Mapping of built up areas in the Cameroonians shores of Lake Chad and its hinterland through based object classification of Sentinel 2 images

Paul Gérard Gbetkom (1, 2) , Sébastien Gadal (1, 2) , Ahmed El Aboudi (3) , Homère Ngandam (4)

1 AMU - Aix Marseille Université
2 ESPACE - Étude des Structures, des Processus d’Adaptation et des Changements de l’Espace
3 UM5 - Université Mohammed V de Rabat [Agdal]
4 Stats n' Maps, Private Consulting Firm

The Lake Chad is now one of the major territories of concentration of people in the central Sahel region. Since the beginning of the 21st Century, the strong human settlements, especially built-up area, have expanded as a spontaneous adaptation to the upsurge of border conflicts; the main purpose being the access to best lands for agriculture, cattle farming and fishing. Nowadays, the Cameroonian side of the Lake Chad and its hinterland constitute a growing habitable space where we find few towns and numerous hamlets. However, there is no updated map of the built-up sprawl. Therefore, the aim of this study is to bring out the built up areas distribution in that space through orientated-object classification. The methodology implemented combines statistical and spectral processing’s: The third and sixth component of Principal Component Analysis (PCA) which contains respectively 87% and 11 % of information, Normalized Difference Built-up Index (NDBI) and Urban Index (UI), performed on Sentinel 2 images of April 2017. These neo-bands are stacked with originals bands to generate a composite image. The new image obtained is then segmented and classified by Support Vector Machines (SVMs) algorithm and submitted to a threshold analysis in an image on which the built-up areas are clearly separated from the others spatial classes. The results obtained is the delineation the built-up areas, that is the distribution and the structure of the housing space in the Cameroonians shores of Lake Chad and its hinterland.

Mots clés: OBIA; Lake Chad; Built-up areas; Support Vector Machine; Remote Sensing; Sentinel 2 data; NDBI

Domaines: Traitement des images [eess.IV]; Méthodes et statistiques; Environnement et Société; Géographie; Modélisation et simulation

Paul Gérard Gbetkom, Sébastien Gadal, Ahmed El Aboudi, Homère Ngandam. Mapping of built up areas in the Cameroonians shores of Lake Chad and its hinterland through based object classification of Sentinel 2 images. GEOBIA 2018: From pixels to ecosystems and global sustainability, Jun 2018, Montpellier, France. . ⟨hal-01826863⟩