一种基于微阵列镜头的多目标的图像置换方法

A Multi Target Image Replacement Method Based on Microarray Lens

本文提出了一种基于微阵列镜头的利用图像深度信息进行图像置换的算法.该算法利用基于马尔科夫随机场理论的图割算法构造能量函数,恢复所拍摄场景的深度图,再根据图像中目标物体所在深度层范围进行置换,从而在保留置换部分完整图像的同时,实现微阵列镜头的图像置换功能.实验结果证明,相较于局限在二维平面且交互复杂的传统图像置换技术,本文算法能够准确提取置换图像,并在置换后不损坏图像本身的内容,能有效地处理图像中物体的前后遮挡问题,减少繁琐的交互操作及对操作人员经验的依赖.

A new image replacement algorithm based on microarray lens through using the depth information of image is proposed in this paper.With the array image,it constructs the depth map using the energy function with Graph Cuts algorithm based on the Markov random field theory,and it replaces the objects according to their depth layer range.The experimental results demonstrate that,compared with the traditional image replacement techniques which is confined to two-dimensional plane and has complex interactive operation,our method can extract the exact replacement image,and the content of the image itself will not be damaged after the replacement.Besides,it effectively deals with the front and rear occlusion problems in image and requires less interactive operations and operator＇s experiences.