Nada Milisavljevic

Filtering Soil Surface and Antenna Effects From GPR Data to Enhance Landmine Detection

The detection of antipersonnel landmines using ground-penetrating radar (GPR) is particularly hindered by the predominant soil surface and antenna reflections. In this paper, we propose a novel approach to filter out these effects from 2-D off-ground monostatic GPR data by adapting and combining the radar antenna subsurface model of Lambot with phase-shift migration. First, the antenna multiple reflections originating from the antenna itself and from the interaction between the antenna and the ground are removed using linear transfer functions. Second, a simulated Greens function accounting for the surface reflection is subtracted. The Greens function is derived from the estimated soil surface dielectric permittivity using full-wave inversion of the radar signal for a measurement taken in a local landmine-free area. Third, off-ground phase-shift migration is performed on the 2-D data to filter the effect of the antenna radiation pattern. We validate the approach in laboratory conditions for four differently detectable landmines embedded in a sandy soil. Compared to traditional background subtraction, this new filtering method permits a better differentiation of the landmine and estimation of its depth and geometrical properties. This is particularly beneficial for the detection of landmines in low-contrast conditions

Sensor fusion in anti-personnel mine detection using a two-level belief function model

A two-level approach for modeling and fusion of antipersonnel mine detection sensors in terms of belief functions within the Dempster-Shafer framework is presented. Three promising and complementary sensors are considered: a metal detector, an infrared camera, and a ground-penetrating radar. Since the metal detector, the most often used mine detection sensor, provides measures that have different behaviors depending on the metal content of the observed object, the first level aims at identifying this content and at providing a classification into three classes. Depending on the metal content, the object is further analyzed at the second level toward deciding the final object identity. This process can be applied to any problem where one piece of information induces different reasoning schemes depending on its value. A way to include influence of various factors on sensors in the model is also presented, as well as a possibility that not all sensors refer to the same object. An original decision rule adapted to this type of application is proposed, as well as a way for estimating confidence degrees. More generally, this decision rule can be used in any situation where the different types of errors do not have the same importance. Some examples of obtained results are shown on synthetic data mimicking reality and with increasing complexity. Finally, applications on real data show promising results.

Improving mine recognition through processing and Dempster–Shafer fusion of ground-penetrating radar data

A methodfor modeling andcombination of measures extractedfrom a ground-penetrating radar (GPR) in terms of belief functions within the Dempster-Shafer framework is presentedandillustratedon a real GPR data set. A starting point in the analysis is a preprocessed C-scan of a sand-lane containing some mines andfalse alarms. In order to improve the selection of regions of interest on such a preprocessed C-scan, a methodfor detecting suspectedareas is developed, basedon region analysis aroundthe local maxima. Once the regions are selected, a detailedanalysis of the chosen measures is performed for each of them. Two sets of measures are extracted and modeledin terms of belief functions. Finally, for every suspected region, masses assigned by each of the measures are combined, leading to a first guess on whether there is a mine or a non-dangerous object in the region. The region selection methodimproves detection, while the combination methodresults in significant improvements, especially in eliminating most of the false alarms.

Characterization of mine detection sensors in terms of belief functions and their fusion, first results

In this paper, characterization of mine detection sensors in terms of belief functions and their fusion are presented. The need for fusion of mine detection sensors is discussed, and reasons for choosing Dempster-Shafer framework are pointed out, taking into account specificity and sensitivity of the problem. This work is done in the scope of the HUDEM project, where three promising and complementary sensors are under analysis. These sensors are presented, and detail analysis is performed in case of fusing the data from them. A way for including in the model influence of various factors on sensors and their results is discussed as well and will be further analyzed in the future. The application of the approach proposed in this paper is illustrated on the frequent case of detecting metallic objects, but the possibility for modifying it to some other situations exists.

Possibilistic Versus Belief Function Fusion for Antipersonnel Mine Detection

Two approaches for combining humanitarian mine detection sensors are presented-one based on belief functions and the other one based on possibility theory. The approaches are described in parallel. First, different measures are extracted from the sensor data. Mass functions and possibility distributions are then derived from the measures based on prior information. After that, the combination of masses and the combination of possibility degrees are performed in two steps, on a separate sensor level and between the sensors. Combination operators are chosen to account for different characteristics of the sensors. The selection of the decision rules is discussed for both approaches. The proposed approaches are illustrated on a set of real mines and nondangerous objects, and promising results have been obtained.

