Project

EnvMon-Short

EnvMon-Short

  • Duration:

Technical evaluation of a robotic measurement system for the spatial-temporal high-resolution in-situ classification of SONAR data using neural networks

Current lake monitoring is limited to punctual measurements, several times during the year and done mostly manual, of quality parameters and visibility depths in the water body to derive the water body condition. However, statements about seasonal fluctuations, the occurrence of macrophytes or the fish population are only possible to a limited extent. Especially in times of algae blooms (summer - early autumn), video observation is only applicable to a very limited extent due to the low visibility and is therefore not very meaningful. Consequently, an additional automated, spatially high-resolution recording and evaluation of the water condition is of high interest. Therefore, the project aims to test various sonar sensors in advance to determine whether they provide sufficient data quality to enable the classification of land cover in particular, but also of macrophytes, macro zoo benthos or fish through a deep neural network. Furthermore, the possibilities of different vehicle platforms and system components for autonomous sonar-based data collection will be evaluated in that context

Partners

Christian-Albrechts-Universität zu Kiel - Institut für Natur- und Ressourcenschutz, Abteilung Hydrologie und Wasserwirtschaft Stein Maritime Consulting (SMC)

Sponsors

Federal Ministry of Education, Science, Research and Technology (BMBF)

02WDG1584B

Federal Ministry of Education, Science, Research and Technology (BMBF)

Share project:

Contact Person
Dr.-Ing. Leif Christensen

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz