Master student (f/m/d) in computer science, engineering, robotics or natural science

Start your career with an interesting master thesis at DLR Solar Research Institute, one of the world’s leading institutions in the field of concentrated solar technology!

Solar tower power plants consist of several hundred to several thousand mirrors, so-called heliostats, which focus solar radiation upon a collector tower. An important step in the qualification of solar tower power plants is the monitoring of the heliostats. Operationally relevant parameters include their alignment and the condition of the mirror panels. To reduce the cost of qualification and automate the monitoring, unmanned aerial vehicles (UAVs) are increasingly being used. These capture aerial images, which are then evaluated by image-processing algorithms. Although conventional image-processing methods already achieve good results, they often still require manual intervention. For this reason, the use of deep learning methods is currently being investigated at our institute.

As part of your thesis, you will support us in further developing our deep learning models for detection tasks in aerial images of heliostat fields. The exact tasks are determined according to the requirements of the university and the applicant's previous knowledge. Possible topics could be model ablation studies, class-agnostic models and object tracking.
You will be part of the AuSeSol-AI project and will be closely integrated into our diverse and highly motivated team in the “Qualification group” located in Almería, Spain. An experienced colleague will closely work together with you and support you throughout the entire thesis.