Clutter reduction in GPR measurements for detecting shallow buried landmines: a Colombian case study

Detection of land-mines from ground-penetrating radar data is a challenging task demanding accurate and useful filtering techniques to reduce soil-surface and antenna reflections, which obscure the land-mine response. In this paper, we apply and adapt a recently proposed filtering approach to enhance the detection of shallow buried anti-personnel land-mines from data acquired in typical mine-affected soils in Colombia. The methodology combines a radar-antenna-subsurface model with phase-shift migration to filter out antenna and soil-surface effects from off-ground monostatic radar two-dimensional data. Firstly, antenna Multiple reflections are removed using linear transfer functions. Secondly, simulated Greens functions accounting for the surface reflection are subtracted. These functions are derived using the relative dielectric permittivity of the surface, which is estimated by full-wave inversion of the radar signal for measurements taken in local land-mine-free areas. Finally, the antenna radiation pattern effect is filtered out by performing phase-shift migration, and information about size and shape is extracted. Data are acquired using a hand-held vector network analyser connected to an off-ground monostatic horn antenna. Typical Colombian targets such as low-metallic anti-personnel land-mines and low- and non-metallic improvised explosive devices are used. Results prove that the proposed technique effectively reduces clutter under non-controlled conditions and yields target features that are useful for detection of these land-mines.

Comparison of three methods for shape recognition in the case of mine detection

Abstract Three methods for shape recognition and their usefulness are analyzed and compared in the scope of the antipersonnel mine detection problem. It is shown that, because of the variety and complexity of the problem, neither of them can be completely discarded.

Comparison of belief functions and voting method for fusion of mine detection sensors

In this paper, two methods for fusion of mine detection sensors are presented, based on belief functions and on voting procedures, respectively. Their application is illustrated and compared on a real multisensor data set collected at the TNO test facilities under the HOM-2000 project. This set contains data acquired by metal detector, infrared camera and ground penetrating radar. The data acquisition and preprocessing are briefly described. For some typical cases presented in this data set, the characteristics extracted and used by both methods are discussed, as well as the answers given by each method and possible causes of potential differences in results. Also, it is shown how the different voting schemes compare to belief functions modeling in various situations, based on the knowledge that is put into the belief functions. Since the roots of the two methods are different, i.e. belief functions involve expert knowledge while voting is a simple approach, the explanations involve these differences. Problems that arise when comparing and evaluating different methods are also addressed. Finally, it is shown that both of the methods have their advantages and drawbacks, depending on the measurement and operational conditions. This paper is a result of a joint work at three European institutions towards a common goal: humanitarian demining.

Robust Techniques for Coherent Change Detection Using Cosmo-Skymed SAR Images

The satellite-borne SAR (Synthetic Aperture Radar) is a quite promising tool for high-resolution geo-surface measurement. Recently, there has been a great interest in Coherent Change Detection (CCD), where the coherence between two SAR images is evaluated and analyzed to detect surface changes. The sample coherence threshold may be used to distinguish between the changed and unchanged regions in the scene. Using COSMO-SkyMed (CSK) images, we show that for changed areas, the coherence is low but not completely lost. This situation, which is caused by the presence of bias in the coherence estimate, considerably degrades the performance of the sample threshold method. To overcome this problem, robust detection in inhomogeneous data must be considered. In this work, we propose the application and improvement of three techniques: Mean Level Detector (MLD), Ordered Statistic (OS) and Censored Mean Level Detector (CMLD), all applied to coherence in order to detect surface changes. The probabilities of detection and false alarm are estimated experimentally using high-resolution CSK images. We show that the proposed method, CMLD with incorporation of guard cells (GC) in the range direction, is robust and allows for nearly 4% higher detection probability in case of low false alarm probability.

Multisensor Data Fusion for Spaceborne and Airborne Reduction of Mine Suspected Areas

The problem of mined area reduction is addressed in this paper. Pieces of information collected using airborne multispectral scanners and airborne full polarimetric SAR, together with context information, all integrated in a geographical information system, are classified and combined in order to find indicators of mine presence and mine absence and provide image analysts with adequate tools to interpret mined scenes during the area reduction process. The paper contains a broad description of the whole problem and of the developed method and focuses on classification and data fusion tools based on the belief function framework and fuzzy sets theory.