基于树状稀疏模型的视觉传感器网络图像数据重构

Image Reconstruction Algorithm Based on Tree Sparsity Model for Visual Sensor Network

针对现有压缩感知算法无法有效利用视觉传感器网络中图像数据相关性的问题,提出一种基于树状稀疏模型的视觉传感器网络数据压缩感知算法。在分析图像数据小波域稀疏特性的基础上,构建了一种视觉传感器网络图像数据的树状稀疏模型,进而针对此模型设计一种新的压缩感知重构算法。理论分析和实验结果表明,相比于传统图像数据压缩感知算法,该算法可有效利用图像数据相关性减少准确重构图像数据所需的测量值,降低视觉传感器网络数据传输能耗。

A tree sparsity model based image compressed sensing （CS） algorithm was proposed to efficiently explore the correlation in visual sensor networks （~SN） data. Based on the analysis of wavelet sparsity, a tree sparsity model for VSN image was established and a new CS recovery algorithm for this model was proposed. Analysis and experimental results demonstrate that the proposed algorithm can significantly reduce the measurement for the accurate recovery, and subsequently lower energy consumption of data traffic in VSN.