Abstract:
This thesis presents a method of underwater surveillance of ferromagnetic objects based on Magnetic Anomaly Detection. A numerical method of calculating electromagnetic fields produced by the elementary dipole sources in a stratified media is built. The method is useful in estimating magnetic signature at the air generated by the ferromagnetic objects submerged in the ocean. Mathematical modifications to match the same method for a means of measuring magnetic noise signals where noise sources are modeled as multiple equivalent electric and magnetic dipoles distributed at the environment are developed. After developing the simulation environment which is aimed at mimicking the real ocean environment, Empirical Mode decomposition technique is used to anlyse the noisy magnetic signals recorded by Tri-axial magnetometer from Unmanned Aerial Vehicle. It decomposes the obtained signals into Intrinsic Mode Functions and Residue, and responds to any anomaly found in the denoised signals. A Triangular geometry of sensor positions is proposed with combined Particle Filter and Unscented Kalman Filter as signal processing algorithms for the classification, localization and tracking of the target in this study.
