A large number of algorithms have been proposed for automatic buried threat detection (BTD) in ground penetrating radar (GPR) data. Convolutional neural networks (CNNs) have recently achieved groundbreaking results on many recognition tasks. This success is due, in part, to their ability to automatically infer effective data representations (i.e., features) using training data. This capability however results in a high capacity model (i.e., many free parameters) that is difficult to train, and more prone to overfitting, than models employing hand-crafted feature designs. This drawback is pronounced when training data is relatively scarce, as is the case with GPR BTD. In this work we propose to combine the relative advantages of hand-crafted features, and CNNs, by constructing CNN architectures that closely emulate successful hand-crafted feature designs for GPR BTD. This makes it possible to apply supervised training to traditional hand-crafted features, allowing them to adapt to the unique characteristics of the GPR BTD problem. Simultaneously, this approach yields a much lower capacity CNN model that incorporates substantial prior research knowledge, making the model much easier to train. We demonstrate the feasibility and effectiveness of this approach by designing a “neural” implementation of the popular histogram of oriented gradient (HOG) feature. The resulting neural HOG (NHOG) implementation is much smaller and easier to train than standard CNN architectures, and achieves superior detection performance compared to the un-trained HOG feature. In theory, neural implementations can be developed for many existing successful GPR BTD algorithms, potentially yielding similar benefits.
The Ground Penetrating Radar (GPR) is a remote sensing modality that has been used to collect data for the task of buried threat detection. The returns of the GPR can be organized as images in which the characteristic visual patterns of threats can be leveraged for detection using visual descriptors. Recently, convolutional neural networks (CNNs) have been applied to this problem, inspired by their state-of-the-art-performance on object recognition tasks in natural images. One well known limitation of CNNs is that they require large amounts of data for training (i.e., parameter inference) to avoid overfitting (i.e., poor generalization). This presents a major challenge for target detection in GPR because of the (relatively) few labeled examples of targets and non-target GPR data. In this work we use a popular transfer learning approach for CNNs to address this problem. In this approach we train two CNN on other, much larger, datasets of grayscale imagery for different problems. Specifically, we pre-train our CNNs on (i) the popular Cifar10 dataset, and (ii) a dataset of high resolution aerial imagery for detecting solar photovoltaic arrays. We then use varying subsets of the parameters from these two pre-trained CNNs to initialize the training of our buried threat detection networks for GPR data. We conduct experiments on a large collection of GPR data and demonstrate that these approaches improve the performance of CNNs for buried target detection in GPR data
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.