Reference : Mapping of riparian invasive species with supervised classification of Unmanned Aeria...
Scientific journals : Article
Life sciences : Environmental sciences & ecology
http://hdl.handle.net/2268/184730
Mapping of riparian invasive species with supervised classification of Unmanned Aerial System (UAS) imagery
English
[fr] Cartographie de plantes invasives des bandes riveraines à l'aide d'imagerie drone (UAV)
Michez, Adrien mailto [Université de Liège > Ingénierie des biosystèmes (Biose) > Gestion des ressources forestières et des milieux naturels >]
Piégay, Hervé [> >]
Lisein, Jonathan mailto [Université de Liège > Ingénierie des biosystèmes (Biose) > Gestion des ressources forestières et des milieux naturels >]
Claessens, Hugues mailto [Université de Liège > Ingénierie des biosystèmes (Biose) > Gestion des ressources forestières et des milieux naturels >]
Lejeune, Philippe mailto [Université de Liège > Ingénierie des biosystèmes (Biose) > Gestion des ressources forestières et des milieux naturels >]
Feb-2016
International Journal of Applied Earth Observation and Geoinformation
Elsevier Science
44
Yes (verified by ORBi)
International
0303-2434
1569-8432
Amsterdam
The Netherlands
[en] Mapping of invasive species ; Unmanned aerial system ; UAS ; Supervised classification ; Random forests
[en] Riparian zones are key landscape features, representing the interface between terrestrial and aquatic ecosystems. Although they have been influenced by human activities for centuries, their degradation has increased during the 20th century. Concomitant with (or as consequences of) these disturbances, the invasion of exotic species has increased throughout the world’s riparian zones.
In our study, we propose a easily reproducible methodological framework to map three riparian invasive taxa using Unmanned Aerial Systems (UAS) imagery: Impatiens glandulifera Royle, Heracleum mantegazzianum Sommier and Levier, and Japanese knotweed (Fallopia sachalinensis (F. Schmidt Petrop.), Fallopia japonica (Houtt.) and hybrids). Based on visible and near-infrared UAS orthophoto, we derived simple spectral and texture image metrics computed at various scales of image segmentation (10,30, 45, 60 using eCognition software). Supervised classification based on the random forests algorithm was used to identify the most relevant variable (or combination of variables) derived from UAS imagery for mapping riparian invasive plant species. The models were built using 20% of the dataset, the rest of the dataset being used as a test set (80%).
Except for H. mantegazzianum, the best results in terms of global accuracy were achieved with the finest scale of analysis (segmentation scale parameter = 10). The best values of overall accuracies reached 72%, 68%, and 97% for I. glandulifera, Japanese knotweed, and H. mantegazzianum respectively. In terms of selected metrics, simple spectral metrics (layer mean / camera brightness) were the most used. Our results also confirm the added value of texture metrics (GLCM derivatives) for mapping riparian invasive species.
The results obtained for I. glandulifera and Japanese knotweed do not reach sufficient accuracies for operational applications. However, the results achieved for H. mantegazzianum are encouraging. The high accuracies values combined to relatively light model-inputs needed (delineation of a few umbels) make our approach a serious contender as a cost-effective tool to improve the field management of H. mantegazzianum.
Service public de Wallonie : Direction générale opérationnelle de l'agriculture, des ressources naturelles et de l'environnement - DG03
Researchers ; Professionals ; Students
http://hdl.handle.net/2268/184730
10.1016/j.jag.2015.06.014

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Restricted access
1-s2.0-S0303243415300040-main.pdfPublisher postprint1.25 MBRequest copy

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.