Skip to Main content Skip to Navigation
Master Thesis

Développement et comparaison de protocoles semi-automatisés de détection d’objets géomorphologiques : les surfaces d’aplanissement

Abstract : Planation surface (PS) maps are mainly built either by manual digitalization or through semiautomated pixel classification, which does not consider relationships between neighboring pixels, can be highly time-consuming and subjective. However, Object Based Image Analysis avoid these biases by using objects. This semi-automated method relies on i) a segmentation step followed by ii) classification. We developed two different protocols using SRTM 30arcsecond data and its derivatives. Both protocols differ in their classification step. The first uses an unsupervised classification based on a clustering algorithm. The second uses a supervised classification based on a machine-learning algorithm from user-defined training samples. Both protocols allow us to identify PS from their slope and curvature with accuracy around 70%.
Document type :
Master Thesis
Complete list of metadata

https://dumas.ccsd.cnrs.fr/dumas-02899768
Contributor : Cnam - Service Commun de la Documentation - Esgt <>
Submitted on : Wednesday, July 15, 2020 - 2:47:30 PM
Last modification on : Friday, January 8, 2021 - 3:15:45 AM

File

LE CADRE Adrien.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : dumas-02899768, version 1

Collections

Citation

Adrien Le Cadre. Développement et comparaison de protocoles semi-automatisés de détection d’objets géomorphologiques : les surfaces d’aplanissement. Sciences de l'ingénieur [physics]. 2019. ⟨dumas-02899768⟩

Share

Metrics

Record views

20

Files downloads

13