Ground Penetrating Radar B-Scan Data Modeling and Clutter Suppression

Abstract

Clutter suppression is a vital problem for Ground Penetrating Radar (GPR) signal process. The clutter suppression algorithm of GPR based on Principal Component Analysis (PCA) is concerned with the good adaptability of the ground irregularity, but there is still a problem of selecting the main components which exist in the algorithm. To settle this problem, we define the principle of matrix inner product at first, and analysis the approximate mutual orthogonality of the ground surface clutter matrix, the target signal matrix and the background noise matrix. And then, we present a method of modeling the GPR B-scan data. Combining the data model with GPR image entropy, we propose a method of estimating the principal component of the clutter suppression algorithm. Finally, we use the simulation data generated by the GPRMax software to verify the validity of the algorithm.