Analysis of Imaging Spectrometer Data with Lake Environment Applications
1999 (English)Doctoral thesis, monograph (Other academic)
In this thesis the image processing and analysis aspects of imaging spectrometer (IS) data have been investigated for water and wetland applications. The Compact Airborne Spectrographic Imager (CASI) has been the main instrument in the evaluations. To fully benefit from the high spectral and spatial resolution data in the analysis phase, the preprocessing of data, is important and has been a focus of this thesis. To restore, improve and evaluate the data, the radiometric calibration, wavelength band positioning, noise and other radiometric anomalies, geometric calibration and atmospheric calibration have been studied. Existing methods have been evaluated, and new ones proposed, and the most appropriate methods applied to the data.
On the image analysis aspects of hyperspectral data sets, spatial true physical structures in the images were studied using data compression and segmentation methods, and a new technique combining compression and colour transformation. The latter was shown to be a fast and objective method to visualise the spatial structures in a large data set.
The usefulness of IS data in water quality applications was evaluated developing statistical relationships between image data and data collected in the field. A comprehensive in situ data set, collected along a transect in Lake Erken, Sweden, during a bloom of the cyanobacteria Gloeotrichia echinulata was used. It was found that a correlation of the image data to chlorophyll a and phaeophytine a could be established, but also that the preprocessing of images is important, and that the dynamic character of water is a complicating factor. Aquatic macrophytes in Lake Mälaren, Sweden, were classified. IS data was found to be powerful for these kinds of applications, but the analysis suffered from poor data.
Place, publisher, year, edition, pages
Uppsala: Acta Universitatis Upsaliensis, 1999. , 121 p.
Uppsala Dissertations from the Faculty of Science and Technology, ISSN 1104-2516 ; 17
Computer Vision and Robotics (Autonomous Systems)
Research subject Computerized Image Analysis
IdentifiersURN: urn:nbn:se:uu:diva-940ISBN: 91-554-4384-2OAI: oai:DiVA.org:uu-940DiVA: diva2:172897
1999-03-12, Room 2005, Ångström Laboratory, Lägerhyddsvägen 1, Uppsala, 10:15 (English)
Lindell, Tommy, Docent