This work proposes a Machine Learning (ML) approach for the analysis and classification of Ground Penetrating Radar (GPR) given a limited number of B-scan images. Specifically, we consider both a custom Convolutional Neural Network (CNN) and a wellestablished Deep Learning (DL) architecture, DenseNet, that is opportunely scaled down to take into account the small dataset. Those networks are then employed to classify B-scan simulations from buried cylinders in order to retrieve the host media permittivity, the cylinder depth respect to surface, and cylinders radius. A prediction based on the mean-square error (MSE) is applied. The main aim of the proposed work is to test the applicability of a scaled-down version of DenseNet architecture to the analysis of B-scan images and compare the performance respect to a classical CNN. The architecture chosen has shown interesting results in retrieving information from a limited data set. Limitations of the considered approach are also discussed.

这项工作提出了一种机器学习 (ML) 方法，用于在有限数量的 B 扫描图像的情况下对探地雷达 (GPR) 进行分析和分类。具体来说，我们同时考虑了自定义卷积神经网络 (CNN) 和完善的深度学习 (DL) 架构 DenseNet，该架构适时缩小以考虑小数据集。然后使用这些网络对埋入圆柱体的 B 扫描模拟进行分类，以检索宿主介质介电常数、圆柱体相对于表面的深度和圆柱体半径。应用基于均方误差 (MSE) 的预测。拟议工作的主要目的是测试 DenseNet 架构的缩小版本对 B 扫描图像分析的适用性，并比较经典 CNN 的性能。所选择的架构在从有限的数据集中检索信息方面显示出有趣的结果。还讨论了所考虑方法的局限性。