Multi-hypothesis distributed video compression sensing based on key frame secondary reconstruction

Abstract

Reconstruction algorithms are the key technology of distributed video compressed sensing. The research focus of traditional distributed video compressed sensing reconstruction algorithms is mostly on improving the reconstruction quality of non-key frames, ignoring the reconstruction quality of key frames, and the information of key frames are not Underutilized. In view of the above problems, a distributed video compression sensing algorithm based on secondary reconstruction of key frames is proposed. Firstly, the fractional order total variation algorithm is used for the initial reconstruction of the key frame, and the reconstructed frame is used as the reference frame to assist the secondary reconstruction of the key frame, which improves the reconstruction quality and reduces the calculation complexity. Then, a multi-reference frame bidirectional prediction hypothesis set optimization algorithm is proposed to increase the number of reference frames and improve the quality of the hypothesis set through optimization without expanding the size of the hypothesis set. Experimental results show that the overall performance of the proposed algorithm is better than the most advanced methods.