Lossless Compression of Pipeline Magnetic Flux Leakage Inspection Data Based on Integer Lifting Wavelet Transform

摘要

Pipeline Magnetic Flux Leakage (MFL) inspection device with high resolution generates a great deal of data. An effective technique of online data compression is urgently needed. Loss data compression will ignore both noise and useful weak MFL signal, so lossless data compression is necessary. First-order differencing is used to eliminate signal variance generated by the variance of sensors lift-off. Differenced MFL image is integer lifting wavelet transformed with integer CDF(2,2) biorthogonal wavelet. SPIHT (Set Partitioning in Hierarchical Tree) algorithm is used to code the transformed data, and then lossless data compression is implemented. The compression ratio of inspection data for the pipeline with defect is above 3:1, and for the pipeline without defect the compression ratio is above 6:1. The algorithm has a good real-time characteristic and is easy to hardware implementation.