Object Identification for GPR images with Convolutional Neural Network and Generative Adversarial Network

Abstract

In this study, to automatically detect underground objects from the ground penetrating radar (GPR) images by the deep neural network (DNN), we have generated GPR images for training the DNN using a fast finite-difference time-domain (FDTD) simulation with graphics processing units (GPUs). Also, we have studied about identification characteristics of the underground objects with the generated GPR images by a convolutional neural network (CNN) and a fine-tuning which is modified VGG16 trained by the ImageNet. In this work, we have investigated to identify experimental GPR images by the generative adversarial network (GAN) which transfers from simulated images to artificial images.