Signal Processing Techniques for Improved Performance of a SQUID-Based Metal-Detector

A SQUID-based prototype sensor system for the detection of stainless steel contaminants in food has been previously reported . In this work, we describe the use of adaptive filters, matched filter correlation and component estimation algorithms for automatic contaminant detection for this system. The specification of the probability of false positives-using statistical analysis-allows the resulting contaminant detection limit to be determined. These improvements will enable commercial operation of SQUID-based stainless-steel detection systems in standard factory environments with acceptable correct and false positive detection rates.