Mandjo Béa BoréData analyst - Developer

Mapping forest ecosystems represents a major challenge for natural resource conservation. In Africa, Raffia forests play a crucial ecological role, but their precise delineation remains a technical challenge due to the heterogeneity of tropical rainforests.

In this research project conducted in collaboration with IRD (Institut de Recherche pour le Développement) and the AMAP Laboratory, my objective was to assess the potential of very high spatial resolution satellite images for mapping Raffia stands in southern Cameroon.

To achieve this, I applied various satellite imagery analysis methods. I used object-based image analysis with eCognition to segment the images, combined with processing in ENVI and spectral analyses. I also implemented the FOTO method (Fourier-based Textural Ordination) in Matlab to characterize canopy texture. Spatial results were integrated and visualized in ArcGIS.

The results demonstrated that the Fourier method proves particularly promising for isolating Raffia stands from other vegetation types. However, the high heterogeneity of tropical forests, accentuated by very high spatial resolution, constitutes a major constraint requiring adapted methodological approaches.