A preliminary analysis of a sparse reconstruction based classification method applied to GPR data

Abstract

This work investigates the performance of a sparse decomposition based approach applied to measured Ground Penetrating Radar (GPR) datasets for landmine recognition. The decomposition of the datasets is achieved via the solution of a constraint-relaxed convex optimization problem known as Basis Pursuit Denoise (BPDN). We demonstrate that it is crucial to appropriately construct a database of known scattering responses from mines and clutter, which will form the so-called dictionary. The robustness and accuracy of the methodology are evaluated against different parameters such as the size of the dictionary, the number and selection of time samples and the regularization parameter (noise estimate). Achieved performances are then assessed using the probability of detection and false alarm rate. As figure of merit for the classification accuracy, we use an additional measure, the so called Sparsity Concentration Index (SCI). For validation purposes, we finally compare the classification performance of the presented strategy with another sparse reconstruction based technique (Orthogonal Matching Pursuit, OMP) and an algorithm based on Support Vector Machines (SVM). The obtained results evidence that the proposed method is not only able to discriminate between targets and clutter, but also to recognize the particular type of mine simulants present in the evaluated surveys

On Choosing Training and Testing Data for Supervised Algorithms in Ground-Penetrating Radar Data for Buried Threat Detection

“…In this scenario, the classifier is applied at regular intervals along the A-scan and keypoints at the locations with the largest classifier decision statistics are utilized [6], [7], [11], [28]. The strategies in [8], [10], [20], [35], [36], [39], [45] set = , so that the same number of keypoints are utilized in training and testing.…”

Dictionary Learning for Adaptive GPR Landmine Classification

“…Sparse representation (SR) is effective in extracting the mid-or high-level features in image classification [14,15]. In the context of anti-personnel landmines recognition using GPR, our prior work [16,17] has shown that frameworks based on SR improve the performance of Support Vector Machine (SVM) classifier in distinguishing different types of mines and clutter in highly corrupted GPR signals. In this approach, the signal-of-interest is transformed into a domain where it can be expressed as a linear combination of only a few atoms chosen from a collection called the dictionary matrix [18,19].…”

“…Sparse representation (SR) is effective in extracting the mid-or high-level features in image classification [14,15]. In the context of anti-personnel landmines recognition using GPR, our prior work [16,17] has shown that frameworks based on SR improve the performance of Support Vector Machine (SVM) classifier in distinguishing different types of mines and clutter in highly corrupted GPR signals.…”

“…and normalize its columns 14 Replace the updated atoms D A i into D and normalize its columns 15 Perform SR of selected signals used in the previous step using entropy-thresholded batch OMP 16 Eliminate previous training set elements which have not been used for the last Nu iterations…”

Online dictionary learning aided target recognition in cognitive GPR

“…Therefore, our immediate goal is to find an efficient sparse representation (SR) which accurately represents the scattering behaviors related to soil type and targets. This has been shown to improve the classifier performance in discriminating mines from clutter [16,17].…”