Using lidar and aerial photography to build a geographic object database tuned for ecological model

Ecological models require a large variety of variables in order to describe the biotic and abiotic conditions that define wildlife habitats and biotope distribution. Instead of providing this information using regular grids or categorical maps, geographic object-based image analysis allows to describe the land cover based on meaningful spatial regions which are closer to the natural habitats than regular grids and more flexible to the specific needs of models than categorical data. In the frame of the Lifewatch/Wallonia-Brussels project, we use GEOBIA to create a geographic object database closely related to the ecological concept of ecotope. Through an iterative process with ecological modellers, this database has been designed in order to define its key characteristics : it had to provide a quantitative description of the land cover within topographically relevant spatial regions and it integrates contextual information from different scales. The proof of concept of this new type of geographic database was done over the Wallon region in Belgium. This study area covers approximately 16800 square kilometers with a very fragmented landscape. A dataset including aerial photographs at 0.25 cm resolution and LIDAR at 0.8 pts/m has been provided by theWalloon region for the study. The data were resampled at 2m resolution for the purpose of the analysis. The data processing workflow includes three steps : pixel-based image classification, image segmentation and object-based integration. Pixel-based image classification consists in a supervised classification with the spectral values from the aerial photographs (NIR/Red/green/blue), the Digital Height Model extracted from the LIDAR and the intensity of the first LIDAR return. This yielded a classification into broadleaved trees, needleleaved trees, grass, bare soil, crop, pavement, building, water and shadows with more than 80% overal accuracy. The image segmentation approach is the main novelty of this research. In order to fit with the biotopes, image segments indeed had to take the type of slope into account. This was achieved by computing pseudo-hillshades for North-South andWest-East orientation and including those two files together with the spectral information from the aerial photographs. The result of this analysis is a set of topographically relevant ecotope delineation. The last step applied contextual decision rules to consistently aggregate the land cover information at the ecotope level and add more information from ancillary datasets.