针对252Cf源驱动核材料产生裂变中子脉冲信号具有脉冲序列特殊的“0,1”稀疏结构之特点,采用压缩感知理论,通过巧妙引入图论中的二分图模型,同时结合二分图的最小覆盖性质,适当添加约束条件,构建了稀疏均匀的观测矩阵。研究结果表明,利用压缩感知理论对“0,1”中子脉冲序列特殊稀疏结构的信号重构算法不仅可行,而且还获得了优于l1范数最小化方法重构结果,这对252Cf驱动核材料的中子脉冲信号分析与处理提供了一种新的途径或方法。

The 252Cf-Source-Driven neutron pulse series have special sparse structure of “0, 1”. There is no doubt that it satisfies the general condition of compressive sensing(CS). Unlike conventional compressive sensing approaches, we model the compressive sensing process with bipartite graph, combining with minimal covering of bipartite graph and adding some suitable conditions to the bipartite graph to construct a sparse and uniform measurement matrix. The results show that this new CS reconstruction algorithm is feasible for the 252Cf-source-driven neutron pulse series, and has a better performance than l1method. Furthermore, this research provides a new approach or method for the signal processing of 252Cf-source-driven neutron pulse series.