Abstract:For pipelines, tanks and other equipment defects, magnetic flux leakage inspection is an important nondestructive testing method, but the huge amount of signal data will be generated in the testing process, and it will produce a variety of interference and noise, occurring problems such as signal distortion, drift, and annihilation etc. Signal noise reduction filter is key for magnetic flux leakage defect detection. The signal noise reduction compression method was explored. The compressed MFL signal denoising wavelet packet-Haar wavelet algorithm was studied. By this algorithm, signal data were compressed to reduce noise while the high frequency signal characteristics were retained, resulting in avoiding the phenomenon of signal distortion aberration in data decompressing process.