Feature Extraction and Classification of Buried Landmine Signals

This paper presents a new technique that is capable of detecting and classifying buried landmines at different depths in sandy soil with different soil conditions. Prony's method is used to extract the complex natural resonance (CNRs) frequencies from the reflected landmine signals. These CNRs are used as new features for detection and classification of buried landmines. Different classification methods are employed to investigate the validity of the proposed method.