Shallow-Water Mine Detection Based on Nonlinear Signal Enhancement by Computational Fluid Dynamics Techniques

Detection is the most important part of a countermine operation. A high- resolution mine-detection technique will allow individual mines to be targeted, with a consequent enormous saving of resources. Most imaging techniques produce blurred images with only a few pixels per mine, in which the mines are nearly indistinguishable from their surroundings. The nonlinear edge-sharpening method proposed here, based on the second-order Godunov technique employed in computational fluid dynamics, will significantly improve mine-detection capabilities. The method effectively mitigates instrumental errors, channel noise, and the low-amplitude high-frequency noise component introduced by the imaging technique. It is a hands-off technique that produces qualitative improvement even without image-specific pre- or postprocessing. This is especially valuable for rapid surveillance methods. We propose to develop an optimized PC-based menu-driven package which will permit images to be enhanced in real time. The same techniques can also be exploited in other fields of commercial importance, such as medical science: essentially every CAT scan and MRI facility is a potential customer.