The invention provides a method for increasing
wavelet packet
decomposition speed based on Mallat
algorithm, relates to the technical field of
signal treatment, and solves the problems that in the prior art, in the process of extracting characteristics,
wavelet decomposition is adopted for only decomposing
low frequency signals, and information about the
low frequency signals can lose. The methodcomprises four steps of selecting
wavelet decomposition types and decomposition layer numbers; calculating
signal frequency; through a backcasting reckoning method, obtaining all decomposed signals; and reconstructing wavelets so as to obtain useful target frequency signals. The Daubechies wavelets
algorithm is adopted, the wavelets and the decomposition layer number thereof are selected, a Shannon sampling theory is utilized, so that the frequency range of each fragment after
wavelet packet decomposition can be further obtained. An
algorithm for combining semi-wavelet packets and an improvement
convolution process is adopted, so that the defect that
wavelet decomposition can only realize decomposition of the
low frequency signals, is overcome, and redundant decomposition of wavelet packets is also solved; and therefore, the
signal decomposition precision is improved, the complexity of the algorithm is reduced, and the effect of accelerating
wavelet decomposition is achieved.