Skip to Main content Skip to Navigation
Master Thesis

Évaluation comparative de méthodes pour la cartographie et le suivi des renouées asiatiques par télédétection

Abstract : Japanese Knotweed (Fallopia spp.) are one of the most invasive species in the world. Despite of remote sensing works conducted in recent years, overall accuracy is still insufficient to establish a spatiotemporal tracking.
This study suggests, with a comparative approach, the identification of the best way to identifying Fallopia spp. by remote sensing. Different classification algorithms will be used: NN, SVM, RF, and a rule-based classification with fuzzy logic. Two types of images will be used: Pleiades satellite images (spatial resolution: 50cm) and orthophoto RVNIR (made by photogrammetry, 5 cm). On top of that, various types of indices are integrated in these algorithms: spectral, textural and space-time indices. Finally, all result’s accuracy will be compared.
For the time being, only one type of classification has been applied on orthophoto (red, green and NIR ranges) and on Pleiades image (visible and near infrared range): NN. Results show a higher accuracy with classification on orthophoto RVNIR and the use of various indices (textural, BTBR).
These encouraging results will be compared with findings of NN, RF and based-rules classification with fuzzy method, applied on satellite images and orthophoto RVNIR.
Finally, the use of the best classification will justify a spatiotemporal tracking of Fallopia spp. and will facilitate their management.
Document type :
Master Thesis
Complete list of metadatas
Contributor : Iuga Uga <>
Submitted on : Wednesday, April 26, 2017 - 3:30:45 PM
Last modification on : Friday, July 10, 2020 - 7:58:46 AM
Long-term archiving on: : Thursday, July 27, 2017 - 12:52:43 PM


Distributed under a Creative Commons Attribution - NonCommercial - NoDerivatives 4.0 International License


  • HAL Id : dumas-01514683, version 1


Manon Racle. Évaluation comparative de méthodes pour la cartographie et le suivi des renouées asiatiques par télédétection. Géographie. 2016. ⟨dumas-01514683⟩



Record views


Files downloads