基于共聚焦图像序列的深度估计方法

被引：1

作者：

吴文杰

肖照林

王庆

杨恒

机构：

[1] 西北工业大学计算机学院

来源：

计算机应用与软件 | 2012年 / 29卷 / 12期

关键词：

深度估计; 聚焦测距; 相机阵列; 共聚焦图像;

D O I：

暂无

中图分类号：

TP391.41 [];

学科分类号：

080203 ;

摘要：

提出一种基于共聚焦图像的深度估计方法。采用虚拟孔径技术把相机阵列获得的多视角图像合成得到共聚焦图像序列,并将其作为聚焦测距的数据源进行聚焦分析,实现对场景深度信息的估计。针对共聚焦图像的特点,将传统聚焦测距方法中的清晰度评价算法与颜色一致性评价算法相结合,提出一种聚焦度测量算法,该测量算法同时适用于图像中强纹理区域和弱纹理区域的聚焦分析。实验结果表明,该方法对聚焦判别的有效性较高,可以获得较准确的场景深度估计。

引用

页码：43 / 46

页数：4

共 12 条

[1]

A novel depth-from-focus-based measurement system for the reconstruction of surface morphology with depth discontinuity[J] . Gyung Bum Kim,Gui Yun Tian. &nbspThe International Journal of Advanced Manufacturing Technology . 2009 (11)

[2]

Shape from focus using fastdiscrete curvelet transform. Minhas R,Mohammed A A,Wu Q M J. Pattern Recognition . 2011

[3]

A novel depth-from-hybrid-focus method for surface morphology with depth discontinuity[J] . Gyung-Bum Kim. &nbspThe International Journal of Advanced Manufacturing Technology . 2010 (1)

[4] 基于图像的几何建模技术综述 [J].

束搏 ;

邱显杰 ;

王兆其 .

计算机研究与发展, 2010, 47 (03) :549-560

[5]

Fusion of combined stereo and focus se-ries for depth estimation. Gheta I,Frese C,Heizmann M. GI . 2006

[6]

Passive 3D sensing,and reconstruction using multi-view imaging. Sadjadi F,Ribnick E. Computer Vision and Pattern Recognition Workshops,2010 . 2010

[7]

Real-Time Video-Based Modeling and Rendering of 3D Scenes. Takeshi Naemura,Jinji Tago,Hiroshi Harashima. Computer Graphics . 2002

[8]

Focus measurement onprogrammable graphics hardware for all in-focus rendering from lightfields. Sugita K,Takahashi K,Naemura T,et al. Proceeding of the 2004 virtual reality . 2004

[9]

Light fields and computational imaging. Levoy M. IEEE Computer . 2006

[10]

Synthetic aperture confocal imaging. Levoy M,Chen B,Vaish V,et al. ACM Transactions on Graphics . 2004

← 1 2 →