Novel algorithm for landmine detection using C-scan ground penetrating radar signals

摘要

This paper introduces a novel algorithm that can achieve close-range landmine detection using Ground Penetrating Radar (GPR) data. The proposed GPR feature-based algorithm combines spectral and spatial edge features of the GPR signals to form a new oscillation feature, which can represent the reflection value change between the layers of a C-scan GPR image. The GPR feature-based algorithm detects landmines through comparing the oscillation feature of each A-scan GPR signal in the 3D GPR image to generate a 2D landmine detection result image. Performance evaluation results of the proposed algorithm in comparison with existing methods are reported in terms of detection sensitivity, false alarm rate and computational complexity. The test results show that the GPR feature-based algorithm achieves a high detection performance and can be used in real-time landmine detecting systems.