Application of Freeman decomposition to full polarimetric GPR for improving subsurface target classification

作者：Feng, Xuan^*; Liang, Wenjing; Liu, Cai; Nilot, Enhedelihai; Zhang, Minghe; Liang, Shuaishuai

来源：Signal Processing, 2017, 132: 284-292.

DOI：10.1016/j.sigpro.2016.07.030

• Ground penetrating radar (GPR)
• Freeman decomposition
• Subsurface target
• Geometrical feature
• Polarimetric information
• Classification

Migration technique can reconstruct the subsurface target imaging from record ground penetrating radar (GPR) data. Depending on the reconstructed imaging, we possibly distinguish subsurface target by the geometrical feature. But some different targets have similar reconstructed imaging, which will confuse us. Freeman decomposition is a technique for fitting a three-component scattering mechanism model to polarimetric synthetic aperture radar observations, which is successfully used in the classification of terrain objects. This paper applies the model-based Freeman decomposition to full polarimetric GPR data to improve the subsurface target classification. We use a full polarimetric GPR, which is constructed in laboratory with three types of antenna combinations, HH, VV and VH combinations, to acquire experiment data sets for testing the method. Metallic plate, ball, dihedral and volume scatter with many branches, are buried in the homogenous dry sand under flat ground surface, and three dimensional data sets are acquired above each buried target. After signal processing, we obtained subsurface target imaging by migration and color-coded polarimetric information by Freeman decomposition. Results showed that it can improve the classification capability of GPR for the subsurface target to use both the geometrical feature and polarimetric information.

• 出版日期2017-3
• 单位MIT; 吉林大学