Magnetic Anomaly Detection Using a Parallel Stochastic Resonance System

Abstract

Magnetic anomaly detection (MAD) is a useful method for the detection of visually obscured ferromagnetic targets. However, when the prior information of the target signal is lacking or inaccurate, the performance of the traditional method based on orthonormal basis functions may degrade rapidly. Therefore, a novel detection method based on a parallel stochastic resonance (PSR) system is proposed, which is named the PSR detector. This method uses a PSR system to improve the signal-to-noise ratio (SNR) and then reveals anomaly signals by energy detector. Experimental results show that the detection probability of PSR detector reaches 70% even if input SNR is −3 dB and approximated to 100% if SNR is 0 dB under the false alarm probability (FAR) of 5%. In particular, the PSR detector provides a higher detection probability than the minimum entropy (ME) detector and a more balanced performance than the detector using a single stochastic resonance (SSR) system when the deterministic parameter of the target signal is lacking.