Abstract
Detection of Landmines, especially anti-tank mines, bombs, and unexploded substances, is one of the major challenges facing humanity. The devastation and human tragedy associated with undetected explosives are self-evident in war-torn communities. To deal with this problem, we are only left with proactive measures that such substances must be detected and dealt with before the fallout. Most available solutions have major shortcomings, such as cost, efficiency, and accuracy, where the trade-offs among them are inversely related. On the other hand, advances in deep learning, unmanned aerial vehicle, and sensing are making their way as potential technologies to revolutionize the detection and removal of landmines. In this paper, we go through the literature reviewing the most recent work featuring computerized technologies to detect landmines. To our knowledge, no such study has taken place in this respect. Our aim is to find out how deep learning can be integrated with landmine detection. We identify open challenges toward viable automated solutions that enable deep learning to optimize performance effectively.