文献类型：期刊文献

中文题名：视觉传感网络压缩图像破损数据重构方法仿真

英文题名：Visual Sensing Network Compression Image Damage Data Reconstruction Method Simulation

作者：李恩林[1];刘洋[2]

第一作者：李恩林

机构：[1]河南财经政法大学计算机与信息工程学院;[2]河南财经政法大学云计算与大数据研究所

第一机构：河南财经政法大学计算机与信息工程学院

年份：2018

卷号：35

期号：1

起止页码：184-187

中文期刊名：计算机仿真

外文期刊名：Computer Simulation

收录：CSTPCD;;北大核心:【北大核心2017】；

语种：中文

中文关键词：视觉传感网络;压缩图像;破损数据重构

外文关键词：Visual sensor network ; Compressed image ; Reconstruction of damaged data

摘要：为了提升视觉传感网络服务质量,需要进行视觉传感网络压缩图像破损数据重构方法的研究。但是采用当前方法进行破损数据重构时,无法确定破损数据区域边界重构的最大优先权点,存在图像破损数据重构误差大的问题。为此,提出一种基于非降采样轮廓波变换的视觉传感网络压缩图像破损数据重构方法。该方法先采用局部离散余弦得到压缩图像破损数据的结构部分和纹理部分,利用贝叶斯压缩感知得到破损数据的分布函数,获取分布函数的均值和方差,利用非降采样轮廓波变换确定破损数据区域边界重构的最大优先权点,设计非降采样方向滤波器,给出图像破损数据低频分量和高频分量信息的相似性和一致性,由此实现视觉传感网络压缩图像破损数据重构。仿真证明所提方法重构精度较高,有效地提升了视觉传感网络的服务质量。
ABSTRACT：A reconstructing method of damaged data of compressed image in visual sensor network is proposed based on the non-reduced sampling contourlet transform. Firstly, the local discrete cosine is used to get structure and texture of damaged data of compressed image and Bayesian compressive sensing is adopted to get distribution function of damaged data and the mean value and variance of distribution function. Secondly, the non-reduced sampling cont- ourlet transform is used to determine the maximum priority point for reconstructing the regional boundary of damaged data. Then, the non-reduced sampling directional filter is designed, and the similarity and consistency of low fre- quency and high frequency components of damaged data are given, thus the reconstruction of damaged data of com- pressed images in vision sensing network is realized. Simulation results prove that the proposed method can improve the service quality of visual sensor network effectively, which has high precision in reconstruction.