Phase 1: Development of a soil assessment method for GPR sensors within the context of counter-IED (Detection of improvised explosive devices)
Phase 2: Generating sythetic model data for the detection of buried objects with GPR sensors
For some years now, for the detection of buried objects, ground-penetrating radar (GPR) sensors are being used increasingly more often, parallel to already being used for landmines, for the detection of improvised explosive devices (IEDs). Metal detectors fall short because IEDs typically do not contain metals. Therefore complementary GPR sensors are employed (so called “dual sensors”).
In principal, GPR is able to differentiate between any material, given that the contrast of electric or dielectric properties is high enough. However, certain subsoil properties hinder the GPR performance so much, that buried objects can be overlooked. Since the knowledge of the limits of a detection system is essential to survival when looking for buried objects, a reliable method is necessary to accurately predict the performance of a GPR sensor in dependance on the soil properties.
The performance of a GPR is determined by the attenuation of the electromagnetic (EM) waves, the contrast of electric and dielectric properties between object and soil, as well as soil heterogeneity. IEDs and unexploded ordnances (UXOs) are often buried deeper than antipersonnel mines, making the wave attenuation a limiting factor. The attenuation depends on the soil material and, for example, increases with rising salt, clay and water content. The attenuation is often frequency dependent, which can lead to a deformation of the radar signal in time.
The goal of the projects is the development of an assessment method for predicting GPR performance. Here fore we use the time-domain reflectometry (TDR) method to derive the attenuation and EM propagation velocity in the soil from the traveltime and amplitude of the TDR pulses. The TDR method can, due to its quick handling, also be used in the field during clearance operations to directly obtain GPR performance for the current and local soil properties.
Besides the attenuation, the soil heterogeneity and surface roughness are limiting factors, producing noise in data. This presents a problem particularly then, when the contrast of the object to the soil is very weak and the reflecting radar amplitudes correspondingly small.
The setup of test sites for studying GPR sensor performance is very complex and expensive, and reflects only a limited number of combinations of soils and target objects. Therefore, in the current project, simulations of various scenarios are made and synthetic model data are generated to systematically analyze the different influencing variables and specify the limits of GPR sensors in heterogeneous soils. The synthetic data can be used to train UXO-detection personnel or pattern-recognition algorithms.
Dr. Jan Igel
+49 511 643-2770