Urban areas house vegetation cover in several forms, fulfilling several ecological functions like thermal regulator, biodiversity, air quality, etc. However, their extent is often not very well known, especially in Af...Urban areas house vegetation cover in several forms, fulfilling several ecological functions like thermal regulator, biodiversity, air quality, etc. However, their extent is often not very well known, especially in African cities, making it sometimes difficult to assess their real impact on the urban ecosystem functioning. This work aims to analyse the capacity of satellite sensors for mapping vegetation and wetlands in urban areas. The data produced by the MSI sensors of Sentinel 2 and OLI of Landsat-8 are used to identify and map the vegetation cover in the Dakar region through a supervised classification with the Support Vector Machine (SVM) algorithm. The results show that it is sometimes not very easy to analyse urban vegetation with high spatial resolution images (HRS) resulting from the configuration of the vegetation in an urban environment, sometimes characterized by isolated trees or small green spaces. This explains why Sentinel-2 data which spatial resolution of 10 meters gives a better result compared to Landsat-8 data which is 30 meters. However, a good rendering is noted for the vegetation around the wetlands area for the two sensors resulting from the high density and the size of the vegetated perimeters in this part of the capital. Overall, there is an underestimation of urban vegetation cover, particularly for Landsat-8. The use of very high spatial resolution images could be necessary to better assess the potential of satellite data for monitoring urban vegetation in Sahelian context.展开更多
文摘Urban areas house vegetation cover in several forms, fulfilling several ecological functions like thermal regulator, biodiversity, air quality, etc. However, their extent is often not very well known, especially in African cities, making it sometimes difficult to assess their real impact on the urban ecosystem functioning. This work aims to analyse the capacity of satellite sensors for mapping vegetation and wetlands in urban areas. The data produced by the MSI sensors of Sentinel 2 and OLI of Landsat-8 are used to identify and map the vegetation cover in the Dakar region through a supervised classification with the Support Vector Machine (SVM) algorithm. The results show that it is sometimes not very easy to analyse urban vegetation with high spatial resolution images (HRS) resulting from the configuration of the vegetation in an urban environment, sometimes characterized by isolated trees or small green spaces. This explains why Sentinel-2 data which spatial resolution of 10 meters gives a better result compared to Landsat-8 data which is 30 meters. However, a good rendering is noted for the vegetation around the wetlands area for the two sensors resulting from the high density and the size of the vegetated perimeters in this part of the capital. Overall, there is an underestimation of urban vegetation cover, particularly for Landsat-8. The use of very high spatial resolution images could be necessary to better assess the potential of satellite data for monitoring urban vegetation in Sahelian context.
