Research On Reconstruction Algorithm Of Distributed Compressive Video Sensing

Compressed sensing(CS) which integrates the sampling and the compressing stage of traditional sampling is a new sampling technique, breaking the limits of Nyquist sampling theorem. Distributed compressive video sensing(DCVS) brings CS into video encoding. By using CS for sampling, DCVS reduces the data amount at the encoder thus decreasing the energy consumption. Reconstruct the video signal exactly at the decoder utilizing the reconstruction algorithm. In this paper, I aim at the optimization of the reconstruction algorithm at the decoder.Existing distributed compressive video sensing uses more motion estimation and motion compensation to produce side information in the pixel domain. The prediction quality of pixel motion estimation and motion compensation has a strong dependence reference image for quality. However, after the use of compressed sensing technology for encoding the resulting perception is measured values, can not get the image information of the pixel region, so if we can direct to be reconstructed frame prediction CS domain measurements, will greatly simplify the forecasting process, improve forecast also is the edge quality of the information. Therefore, the proposed reconstruction algorithm to optimize compressed video distributed sensing systems, the use of multi-hypothesis generating side information technology, and the side information as a TV reconstruction iteration initial value, compared to the previous use of motion estimation, motion compensation technology to produce side information, since the multi-hypothesis generating side information of higher quality, thereby improving the quality of TV reconstructed. Simulation results show that the proposed combination of multiple hypothesis TV technology to optimize reconstruction algorithm for distributed compressive video sensing to improve the quality of the reconstruction.