Slides of review talk " The Challenges in Spectral Image Analysis: an Introduction, and Review of ANN Approaches", introducing a special session on "Spectral Image analyses with ANNs", organized by E. Merényi, at ESANN'99, European Symposium on Artificial Neural Networks, Bruges, Belgium, April 21-23, 1999. See Paper for text and references.
Slide 1
Slide 2
Concept of (hyper)spectral imaging. From Campbell, 1996.
Slide 3
The difference between traditional broad band and hyperspectral imaging: bandpasses of Landsat TM (at bottom), and AVIRIS (top).
Slide 4
Hyperspectral (AVIRIS) image of Mammoth Mountans, California. 1024 x 512 pixles, 224 bands, approximately 300 Mbytes od data.
Slide 5
Another illustration of the difference between broad band multi-spectral and hyperspectral resolution of the spectral signatures of surface materials. From Clark, 1999.
Slide 6
Methematical model for spectral data.
Slide 7
Slide 8
Slide 9
Slide 10
Example of the advantage of Self-Oragnizing Artificial Neural Net for clustering a spectral image. For more explanation of this slide see AISRP'99 Workshop presentation.
Slide 12
Slide 13
Slide 14
Slide 15
Slide 16