Search

Greedy algorithms represented by orthogonal matching pursuit (OMP) and subspace pursuit (SP) algorithms are practically used in image processing based upon compressed sensing theory. However, there are two disadvantages: 1)Relatively poor signal reconstruction accuracy; 2) High computation complexity and measurements time. This paper proposes a frame of greedy algorithms obtaining a novel fusion of matching pursuit (FMP), combining the OMP and SP algorithms. FMP unites the two support sets from OMP and SP selecting the most appropriate atoms to achieve secondary screening of the original two support sets, finally realizing the accurate signal reconstruction. Using same test conditions, image reconstruction experiments and stability of Frame, the proposed FMP algorithm can effectively improve signal-to-noise ratio (SNR) with improved reconstruction error. Reconstruction effects using proposed FMP are better than separately using other two greedy algorithms for both high and low resolution images.

In the present, Xintan Kendong 451-the 4th Kendong union station gathering is heavy oil mixed with hot water process and open operation mode. In the pipeline, without the variation of water content, in closed gathering means, mixed with hot water; mixing activity water and emulsifying gathering process are tested. The results show that the heavy oil gathering process of selection and gathering distance are related, for little variation of the water content and the flow, at about 50 ℃. The distance is shorter in the gathering, the use of hot water mixed gathering process more reasonable, when gathering distance is longer, selection of emulsion pipeline gathering process more reasonable.