Quantum Support Vector Machine for Landmine Detection based on UAV Terahertz imaging system

Abstract

Landmines pose a persistent threat to human safety and infrastructure in post-conflict regions. Conventional landmine detection methods often require direct human involvement, exposing personnel to significant risk while facing limitations in accuracy and scalability. In this study, we propose a Quantum Support Vector Machine (QSVM)-based detection system integrated with UAV-mounted terahertz imaging technology to enhance landmine identification efficiency. Leveraging quantum computing principles, QSVM provides improved classification accuracy by optimizing feature separability in high-dimensional spaces. Additionally, terahertz imaging enables remote sensing, reducing operational risks and enhancing detection precision. Experimental findings show that QSVM achieves better precision and recall than traditional classifiers while separating landmines from background noise.. Furthermore, comparative analysis with Quantum K-Means (QK-Means) and Quantum Neural Networks (QNN) underscores the robustness of QSVM in real-world applications. The integration of UAV platforms with quantum-enhanced algorithms offers a scalable, non-invasive, and high-accuracy solution for landmine detection, paving the way for safer and more efficient demining operations.