For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Magnetic Anomaly Detection with Empirical Mode Decomposition Trend Filtering
Han ZHOU Zhongming PAN Zhuohang ZHANG
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences
Publication Date: 2017/11/01
Online ISSN: 1745-1337
Type of Manuscript: LETTER
Category: Digital Signal Processing
MAD, magnetometer, EMD, trend filtering,
Full Text: PDF(650.7KB)
>>Buy this Article
Magnetic Anomaly Detection (MAD) is a passive method for the detection of ferromagnetic objects. Currently, the performance of a MAD system is limited by the magnetic background noise that is non-stationary and shows self-similarity and long-range correlation. In this paper, we propose an empirical mode decomposition (EMD) trend filtering based energy detector for adaptively detecting the magnetic anomaly signal from the background noise. The input data is first detrended adaptively with the energy-ratio trend filtering approach. Then, the magnetic anomaly signal is detected using an energy detector. The proposed detector does not need any a priori knowledge about the target or assumptions regarding the background noise. Experiments also prove that the proposed detector shows a more stable performance than the existing undecimated discrete wavelet transform (UDWT) based energy detector.