The DLR Institute of Planetary Research carries out research programs on the internal structure, formation and evolution of the planets and their moons, asteroids and comets of our Solar System and other planetary systems. Techniques employed include remote sensing and in-situ investigations using instruments carried on spacecrafts, astronomical observations from the ground, laboratory experiments, and theoretical modelling. The Department of Planetary Geodesy has a long-standing experience in investigations on the shape and surface morphology of planets, moons and asteroids. We apply optical mission data from various sensors for these studies, with a strong focus on the acquisition and analysis of high-resolution stereo images.

Within the project SaiNSOR, we will develop AI-supported solutions to strengthen the connection between sensors and data processing in the context of Earth and planetary remote sensing from orbiters, airborne platforms and rovers. Our approach will be based on coupling data-driven and physical computational models, which must be supported by suitable pre-processing and referencing techniques. The main focus lies on the spatial co-registration of diverse image data, which allows e.g. the analysis of time series. Furthermore, it provides the basis for ingesting and exploiting new observation data in a continuous manner for updating existing models, such as digital surface models. The project will be carried out at the Department of Planetary Geodesy of the Institute of Planetary Research in Berlin, in close cooperation with further partner institutes of DLR. The software solutions and application tests will be further applied e.g. for the development of new sensor concepts for planetary exploration.

To support our group in this exciting project that will combine development and experimental tasks, we are looking for a highly motivated candidate. The ideal candidate will have strong skills in geometric image processing and analysis and in programming, and experience in image analysis for computer vision or remote sensing/photogrammetry. Previous experience in planetary research, geoinformatics, and machine learning is welcome but not necessary. The successful candidate will focus on developing and testing methods for co-registration and model updating to be applied for future mission data. The candidate will be working in a team with extensive expertise in remote sensing, digital mapping techniques, and planetary science. This is a fixed-term appointment for 3 years with remuneration according to the TV-L13 level (75%).

Your tasks:

• development of methods for co-registration and model updating, applied to images in different remote sensing contexts
• software implementation and experimental evaluation of basic scenarios (same sensor, similar observation conditions) and more complex scenarios (different sensors, changes of observation conditions/ the observed surfaces)
• development of a generalized workflow for registering and incorporating new data and for model updating
• interpretation of evaluation results, presentation of the work at national and international conferences, publication in scientific journals