The basic idea of Matching Pursuit algorithm is based on the possibility of decomposition and reconstruction of seismic data.It is a process in which a series of functions can be chosen from a certain function set to represent a given signal,and the each and every function within the function set is called as an atom.This paper uses singular value decomposition method to revise matching pursuit algorithm,and improves the speed of convergence and computation,and then compares the time-frequency distribution ...