Character Recognition of Metal Magnetic Memory Signal Based on Wavelet Packet Transposition

The metal magnetic memory technology is the only tried and true method process early damage detection for metal parts. The paper put forward a new filtering method based on wavelet packet decomposition and reconstruction. Wavelet packet decompose signals on each different frequency range, we can achieve filtering goal through intercepting interested frequency range again reconstructing signals. Simulation verify the feasibility of the method and get preferable effect in actual signal processing that make character more clear in stress concentration fields and depress probability of misjudge.