Our group focuses on the development of future EO sensors and state of the art data pre-processing algorithms. Our activities cover a broad spectrum of research fields such as hyperspectral sensors, sensor end-to-end simulations, night illumination, radiative transfer, atmospheric correction, big data processing, as well as geometric fusion of hyperspectral and Lidar data.
The Environmental Mapping and Analysis Program (EnMAP) is a German hyperspectral mission, scheduled for launch in spring 2021. The primary goal of EnMAP is to offer accurate, diagnostic information on the state and evolution of terrestrial ecosystems on a timely and frequent basis, and to allow for a detailed analysis of surface parameters with regard to the characterization of vegetation canopies, rock/soil targets and coastal waters on a global scale. EnMAP is designed to record bio-physical, bio-chemical and geo-chemical variables to increase our understanding of biospheric /geospheric processes and to ensure the sustainability of our resources.
The main objective of the project CHIME E2E is the development of an end-to-end satellite simulator for ESA, which is able to simulate realistically and very accurately the whole chain starting from data recording, sensor calibration and data pre-processing to sensor products up to final surface parameter maps.
GeoMultiSens exploits current scientific and technological advances in Big Data infrastructures, parallel computing environments, Visual Analytics and links them with the potential of satellite remote sensing data to address global challenges such as deforestation, loss of biodiversity, and mega cities.
Artificial light is a unique signature of human activity, with strong correlations to population, economic development level, and electrification rate. It is also a form of global change that remains poorly understood, with consequences for ecosystem services and loss of biodiversity. The spectral, spatial, and temporal patterns of artificial light emission of cities are examined.
Airborne multi-sensor remote sensing data requires careful processing using automatic, consistent, flexible and robust methods. In particular, the fusion of passive hyperspectral with active LIDAR sensor characteristics represents a current research focus. By integrating a high spectral with a high spatial and structural resolution, this type of data fusion allows a more comprehensive object characterization. All synergies regarding geometric and spectral adaptation are exploited to improve both data quality and information content. For this purpose, a variety of modern algorithms and methods are used. A combination of physically driven methods (photogrammetry, ray tracing, radiative transfer modelling) and data driven methods (computer vision, deep learning, point cloud processing) allow for the most realistic measurement of surface reflectance free of geometrical, illumination and viewing angle influences.