Segmentation d'orthophotoplans et invariants colorimétriques pour la détection de toitures
Résumé
In this paper, a segmentation method based on the watershed algorithm using an optimal gradient and colorimetic invariant is presented. This segmentation technique is used in the context of roof detection in orthophotoplans images. To achieve that, eight gray level gradients, six color gradients and 24 colorimetric invariant are tested on a database of 67 roofs. The results shows the quality of the segmentation according to the gradient/colorimetric invariant used. To conclude, the best gradient/colorimétric invariant for roofs detection is given.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|
Loading...