Artifical neural network based buried land mine imaging

Our buried land mine imaging system is based on transmitting and receiving antennas mounted within a cylindrical auger, a stepped frequency CW transmitter/receiver, and three artificial neural network (ANN) imaging processors, all controlled by a small computer. Stepped frequency, CW microwave signals reflected from objects within the view of the transmitter and receive antennas are used as inputs to the first ANN processor. This paper briefly describes the design of broad RF bandwidth cylindrical waveguide antennas that are mounted in the opposite spades of a dual spaded auger, the training patterns used for ANN land mine imaging and experimental land mine images using data recorded in April 1999 at the USA army´s A.P.Hill test site.