AcoGPR - Adaptive Contactless Ground Penetrating Radar
Todays urban plannings face problems that have never been thought to become true: While nearly all cities and municipalities know for sure that supply line (such as Gas, Water pipelines) cross and follow urban streets is some directed way, their exact position is almost always either unsure, or, in the worst case, completely unknown. The problems resulting from the current situations are obvious: Once the street undergoes maintenance works (e.g. digging is needed to enhance the current infrastructure), pipes are broken easily. This not only causes heavy problems with residents being cut off from some supplies for a period of time, but more severely causes high costs in rebuilding the pipes to the previously working state.
The Information Systems and Machine Learning Lab (ISMLL), headed by Prof. Schmidt-Thieme, in junction with the Institute for High-Frequency Technology at the University of Braunschweig and Detectine GmbH are now starting their collaboration to face this challenging problem. The ISMLL is devoted to the task of building and advancing current state of the art Machine Learning techniques to improve the accuracy when locating both position and radius of pipes of different kind. The Institute for High-Frequency Technology researches in advanced concepts when building radar systems and antenna types, to improve the quality of the signal, making it possible to deduce additional characteristics of the pipes and its surroundings, as well as moving from currents 'near-group' positioning of the sensor towards a higher positioning, making the overall approach more applicable on cluttered surfaces. Detectino GmbH as our local application partner evaluates our prototypes in real-world settings.
Partner:
Kontact Information:
Projects:
- Andre Busche, Ruth Janning, Lars Schmidt-Thieme (2013):
Analysing the Potential Impact of Labeling Disagreements for Engineering Sensor Data, in Proceedings of the LWA 2013 - Lernen, Wissen und Adaptivität, Knowledge Discovery and Machine Learning track, Bamberg. - Ruth Janning, Carlotta Schatten, Lars Schmidt-Thieme (2013):
HNNP - A Hybrid Neural Network Plait for Improving Image Classification with Additional Side Information, in Proceedings of the IEEE International Conference on Tools with Artificial Intelligence, ICTAI 2013.
- Andre Busche, Ruth Janning, Tomas Horvath, Lars Schmidt-Thieme (2012):
A Unifying Framework for GPR Image Reconstruction, in 36nd Annual Conference of the Gesellschaft für Klassifikation (GfKl 2012). - Andre Busche, Ruth Janning, Tomas Horvath, Lars Schmidt-Thieme (2012):
Some Improvements for Multi-Hyperbola Detection on GPR Data, in Workshop on Knowledge Discovery, Data Mining and Machine Learning (KDML 2012). - Ruth Janning, Tomas Horvath, Andre Busche, Lars Schmidt-Thieme (2012):
Pipe Localization by Apex Detection, in Proceedings of the IET international conference on radar systems (Radar 2012), Glasgow, Scotland. - Ruth Janning, Tomas Horvath, Andre Busche, Lars Schmidt-Thieme (2012):
GamRec: a Clustering Method Using Geometrical Background Knowledge for GPR Data Preprocessing, in Artificial Intelligence Applications and Innovations (IFIP Advances in Information and Communication Technology 381), Springer, Heidelberg, Halkidiki, Greece, pp. 347--356. - Daniel Seyfried, Andre Busche, Ruth Janning, Lars Schmidt-Thieme, Jörg Schöbel (2012):
Information Extraction From Ultrawideband Ground Penetrating Radar Data: A Machine Learning Approach, Proceedings of the 7th German Microwave Conference .
