Exploitation of spectral indices NDVI, NDWI & SAVI in Random Forest classifier model for mapping weak rosemary cover: application on Gourrama region, Morocco

This work aims to present an efficient and practical method to mapping rosemary cover, which belongs to esparto grasslands. The approach consists of merging tow technics: remote sensing and machine learning. At first, three indices, normalized differenced vegetation index (NDVI), normalized differenced water index (NDWI), and soil adjusted vegetation index (SAVI), were calculated using Sentinel 2A MSI image clipped for the study area. In a second place, a set of terrain truth samples was used to train the model. In the end, the model was used to classify the RGB image built with the three spectral indices. The model was run for the study area, then validated by running it on another region and using samples. Results are maps of rosemary cover densities for both regions. The validation test showed a score of 90%, proving the efficiency of the model.