing
3D reconstruction
single image model
stereo sensation
3D printing
