Hendrik Krüger

New approach of signal processing for classification problems using a-priori information

Metal detectors based on the eddy current principle are widely used in humanitarian demining, but their main disadvantage is the high false alarm rate caused by harmless clutter objects or ¿uncooperative¿ soils. In most cases, the types of mines in the field are known a-priori. So their influence on the metal detector can be measured and stored in a database. In this paper a new method is described that uses this information to adapt the soil compensation and the signal classification to achieve a significant reduction of false alarms. Therefore a fuzzy classifier is used that adapt to the signal quality and feature variance of the known objects. The ability to lower the false alarm rate was verified by trials on the test lane for mine detection provided by the Joint Research Center in Ispra (Italy).

Solutions for 3D position referencing for handheld metal detectors used in humanitarian demining

Handheld metal detectors based on the eddy current principle are widely used in humanitarian demining, but their main disadvantage is the high false alarm rate caused by harmless clutter objects or “uncooperative” soils. The key to lower the false alarm rate is to provide more information to the deminer by imaging, the use of a-priori information for a better soil compensation and signature classification. Therefore position referenced measurement data is needed. This paper is focused on solutions for the acquisition and processing of position referenced measurements, which are applicable to handheld metal detectors and usable in the field.

Gas Sensing of Nitrogen Oxide Utilizing Spectrally Pure Deep UV LEDs

In this paper, we will present the development of a compact LED-based optical gas sensing system in the ultraviolet-C spectral region. This includes the design of the LED heterostructure emitting near 226 nm, the development of an LED chip, and the implementation into a gas sensing system capable of detecting nitrogen oxide concentrations in the ppm range.

Micromirror array based optical spatial filter technique for fast and flexible velocimetry

Spatial filtering technique is an optical method for measurement of the relative velocity between the sensor and the observed surface(s). This technique is well known for tens of years, its benefits are a potentially simple optical setup and low requirements to the data acquisition and analysis. Therefore it is a valuable method for online contactless velocimetry. The classical spatial filtering setup is based on a static grating in the optical detection path with at least one or two photodetectors. Newer developments are using line array or two-dimensional detectors (e.g. CCD or CMOS). Here the grating structure is realized by weighting the pixels in a signal pre-processing. The flexibility of adapting the grating to the application is much higher but the detector speed of the array sensors is much lower than that of the photodetectors in the classical setup. Both properties are important for a spatial filter sensor that can be used in a wide range of application. The novel micromirror array based optical spatial filter design that is presented in this paper combines the advantages of both state of the art systems. It utilizes fast photodetectors in combination with a very precise, flexible and high resolved differential grating that is realized by the micromirror array. The present paper gives a short overview on the spatial measurement technique, explains the micromirror array based sensor approach and shows the proof of concept.

2D image reconstruction from blurred and disturbed multi parameter measurements at the example of an inductive metal detector

Metal detectors (MD), based on the eddy current principle, are widely-used in humanitarian demining. The sensor signal (i.e. the induced complex coil voltage) is influenced by the object properties (material, shape). This effect was used in previous research projects to classify suspicious objects. But the identification of multi-target situations (a mine covered by a neighboring metal object) was still unsolved. From a 2D raw data image only shapes larger than the coil diameter can be visualized. In this paper the deconvolution of multivariate spatial referenced MD-data is used for imaging object distributions, shapes and material properties. The key to a high resolution image reconstruction is the reduction of disturbance influences caused by soils with magnetic properties. The ability to identify multi-target situations and recognize mines by their characteristic metal distribution was demonstrated on test lanes for mine detection provided by the University of Rostock and the CTRO-Benkovac.

Multivariate data analysis for accuracy enhancement at the example of an inductive proximity sensor

The sensitivity of inductive proximity sensors decreases exponential with increasing distance of the target to the sensing element. The temperature effect and cable influence dominates the measured impedance at the end of the sensing range. For precise measurements, also under a wide specified range of temperature and cable length, disturbance influences must be rejected. Therefore methods, known from the field of non destructive testing and metal detection can be adapted. If the sensing range is defined by a maximal tolerated deviation, a better disturbance rejection leads to a higher range for the proximity sensor.

Mode filter for LED-based absorption spectroscopy

A mode filter for LED-based gas absorption spectroscopy is shown. Two LEDs with different wavelengths are used for the measurement of gas concentrations. One wavelength for measurement and a second wavelength as reference. This allows to measure gases like NO2 (405nm) and SO2 (285nm). Different disturbances, like temperature, vibrations and the optical properties of LEDs require a mode filter to reduce their influences. The mode filter is eight-shaped. It is examined, how many windings on the mode filter are required to suppress the influence of disturbances. Three kinds of disturbances were evaluated: bending a fibre, shifting a fibre connector and changing the temperature of a fibre connector. An improvement with the mode filter is determined for different number of windings. For more than 7...8 windings on the mode filter, there is no significant improvement. Only the losses of the fibre system increase with an increasing number of windings. Mode filters are indispensable for a high sensitive sensor system.

New setup for a real time high resolution UV-LED absorption spectroscopy

A real time, high resolution LED based, spectroscopic gas concentration measurement system is presented. This kind of measuring techniques was enabled by the advances in the fabrication of ultraviolet LEDs (UV-LEDs). The system presented here has a relatively compact design and small application efforts. Different hardware setups have been evaluated to design a system for real time high resolution LED absorption spectroscopy. Applied are two concave mirrors, two photo detector diodes and a 405nm LED mounted in a small device. A standard low power microcontroller was applied as signal processor. The few, robust, reasonable priced components enables high volume applications for this sensor. Different optical simulations provide optimized system parameters for gas concentration measurements in the low ppm range. NO2 concentrations of 200 ppm have been measured with a resolution of ~1.5% counts (16-Bit-ADC). Therefore the ppb range becomes an available measurement range with higher ADC-Resolutions.