Hyoung-sun Youn

Automatic GPR target detection and clutter reduction using neural network

Ground penetrating radar (GPR) has been widely used for the detection and location of buried objects. However, the detection method is often subjected to operators interpretation due to large quantities of data and undesired clutter and noise. Such a detection method is neither reliable nor efficient.

An internet based interactive telemedicine system for remote healthcare

To enable remote monitoring and diagnosis of patients located far from medical facilities, a prototype Internet based telemedicine system is developed. The developed telemedicine system establishes an audio-visual connection as well as a data path from the patients home to the medical personnels office through a web browser interface. A remote controlled wireless camera installed on the patient side can be tilted, zoomed, and panned by the doctor, at the medical facility, to examine different parts of the patients body through the web interface. Unlike many of the existing systems, the established data path enables additional diagnosis functions such as electrocardiogram, blood pressure, respiration, temperature, and stethoscope. Results from laboratory and field testing of the developed system are presented and avenues for broader implementation in integrated facilities using low cost wireless networks will be discussed.

Special GPR antenna developments for landmine and UXO surveys

Ground penetrating radar (GPR) is a promising technique for detecting and identifying buried landmines and UXOs. The antenna plays a critical role in determining the performance for tasks and requires to have a broad bandwidth, minimal ground dependency, fast or no antenna ring-down, desirable polarization and pattern characteristics. Three unique UWB dual-polarization GPR antenna designs are discussed. A dielectric rod antenna, covering from 1 GHz to 6 GHz, is suitable for shallow (less than 12 inches) targets, such as landmine detection/identification. The other two antennas, covering frequencies from 20 MHz to 800 MHz, are suitable for general GPR applications, such as the detections of pipes, utility lines, UXOs, AT mines, etc.

Neural detection for buried pipes using fully-polarimetric ground penetrating radar system

Ground penetrating radar (GPR) has been widely used for detecting and locating buried objects. However, the detection method using GPR is often subjected to operator interpretation due to large quantities of data and the presence of undesired clutter and noise. The artificial neural network (ANN) technique gives a promising approach to a more systematic and autonomous detection system. An automatic buried pipe detection algorithm, using a two-step ANN scheme on GPR data, is proposed. The detection performance of each ANN in the presence of different signal-to-noise and signal-to-clutter ratios is discussed. Estimating the linearity and orientation of the pipe by fully-polarimetric GPR is reviewed, and applying these factors to pipe detection is discussed. Examples of the two-step ANN detection application to actual field data measured by fully-polarimetric GPR is also presented.

Automatic UXO classification for fully polarimetric GPR data

This paper presents an automatic UXO classification system using neural network and fuzzy inference based on the classification rules developed by the OSU. These rules incorporate scattering pattern, polarization and resonance features extracted from an ultra-wide bandwidth, fully polarimetric radar system. These features allow one to discriminate an elongated object. The algorithm consists of two stages. The first-stage classifies objects into clutter (group-A and D), a horizontal linear object (group-B) and a vertical linear object (group-C) according to the spatial distribution of the Estimated Linear Factor (ELF) values. Then second-stage discriminates UXO-LIKE targets from clutters under groups B and C. The rule in the first-stage was implemented by neural network and rules in the second-stage were realized by fuzzy inference with quantitative variables, i.e. ELF level, flatness of Estimated Target Orientation (ETO), the consistency of the target orientation, and the magnitude of the target response. It was found that the classification performance of this automatic algorithm is comparable with or superior to that obtained from a trained expert. However, the automatic classification procedure does not require the involvement of the operator and assigns a unbiased quantitative confidence level (or quality factor) associated with each classification. Classification error and inconsistency associated with fatigue, memory fading or complex features should be greatly reduced.