Landmine Detection with UAV-based Optical Data Fusion

Abstract

A general approach to landmine detection using drones equipped with multiple optical imaging sensors is described. The height of the drone's aerial survey is chosen so that enough pixels are captured within the site of landmine installation. This circumstance ensures the statistical representativeness of detection by each separate sensor. Next, the decision-level data fusion for different sensors is performed. Preliminary estimates of landmine detection probability by different statistical methods in multispectral and infrared imagery are presented.