Non-linear filtering for landmine detection on ground penetration radar data

摘要

A ground penetration radar is a geophysical method used in landmines detection method. The ground penetration radar sends a continuous wave in a frequency range and detects reflected signals from subsurface structures, even noise commonly called clutter. This paper proposes an automated landmine detection method based on ground penetration radar data segmentation using histogram thresholding and a non-linear filtering. Wavelet clutter reduction process is firstly performed to facilitate the segmentation step. Then, a decision algorithm is applied for the landmine detection task. The proposed method tests are performed on the MACADAM database.