|
Multispectral Image AnalysisMultispectral imaging is an important tool for better
understanding of image formation and reflectance
phenomena.
For this, a multispectral (or, hyperspectral) image
combines
the benefits of spectroscopy with the topological
information of a two-dimensional image. Captured data is
complex and often imperceptible by the human eye. Its
interpretation can be more robust and happen in a broader
scope in comparison with a regular color image. Research
on
computer vision methods that interpret, or rely on, scene
reflectance often profits from analyzing this data.To process this amount of high-dimensional data we need
more
sophisticated procedures for image analysis as well as
efficient ways of handling the large amounts of
information
and visualization in an intuitive way. In this research
project, a novel visualization method is developed that
makes feasible the interactive inspection of the data
prior
to any further processing (e.g. application-dependent
data
reduction). Furthermore, descriptors are studied and
implemented that help separation of geometry,
illumination
and material features. New images are captured using a
specifically acquired hyperspectral camera with high
spectral and spatial resolutions. These are then put to
use
in evaluation and improvement of existing analysis
methods.
| Project manager: Elli Angelopoulou, Ph.D., Akad. Rat
Project participants: Dipl.-Inf. Eva Eibenberger, Dipl.-Inf. Johannes Jordan
Keywords: Multispectral Imaging; Reflection Analysis; Computer Vision
Duration: 1.3.2010 - 31.3.2015
Sponsored by: European Space Agency
Contact: Jordan, Johannes Phone +49 9131 85 27891, Fax +49 9131 85 27270, E-Mail: johannes.jordan@cs.fau.de
| Publications |
---|
Jordan, Johannes ; Angelopoulou, Elli: Gerbil - A Novel Software Framework for Visualization and Analysis in the Multispectral Domain. In: Koch, Reinhard ; Kolb, Andreas ; Rezk-Salama, Christof (Ed.) : VMV 2010: Vision, Modeling & Visualization (15th International Workshop on Vision, Modeling & Visualization Siegen 15.-17.11.2010). Vol. 1, 1. Edition Goslar : Eurographics Association, 2010, pp 259-266. - ISBN 978-3-905673-79-1 | Jordan, Johannes ; Angelopoulou, Elli: Edge Detection in Multispectral Images Using the N-dimensional Self-organizing Map. In: IEEE (Ed.) : 18th IEEE International Conference on Image Processing (ICIP) (18th IEEE International Conference on Image Processing (ICIP) Brussels Sept. 2011). 2011, pp 3181 -3184. [doi>10.1109/ICIP.2011.6116344] | Jordan, Johannes ; Angelopoulou, Elli: Supervised Multispectral Image Segmentation With Power Watersheds. In: IEEE (Ed.) : 19th IEEE International Conference on Image Processing (ICIP) (19th IEEE International Conference on Image Processing (ICIP) Orlando, FL 30.09.2012). 2012, pp 1585-1588. | Jordan, Johannes ; Angelopoulou, Elli: Hyperspectral Image Visualization With a 3-D Self-organizing Map. In: IEEE (Ed.) : Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS), 5th Workshop on (Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS), 5th Workshop on Gainesville, FL June 2013). 2013, pp 1-4. | Jordan, Johannes ; Angelopoulou, Elli: Mean-shift Clustering for Interactive Multispectral Image Analysis. In: IEEE (Ed.) : 20th IEEE International Conference on Image Processing (ICIP) (20th IEEE International Conference on Image Processing (ICIP) Melbourne September 2013). 2013, pp 3790-3794. | Jordan, Johannes ; Angelopoulou, Elli ; Antonio Robles-Kelly: An Unsupervised Material Learning Method for Imaging Spectroscopy. In: IEEE (Ed.) : IEEE WCCI 2014 (2014 International Joint Conference on Neural Networks Beijing 06-07-2014). 2014, pp 2428-2435. |
Institution: Chair of Computer Science 5 (Pattern Recognition)
|
|
|