V. Kovalenko

A Novel Clutter Suppression Algorithm for Landmine Detection With GPR

In this paper, we propose a new algorithm for the enhancement of plastic-cased antipersonnel mine detection using a video-impulse ground-penetrating radar (GPR). The algorithm is implemented as a nonlinear signal processor, which searches for the presence of a reference waveform in a 1D GPR echo return. The reference waveform represents a class of targets within a certain environment. The processor marks the presence of all responses similar to the reference waveform with a sharp mono-cycle. Simultaneously, responses with different waveforms, which presumably correspond to clutter, are suppressed. The reference waveform and other algorithm parameters are determined from training data sets acquired in a controlled environment. After training, the algorithm can be successfully applied at sites where soil, targets, and measurement scenarios are similar but not identical to those of the training site. The processor is integrated into an automated data processing and mine detection scheme as an additional clutter suppression step. The scheme consists of clutter suppression, synthetic aperture radar focusing, construction of a confidence map, and automated detection in it. The suggested algorithm is tested on experimental data, and its performance is compared against schemes where clutter suppression is organized by means of background removal and the cross correlation with a reference wavelet. The performance comparison is done in terms of receiver operating characteristic curves. It has been found that the suggested algorithm reduces the false alarm rate in about two and a half times in comparison to the cross-correlation-based clutter suppression.

Application of UWB Near-Field Polarimetry to Classification of GPR Targets

Polarimetric properties of the recently developed IRCTR video-impulse radar are analyzed. These properties are used for the Polarimetrie analysis of data acquired over a sand lane containing landmine simulants and clutter. From the measured scattering matrices, the estimated linearity factor and the angle of orientation of different targets are determined as a function of frequency. Focused multi-component images of different targets are formed and analyzed. It is shown that the eigenvalues of the target scattering matrix and multi-component images can be used for the classification of detected targets.

Alternating-sign windowed energy projection of SAR focused GPR data

The automated detection of plastic-cased antipersonnel landmines in 3-D high-resolution UWB radar images is considered. The 3-D images are projected in different ways onto 2-D confidence maps, in which the automated APM detection is made by marking local maxima in the maps followed by thresholding. Specifically, we consider the confidence maps of the following types: energy projection (EP), windowed energy projection (WEP) and introduced here alternating sign windowed energy projection (ASWEP). The latter is shown to allow diminishing of the false alarms number in comparison for the other two techniques for most of the considered cases. The advantages of the new technique are described by means of receiver operator curves (ROC). The ROC curves are built for the datasets acquired over several minefield simulation sites under different environmental conditions

Waveform based detection of anti-personnel mines with an UWB radar

The present paper deals with feature-based landmine detection by means of UWB ground penetrating radar (GPR). A waveform of target response is used as a discriminative feature for two feature-based detection schemes. The output of each DPS is a confidence map obtained by means of energy projection of a 3D distributed feature onto a 2D surface. An automated detection is made over the obtained confidence maps and mines detectability is analyzed on the basis of the receiver operation characteristic curves. Performances of the feature- and energy-based data processing schemes (DPS) applied to the experimental data are compared. It is shown that suggested waveform based detection is stable to the change of the environment.

Polarimetric feature fusion in GPR for landmine detection

A polarimetric multi-feature framework for the detection of antipersonnel landmines with ground penetrating radar (GPR) is suggested. The features result from independently acquired and processed GPR measurements in co- and cross-polar configurations. The initial detection in the confidence maps is made independently after which the coordinates of the detected targets are co-located. The marginal feature distributions are normalized via Johnsons transform prior to the fusion process and a Maximum Likelihood based linear-quadratic classifier is used as a fusion rule. The framework makes use of secondary data acquired from an open test site to train the classifier. The framework performance is illustrated on the data acquired over a specifically designed test- site.

Evaluation of some features for extended target extraction in polarimetric radar

Detection in most surveillance radars is based on the condition of point targets against a more or less homogeneous background. Currently, the resolution of many new types of radar is increasing, at least in the range dimension. Therefore many objects no longer can be considered as points. Also as a consequence, the background is getting more diverse, in statistical terms. The scene addressed in this paper concerns a ground clutter environment, and extended objects observed with a polarimetric radar with modestly high resolution (i.e. 6m resolving power in range). A staged approach is proposed to detection and parameter assessment of extended objects and adding classification based on polarimetric features. The evaluation of this approach is based on recordings of real natural scenes and artificially inserted extended objects. It has been observed that in the multi-stage detection the object classification benefits from several features, amongst which polarimetric ones. It is proposed that the quality of the contour circumscribing the object is the prime factor for quality of the features next to the polarimetric features. Clutter is affecting the edges of the contours, and therefore may have a major impact on features that are dependent on these contours.

