Pr. J.-P. Antoine*, L. Jacques*,+, Dr. J.-F. Hochedez**
12 May 2000
| * | Institut de Physique Théorique, FYMA, UCL, Louvain-la-Neuve. |
| + | Doctorant F.R.I.A. |
| ** | Department of Solar Physics, Royal Observatory of Belgium. |
Here is an academic example of edges and ridges detection with multiscale selection.
Figure 1: Academic example: an angular
gaussian (128x128x8bits)
Figure 1a: Ridges(*)
Figure 1b: Edges(*)
(*) The scales range from 0.2 to 180, that is from 1 to 80 pixels (filter width)
| Original Images |
|
|
|
|
| Multiscale ridges |
|
|
|
|
| Multiscale edges |
|
|
|
|
Here is a color visualisation of a cumulated multiscale ridges detection in the four following wavelengths:
171 Å |
195 Å |
284 Å |
304 Å |
| Figure 2: |
Each color represents the wavelength of the ridges. There is a thresholding of 0.5% in the ridge measure. |
- How to link the information of different wavelengths ?
Between edges and ridges ?
Purposes:
- Find (automatically) magnetic loops's feet;
- Quantify magnetic loops;
- Classify coronal holes;
- Find polar plumes orientation ...
- How to use the wavelet machinery to decrease
the computational time ?
- J.-P. Antoine, P. Carette, R. Murenzi and B. Piette, Image analysis with two-dimensional continuous wavelet transform, Signal Proc. 31 (1993), 241-272.
- J. Canny. A computational approach to edge detection. IEEE-PAMI, 8(6):679-698,1986.
- T. Lindeberg. Edge detection and ridge detection with automatic scale selection. In Proceeding IEEE Comp. Soc. Conf. on Computer Vision and Pattern Recognition, 1996, pages 465-470, San Francisco, California, June 1996. IEEE Computer Society Press.
