Improving the Performance of the PI Systems through the Use of Neural Network

Abstract

Pulse induction systems include the very cheep and often-used metal detectors employed, besides other applications, in geology and mine-detecting robots. these systems only detect the presence of metals but are not capable of distinguishing them from each other in this article a method is introduced in which multiple layer perception networks are used to distinguish metals from each other on the basis of their electromagnetic characteristics . This method has been tested on 200 different samples containing the three metals iron, copper and aluminum. The results obtained show the degree of accuracy of the proposed method is over 95 percent.