作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 多媒体技术 • 上一篇    下一篇

基于三维树状结构和贝叶斯模型的视频压缩感知重构算法

庄燕滨1,2,王尊志1,肖贤建2,张学武3   

  1. (1.河海大学计算机与信息学院,南京 211100; 2.常州工学院计算机信息工程学院,江苏 常州 213002; 3.河海大学物联网工程学院,江苏 常州 213022)
  • 收稿日期:2015-01-27 出版日期:2016-02-15 发布日期:2016-01-29
  • 作者简介:庄燕滨(1964-),男,教授,主研方向为智能信息处理、视频图像处理;王尊志,硕士研究生;肖贤建、张学武,副教授。
  • 基金资助:
    国家自然科学基金资助项目(61273170);江苏省科技厅工业支撑计划基金资助项目(BE2010072)。

Video Compressive Sensing Reconstruction Algorithm Based on 3D Tree Structure and Bayesian Model

ZHUANG Yanbin  1,2,WANG Zunzhi  1,XIAO Xianjian  2,ZHANG Xuewu  3   

  1. (1.College of Computer and Information,Hohai University,Nanjing 211100,China;2.College of Computer and Information Engineering,Changzhou Institute of Technology,Changzhou,Jiangsu 213002,China;3.College of Internet of Things Engineering,Hohai University,Changzhou,Jiangsu 213022,China)
  • Received:2015-01-27 Online:2016-02-15 Published:2016-01-29

摘要:

针对传统水下视频编码对水声信道带宽要求较高,并且水下视频具有场景复杂、不固定等特点,基于三维树状结构和贝叶斯模型,提出一种水下视频压缩感知重构算法。在编码端,模拟彩色编码孔径压缩时间成像系统对视频信号进行编码。在贝叶斯压缩感知模型的基础上,解码端利用小波和离散余弦变换系数的三维树状结构得到贝叶斯压缩感知反变换算法,从单通道的压缩测量值中重构彩色视频帧。实验结果表明,该算法能够精确重构复杂的视频场景。

关键词: 压缩感知, 水下视频, 三维树状结构, 小波变换, 贝叶斯模型

Abstract: In view of the traditional underwater video coding requiring higher underwater acoustic channel and the scenes of underwater video with uneven illumination being complex and not fixed,this paper presents a reconstruction algorithm with Three Dimension(3D) tree structure Bayesian compressive sensing for underwater video.At the encoder side,analog color coded aperture compressive temporal imaging system is used to code the video signal.At the decoding end,based on the model of Bayesian compressive sensing,by exploiting the 3D tree structure of the wavelet and Discrete Cosine Transformation(DCT) coefficients,a Bayesian compressive sensing inversion algorithm is derived to reconstruct color video frames from a single monochromatic compressive measurement.Experimental results show that the algorithm is able to reconstruct the complex video scenes accurately.

Key words: compressive sensing, underwater video, 3D tree structure, wavelet transformation, Bayesian model

中图分类号: