KIRAS Security Research

In the research project MiTrAs new methods for the monitoring, early recognition and trend analysis are developed. The fusion of information from satellite images, Open Source and Social Media from non-European countries are used to estimate developments in dedicated applications areas, while focusing on the following substantial core research areas:

• Evaluation of the feasibility of multi-temporal satellite images for object recognition such as groups of people and vehicles based on the development of automated Deep Learning-Analysis methods (e.g. convolutional neural networks) applied at high resolution panchromatic and multispectral image data. New technologies like high resolution radar satellites and the implementation of satellite video are examined for the first time and evaluated regarding their suitability for the recognition of small objects.
• For the analysis and the formatting of heterogeneous data from Open Source resources and Social Media communication a Deep-Learning method and algorithms are developed. An essential criterion in doing so is to question the above-mentioned areas of application with respect to plausibility, timeliness and geographic origin.
• Algorithms and interpretation concepts for the fusion of spatially and chronologically differing information sources that enable relating the analysed observations regarding plausibility and relevance to each other. The research findings are demonstrated and evaluated by means of easy to use visualization.