Data analysis and interpretation

Today, the data analysis and interpretation in complex technical processes is often a manual process: Data are analysed by humans and interpreted accordingly. High data volume is increasingly becoming a problem that can lead to highly complex situation descriptions and dependencies. Today, not only is the 3D geometry of the surfaces of artificial and natural objects recorded, but also potential deformation variables such as temperature, precipitation, load conditions, etc. The modelling of these complex interrelationships is the object of the research and development work. In the past years, there were great changes in data acquisition: the implementation of mobile systems, which operate free of global reference system, increase. New approaches, e.g. analysis of planar changes, are being prepared and implemented. The insights and results of current sensor development are pursued. At the end, actions for the preservation and safe operation of large natural and artificial structures can be derived. The summary of the research and development goals in this area appears as follows:

• Research and development of novel processes for data analysis (e.g. evaluation of data quality) while taking into account the underlying sensor system.
• Research and development of data interpretation components based on self-learning algorithms (e.g. artificial neural networks).
• Research and development of data processing methods for the evaluation of interpretation results (as a basis for the human decision-making process).

Modern sensors and sensor concepts yield an abundance of data that, after chronological and spatial referencing, yield a comprehensive description of the object. To minimize the required effort, there needs to be a focus on critical areas. The identification and location of these "hotspots" is possible on stored knowledge about the object structure (e.g. tunnel or bridge type), the critical situations and areas identified in the past and the available measurements.