3D GPR Real-Time Automated Detection of Buried Utilities

mini-trench" technique (addressed in ITU L.83 recommendation) requires that non-destructive mapping of buried services enhances its productivity to match the improvements of new digging equipment. Therefore, the development of a fully automated and real-time 3D GPR processing system plays a key-role in overall optical network deployment profitability. We propose a novel processing scheme whose goal is the automated processing and detection of buried targets, that can be applied in real-time to 3D GPR array system (16 antennas, 900 MHz central frequency). After the standard pre-processing steps, the antenna records are continuously focused during acquisition, by the mean of Kirchhoff depth-migration algorithm, to build pre-stack reflection angle gathers G(x, θ; v) at nv different velocities. The analysis of pre-stack reflection angle gathers plays a key-role in automated detection: by the mean of correlation estimate computed for all the nv reflection angle gathers, targets are identified and the best local propagation velocities are recovered. The data redundancy of 3D GPR acquisitions highly improves the proposed automatic detection reliability. The proposed approach allows to process 3D GPR data and automatically detect buried utilities in real-time on a laptop computer, without the need of skilled interpreters and without specific high performance hardware. More than 100 Km of acquired data prove the feasibility of the proposed approach.