Fast Analysis of C-Scans From Ground Penetrating Radar via 3-D Haar-Like Features With Application to Landmine Detection

Abstract

This paper aimed to devise an efficient algorithm applicable to ground penetrating radar (GPR) and to enable an automatic landmine detection. Proposed is a machine learning approach in which we put the main emphasis on fast performance of the scanning procedure analyzing the C-scans, i.e., 3-D images defined over the coordinate system, i.e., along track by across track by time, where the time axis can be associated with depth. The approach is based on our proposition of 3-D Haar-like features. Learning of the detector is carried out by boosted decision trees.Practical experiments on metal and plastic antitank mines in a garden soil are carried out. A prototype mobile platform is designed to scan the subsurface of the ground, equipped with a GPR based on a standard vector network analyzer and our original antenna system. We report the results, particularly the following: detection sensitivity, false alarm rates, receiver operating characteristic curves, and times of learning and detection.

Constant-Time Fourier Moments for Face Detection — Can Accuracy of Haar-Like Features Be Beaten?

“…The proposed approach could be beneficial in other machine learning applications where accuracy is of primary importance (e.g. medical diagnosis, image-based fault dection in production, landmine detection [4,5]), and where one is willing to invest some additional time in the preparation of special integral images in order to improve accuracy.…”

“…The intention of Viola and Jones was to generate a massive multitude of features (e.g. ∼10 5 ), so that some of them might happen to represent good characteristics of target objects (e.g. for faces: differences between forehead and eyes, nose and cheeks, etc.).…”

“…In turn, cumulants of so-called vote matrices allow for extraction of HOG 2 features [5,7]. Yet, other propositions of that kind are scarce and, in generality, approaches which would allow for constant-time extraction of more advanced features, exhibitting better approximation properties, are not known.…”

Statistical moments calculated via integral images in application to landmine detection from Ground Penetrating Radar 3D scans

“…Reported were: % 95% sensitivity, % 0:0048FA=m 2 FAR, indicating the high effectiveness of the method. We first describe briefly after [12,22] how HOG features are extracted; then, we report the results obtained on our data. We programmed the benchmark in C# and integrated it with our software.…”

“…As regards the latter, quite many state-of-the-art methods have been tried out, e.g., Naive Bayes and LVQ in [6], neural networks in [10], least squares curve fitting in [9,26], HMMs in [14,17,26] or ensemble classifiers in [12,22]. Yet, it seems, in general, that the final success is less dependent on the choice of learning algorithm and more dependent on the quality of images and features extracted from them.…”