基于彩色图像局部结构特征的深度图超分辨率算法

被引：6

作者：

杨宇翔 ^{[1
]}

汪增福 ^{[1
,2
]}

机构：

[1] 中国科学技术大学自动化系

[2] 中国科学院合肥智能机械研究所

来源：

模式识别与人工智能 | 2013年 / 26卷 / 05期

关键词：

飞行时间相机; 拉普拉斯矩阵; 局部结构特征; 深度图超分辨率; 最速下降法;

D O I：

10.16451/j.cnki.issn1003-6059.2013.05.003

中图分类号：

TP391.41 [];

学科分类号：

080203 ;

摘要：

运用飞行时间相机来获取场景深度图像非常方便,但由于硬件的限制,得到的深度图像分辨率非常低,无法满足实际的需要.文中结合同场景的高分辨率彩色图像来制定优化框架,将深度图超分辨率问题转化为最优化问题来求解.具体来说,将彩色图像和深度图像在局部小窗口内具有的近似线性关系通过拉普拉斯矩阵的方式融合到目标函数的正则约束项中,运用彩色图像的局部结构参数模型,将该参数模型融入到正则约束项中对深度图的局部边缘结构提供更进一步的约束,再通过最速下降法有效地求解该优化问题.实验表明文中算法较其它算法无论在视觉效果还是客观评价指标下都可得到更好的结果.

引用

页码：454 / 459

页数：6

共 16 条

[1] Recent progress in coded structured light as a technique to solve the correspondence problem[J] . J. Batlle,E. Mouaddib,J. Salvi. &nbspPattern Recognition . 1998 (7)
[2] Range Image Super-Resolution via Guided Image Filter. Yang Yuxiang,Wang Zengfu. Proc of the4th International Conference on Internet Multimedia Computing and Service . 2012
[3] Spatial-Depth Super-Resolution for Range Images. Yang Qingxiong,Yang Ruigang,Davis J,et al. Proc of the IEEE Confer-ence on Computer Vision and Pattern Recognition . 2007
[4] A Revisit to MRF-Based Depth Map Super-Resolution and Enhancement. Lu Jiangbo,Min D,Pahwa R S,et al. Proc of the IEEE International Conference on Acoustics,Speech and Signal Process-ing . 2011
[5] An Application of Markov Random Fields to Range Sensing. Diebel J,Thrun S. Proc of Advances in Neural Information Processing Systems . 2005
[6] Super-Resolution Reconstruction Algorithm for PMD Range Image Based on Regulari-zation. Zhang Xudong,Shen Yuliang,Hu Liangmei. Proc of the International Conference on Intelligent Human-Machine Systems and Cybernetics . 2011
[7] High-quality Scanning Using Time-of-flight Depth Super-Resolution. Schuon S,Theobalt C,Davis J,et al. Proc of the IEEE Comput-er Society Conference on Computer Vision and Pattern Recognition . 2008
[8] Noise Brush:Interactive High Quality Image-Noise Separation. Chen Jia,Tang Chikeung,Wang Jue. ACM Transactions on Graphics . 2009
[9] Fast and Simple Super-Resolution for Range Data. Xiang Xueqin,Li Guangxia,Tong Jing,et al. Proc of the International Con-ference on Cyberworlds . 2010
[10] Closed-form solution to naturalimage matting. Levin A,Lischinski D,Weiss Y. IEEE Transactions on Pattern Analysis andMachine Intelligence . 2008

← 1 2 →