Multiple Magnetic Targets Localization Using Multiscale Cluster Matching Pursuit Method
Guangjing Du,Yao Wang,Lei Chen,Guoqiang Wang
DOI: https://doi.org/10.1109/tgrs.2024.3478757
IF: 8.2
2024-11-01
IEEE Transactions on Geoscience and Remote Sensing
Abstract:Multiple magnetic targets localization (MMTL) has been widely applied in geologic prospecting and vehicle surveillance due to its fast response speed and low cost; however, the MMTL of closely spaced targets is challenging because it is difficult to distinguish the signal of a target submerged in the magnetic anomalous fields of neighboring targets. In addition, current MMTL methods require large numbers of measuring points and precise numbers of targets, making these methods laborious and unfeasible. In this study, we propose a novel MMTL method based on a multiscale cluster matching pursuit (MCMP) algorithm that does not require prior information of the target number and is robust to noises. In this method, the MMTL task is first modeled as a sparse signal recovery problem that is solved with the proposed MCMP algorithm. Specifically, the target is iteratively searched at multiple scales, and the dispersed sources of the target are clustered during updating to avoid false target reconstruction. Using this method with only 4.7% of the measuring points required by previous methods, the localization error of closely spaced targets was reduced by 76% and 94% compared with the point cloud clustering method and data fitting method, respectively. The experimental application of the proposed method to the detection of unexploded ordnances (UXOs) resulted in localization errors of only 0.09 m. Interestingly, the targets could still be distinguished even if their magnetic moments were aligned or their separation distance was approximately equal to the targets' dimensions.
imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geochemistry & geophysics