Landmine contamination remains one of the most pressing humanitarian challenges worldwide. Ukraine alone has approximately 23% of its territory requiring survey, with estimated demining timelines of 5 to 10 years [1, 2], creating urgent demand for faster, safer, and more automated detection solutions. Ground-penetrating radar (GPR) is among the most promising technologies for buried mine detection, particularly for low-metal content mines invisible to electromagnetic induction sensors. As highlighted by Sato [3], the combination of GPR with synthetic aperture radar (SAR) processing and robotic platforms represents a key direction for automated demining. However, conventional GPR systems produce raw subsurface images that require extensive signal processing before targets can be identified, typically including filtering steps to suppress surface reflections and soil clutter, followed by hyperbola detection in the B-scan to localise buried objects, making automatic target recognition (ATR) on mobile platforms particularly challenging.
Lambot, S., Miguel Freixo Gonçalves, Piotr Bielski, Émile le Flécher, & Geert De Cubber. (2026). Model-Based GPR Landmine Detection Using Full-Wave Inversion: Experimental Results from the CONVOY Project. 22nd International Symposium Mine Action 2026, Split, Croatia. https://hdl.handle.net/2078.5/274052