Available at: https://digitalcommons.calpoly.edu/theses/1475
Date of Award
MS in Mechanical Engineering
Acts of insurgency have become an increasing threat resulting in extensive measures being taken by the law enforcement authorities to mitigate their devastating effects on human life and infrastructure. This thesis introduces a magnetometry-based information, and signal processing methodology for detecting concealed ferrous objects in vehicle body panels. From extensive literature research, it was observed that while magnetic sensors have been used in a variety of related applications, but they have not been extensively applied to the on-road detection of firearms and explosives concealed in vehicles. This study utilized an extensive experimental protocol for preliminary concept validation. The main idea behind the approach was that almost all concealed weapons and explosives are made up of a considerable amount of ferrous material, and hence produce a local distortion in the Earth’s magnetic field. This distortion can then be identified by utilizing sensitive magnetic sensors.
To detect concealed ferrous objects, magnetic signatures of a vehicle door panel were obtained by using a scanning assembly design in this thesis project, and compared to a base magnetic signature of the same vehicle door panel. The base magnetic signature is the magnetic field data of the same vehicle where no foreign ferrous objects were present. To analyze the data, a signal processing methodology was designed. To achieve the objective of accurately detecting concealed ferrous objects, simple measures such as magnetic field strength and its energy density were computed. These simple measures were then used in conjunction with more sophisticated statistical methods such as, normalized cross-correlation and Mahalanobis distance. Although all these methodologies were able to detect a magnetic footprint anomaly in the presence of a concealed object, the Mahalanobis distance approach, in particular provided the most conclusive results in all the test cases considered.