Fusion of correlated decisions for GPR based buried landmine detection

The present paper demonstrates the fusion of correlated binary local decisions as applied to Ground Penetration Radar (GPR) based landmine detection. In this procedure, impulse GPR is used as a sensor and various signal processing techniques are used as detectors and decision makers. Decisions made by the signal processors are sent to the fusion center so that a global decision is to be made. In addition, the Bahadur-Lazarfeld Expansion (BLE) is employed to estimate the correlation among individual decisions. An excellent identification performance is achieved with the proposed technique and results of this technique are compared with the AND, OR and Majority voting techniques.