FranceCountry of destination:
The W-Arly-Pendjari (WAP) complex is a vast transboundary protected area located in West Africa. It is a key biodiversity hotspot for conservation scientists. However, agricultural expansion is a major threat to biodiversity in this area. Yet, there is currently no detailed map of the available natural resources, which could enable the quantification of the effects of land cover change across the WAP on biodiversity. Remote sensing is a useful tool to address this issue, because it enables the continuous monitoring of vegetation over large extents and throughout the years. New generation satellites, such as offered by the Copernicus constellation (Sentinel-1 and Sentinel-2), provide freely high spatial and temporal resolution co-registered multispectral and radar images of the terrestrial surfaces. Hence, they offer new opportunities to map the natural habitats in the WAP at no cost. Indeed, the high spatial resolution (10 meters) enables the detection of small elements in the landscapes, while the high temporal resolution (5 days) enables the monitoring of vegetation’s dynamics and improves the discrimination of vegetation’s classes based on their phenology. Additionally, the combined use of radar and optical data has often been shown to enhance the accuracy of the vegetation’s classification because of the complementary characteristics of multispectral and radar sensors. However, the contribution of optical and radar image fusion for vegetation mapping has not really been assessed in a savannah landscape such as found in the WAP. Moreover, promising opportunities offered by these satellites bring new methodological and statistical constraints given the high dimension of data (i.e., large number of spectral and temporal measurements) and the massive number of pixels to process. In addition, very few ground truth data are usually available to train the models. It is particularly true in the WAP complex, which is a remote and relatively inaccessible area representing more than 360 millions of Sentinel-2 pixels. Therefore, classification algorithms robust to high dimensional data and computationally efficient are required. This postdoc project aims first to develop robust methods with low computational costs to map natural resources in the WAP complex and assess the state and quality of these resources. Second, it will investigate image fusion techniques for the classification of natural habitats in a tropical area.
I studied agronomy at the French National School of Agricultural Science and Engineering of Toulouse (ENSAT) where I specialised in geomatics. At the end of my engineering studies, I completed a research internship in remote sensing at the Center for the Study of the Biosphere from Space (CESBIO). Following my graduation in 2013, I worked at CESBIO for one year as a study engineer on research projects for the French and the European Space Agencies. Meanwhile, I got a Young Scientist Contract cofinanced by the French National Institute for Agricultural Research and the French National Institute for Research in Computer Science and Automation (CJS INRA-INRIA) which finances a PhD followed by a two-year postdoc. I did my PhD at the Dynamics and Ecology of Agroforestry Landscapes Lab (DYNAFOR) to pursue research in remote sensing of ecology. My PhD thesis focused on the ecological monitoring of semi-natural grasslands using dense satellite image time series with a high spatial resolution. I completed it in 2017 and then started my AgreenSkills fellowship with Dr. Nathalie Pettorelli at the Institute of Zoology (UK) in 2018. This postdoc gives me the opportunity to apply and to extend my research in remote sensing to natural habitats in a biodiversity conservation area.
M. Lopes, M. Fauvel, A. Ouin, and S. Girard. 2017. SpectroTemporal Heterogeneity Measures from Dense High Spatial Resolution Satellite Image Time Series: Application to Grassland Species Diversity Estimation. Remote Sensing. 9(10):993. Doi: 10.3390/rs9100993.
M. Lopes, M. Fauvel, S. Girard, and D. Sheeren. 2017. ObjectBased Classification of Grasslands from High Resolution Satellite Image Time Series Using Gaussian Mean Map Kernels. Remote Sensing, 9(7):688. Doi: 10.3390/rs9070688.
M. Lopes, M. Fauvel, A. Ouin, and S. Girard. 2017. Potential of Sentinel-2 and SPOT5 (Take5) time series for the estimation of grasslands biodiversity indices, 2017 9th Analysis of Multitemporal Remote Sensing Images (MultiTemp) international workshop, Bruges, Belgium, pp. 1-4. hal-01556786v2.
M. Lopes, M. Fauvel, S. Girard, D. Sheeren. 2016. High dimensional Kullback-Leibler divergence for grassland management practices classification from high resolution satellite image time series, 2016 IEEE International Geoscience And Remote Sensing Symposium (IGARSS), Beijing, China, pp. 3342-3345. Doi: 10.1109/ IGARSS.2016.7729864.
D. Sheeren, M. Fauvel, V. Josipovi?, M. Lopes, C. Planque, J. Willm, and J.-F. Dejoux. 2016. Tree species classification in temperate forests using Formosat-2 satellite image time series. Remote Sensing, 8(9):734. Doi: 10.3390/rs8090734.