BAM - Big-Data-Analytics in Environmental and Structural Monitoring

Active
Flowchart of the project Vanessa Liebler for the i3mainz, CC BY SA 4.0

An interdisciplinary team at Mainz University of Applied Sciences is researching the the potentials of current data mining and machine learning methods for learning methods for problems with a space-time relationship.

Motivation

The aim of the research project is to provide methods that will increase the usefulness of rapidly growing data volumes with a spatial reference. Thus, in the information systems based on machine learning are being developed in the which facilitate decision-making by means of analysis and visualization. facilitate decision-making. Furthermore, the degree of autonomy of optical monitoring systems on the basis of based on image analysis is being increased with the aid of deep learning systems.

Activities

Through the exchange of interests with various project partners from the mobility sector, intersections could be discovered, use cases developed, and promising data promising data were brought together. On this basis Alexander Rolwes and Thomas Müller are implementing a prototype application to for predicting parking garage occupancy in Mainz. They are currently working on the further development of the prediction as well as the integration into a holistic control model for the use of off-street parking spaces. Furthermore, they they are investigating the relationships between spatial factors and parking behavior. parking behavior.

Kira Zschiesche and Denise Becker are implementing promising methods for monitoring changes in structures. Among other things, they are gaining knowledge in connection with methods from structural health monitoring by performing optical vibration measurements on structures and testing initial approaches for automatic crack detection. To extend track safety, they are using a deep learning approach to develop a software solution that identifies trains in images through the use of artificial intelligence. In the future, the team will explore methods for automatic target detection to increase the level of automation in various processes.