26 October 2016 Extraction of urban vegetation with Pleiades multiangular images
Author Affiliations +
Abstract
Vegetation is essential in urban environments since it provides significant services in terms of health, heat, property value, ecology ... As part of the European Union Biodiversity Strategy Plan for 2020, the protection and development of green-infrastructures is strengthened in urban areas. In order to evaluate and monitor the quality of the green infra-structures, this article investigates contributions of Pléiades multi-angular images to extract and characterize low and high urban vegetation. From such images one can extract both spectral and elevation information from optical images. Our method is composed of 3 main steps : (1) the computation of a normalized Digital Surface Model from the multi-angular images ; (2) Extraction of spectral and contextual features ; (3) a classification of vegetation classes (tree and grass) performed with a random forest classifier. Results performed in the city of Rennes in France show the ability of multi-angular images to extract DEM in urban area despite building height. It also highlights its importance and its complementarity with contextual information to extract urban vegetation.
Conference Presentation
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Antoine Lefebvre, Antoine Lefebvre, Jean Nabucet, Jean Nabucet, Thomas Corpetti, Thomas Corpetti, Nicolas Courty, Nicolas Courty, Laurence Hubert-Moy, Laurence Hubert-Moy, } "Extraction of urban vegetation with Pleiades multiangular images", Proc. SPIE 10008, Remote Sensing Technologies and Applications in Urban Environments, 100080H (26 October 2016); doi: 10.1117/12.2241162; https://doi.org/10.1117/12.2241162
PROCEEDINGS
8 PAGES + PRESENTATION

SHARE
Back to Top