Detection of buried mines from GPR array measurement: A statistical approach

Array ground penetrating radar (GPR) has been widely used in landmine and unexploded ordnance remediation. Usually, the exact form of the target signal is not known. We develop a low-complexity, statistical strategy to solve the problem of object detection under these circumstances. At each stop of the array, a high-dimensional analysis of variance (HANOVA) is designed to look for significant difference among the array signals and suppress accumulated noise. By looking at differences among received signals, we circumvent the difficulty of the unknown signal. Results from applying the method on real data are presented