Fully polarimetric measurements over a minefieldlike test site with video impulse ground-penetrating radar

The results of the measurement campaign, which has been held recently at the test facilities for landmine detection systems located at TNO-FEL (The Hague, the Netherlands), are presented. The test facilities give an opportunity to evaluate system performance in different environment (such as grass, sand, clay, etc.) under controlled conditions. The test lanes contain various types of antipersonnel and antitank mines. In this campaign we used the Video Impulse Ground Penetrating Radar that has recently been developed in the IRCTR. The design of the radar allows us to perform simultaneous full-polarimetric measurements in two ultra-wide frequency bands. Furthermore, the scattered from subsurface electromagnetic field is measured in quasi-monostatic and essentially bistatic antenna configurations. The acquired during the measurement campaign data are of high quality in terms of time stability, radar positioning and signal-to-noise ratio. This has allowed to extract full-polarimetric target responses and to analyze them. The obtained results are of importance for target classification.

Evaluation of some features for extended target extraction in polarimetric and non-polarimetric radar

Detection in most surveillance radars is based on the condition of point targets against a more or less homogeneous background. Currently, the resolution of many new types of radar is increasing, at least in the range dimension. Therefore many objects can no longer be considered as points. Also as a consequence, the background is becoming more diverse, in statistical terms. The scene addressed in this paper concerns a ground clutter environment, and extended objects observed with a polarimetric radar with modestly high resolution (i.e. 6 m in range). A step-by-step approach is proposed for the detection and parameter assessment of extended objects and adding classification based on polarimetric features. The evaluation of this approach is based on recordings of real natural scenes and artificially inserted extended objects. It has been observed that in multi-stage detection object classification benefits from several features, including polarimetric ones. It is proposed that the quality of the contour circumscribing the object is the prime factor for quality of features next to polarimetric features. Clutter is affecting, however, the edges of the contours, and therefore may have a major impact on features that are dependent on the shape of these contours. The results also suggest that in the case of large targets with a basically simple shape, like ships, the eccentricity of the shape of the extended object is consistent from scan to scan and probably could support the target tracking.

A SAR-Based Algorithm for Imaging of Landmines with GPR

In this paper we propose a new algorithm for enhancement of imaging of plastic cased antipersonnel landmines using a video-impulse GPR. The algorithm is implemented as a non-linear waveform based signal processor integrated with a SAR focusing procedure. The algorithm constructs a SAR-like image of the subsurface using the outputs of the signal processor. The signal processor searches for presence of reference waveform in raw 1D GPR echo-returns and compresses in time all responses similar to it. Simultaneously, the responses with different waveforms, which presumably correspond to clutter, are suppressed. The shape of the predefined reference waveform depends on the angle at which the imaged point is illuminated. This algorithm is integrated into an automated data processing and mine detection scheme generating a detection list. The reference waveform and other algorithm parameters are determined from training datasets acquired in a controlled environment. The performance of the new algorithm is compared against the performance of the sequence of the signal processor without angle dependency and SAR and against a scheme involving a cross-correlation of the input with the same reference wavelet. The improvement achieved by the suggested algorithm is demonstrated in terms of ROC curves. wavelet is a representation of the target response to the excitation by the probing pulse of the radar and is derived from a set of data acquired in controlled environment conditions. The algorithm is implemented as a two-stage signal processor with a raw A-Scan as an input. The local similarity between the input and the reference wavelet is calculated at the first stage of the processor for each time sample, and the penalty functional is applied to it at the second stage. The output signal of the processor is a range profile, in which position of a target response is marked by a very sharp monopulse, while most of the other reflections are suppressed to nearly zero level. A SAR focusing procedure is applied to the output of the signal processor resulting in subsurface images where the presence of the APM was marked more clearly while some very strong sources of clutter, like pieces of shrapnel or barbed wire were suppressed. However, it has been shown that the targets response wavelet changes with the change of the angle at which it has been illuminated. Due to this, any particular target is marked with highest possible amplitude responses in the output of the PLSM algorithm only for the A-Scans, which correspond to a small range of illumination angles. This shortcoming may diminish the performance of SAR focusing and the following detection. The angle dependency of the reference wavelet can be established either by modeling or from a training dataset but it cannot be readily applied due to an angle ambiguity existing in a time-space C-Scans. More precisely, the illumination angle cannot be tracked in the 3-D domain. That is a wavelet appearing at each time sample of an A-Scan taken from such a domain may correspond to any equidistant object and thus come from any illumination angle. In the present paper we resolve this ambiguity by superimposing the PLSM algorithm with the SAR migration technique. In SAR-focused image of an domain for each imaged point there is only one object placed in it, which is considered as a possible source for the wavelet. Therefore the angle ambiguity is resolved and it is possible to image the given point integrating over the accordingly delayed outputs of PLSM algorithm with reference wavelets dependent on the illumination angle.

Effect of the data pre-processing on GPR mine detector performance

The performance of GPR as an anti-personal mine (APM) detector heavily depends on the quality of the acquired data. A scheme of the quantitative evaluation of the quality of the raw GPR data in terms of stability of certain parameters and the set of the procedures, which compensate the instabilities found are presented. We demonstrate the performance of the data pre-processing chain (DPPC), which includes spectral filtering of A-scans, control of the mutual alignment of B-scans, and time and voltage axiss instabilities compensation in the whole C-scan. The performance of the DPPC is judged both in terms of the stabilization of certain statistical parameters of the data, and by its effect on the receiver operation characteristics (ROC) curve