< Back to previous page


Paul Scheunders

  • Research Expertise  (University of Antwerp):Processing and analysis of multispectral and hyperspectral (signal and image) data for remote sensing and close-range applications. The data is reflectance spectra, in the visible, near and shortwave-infrared wavelength region, acquired by multi-and hyperspectral image sensors, mounted in satellites, airplanes, drones, and in close-range situations. The data-analysis techniques are based on machine learning and AI. A large number of methods was developed: supervised, unsupervised and semi-supervised classification methods and methods based on spectral unmixing.The results are detailed landcover maps of the Earth surface, or information about the fractional abundances of the different materials contained within the field of view of the sensor. Applications are developed for remote sensing (geologie, vegetation monitoring, …) and non-destructive monitoring of materials for e.g. corrosion and concrete damage.
  • Disciplines  (Interuniversity Microelectronics Centre):Display technology, Antennas and propagation, Automation and control systems, Analogue, RF and mixed signal integrated circuits, Neuromorphic computing, Audio and speech processing, Environmental safety and health of nanotechnology, Battery technology, Biomaterials, Ceramic matrix composites, Hybrid composites, Computational materials science, Metals and alloy materials, Polymer processing, Nanomaterials, Functional materials, Biomedical image processing, Biochemical engineering, Biofluid mechanics, Cell, tissue and organ engineering, Arithmetic and logic structures, Automation, feedback control and robotics, Bio-informatics, Analogue and digital signal processing, Engineering instrumentation, Environmental health and safety, Biostatistics, Care for disabled, Bioethics
  • Disciplines  (University of Antwerp):Machine learning and decision making, Physical geography and environmental geoscience, Signal processing, Geomatic engineering
  • Research techniques  (University of Antwerp):Hyperspectral image analysis methods (techniques and algorithms): - supervised methods for spectral classification - contextual classification - multisource classification - image fusion methods - decision fusion methods - semi-supervsed and active learning methods - kernel methods - manifold learning methods - methods for linear and nonlinear spectral unmixing (dimensionality estimation, endmember selection, abundance estimation). - methods for subpixel (superresolution) mapping
  • Users of research expertise  (University of Antwerp):Users of remote sensing products: public and private stakeholders in domains of agriculture, forestery, landmanagement, environmental monitoring. Developers of remote sensing products: data acquisition, collection, processing and distribution, software developers. Developers and users of close-range applications (corrosion monitoring, concrete degradation, oil spill, …)