Using PRISMA Hyperspectral Data for Land Cover Classification with Artificial Intelligence Support

Hyperspectral satellite missions, such as PRISMA of the Italian Space Agency (ASI), have opened up new research opportunities. Using data in land cover classification has yet to be fully explored, and it is main focus this paper. Historically, purposes remote sensing been identify types, detect changes, determine vegetation status forest canopies or agricultural crops. The ability achieve these goals can improved by increasing spectral resolution. At same time, AI algorithms open possibilities. This paper compares three supervised techniques for crop recognition using data: random (RF), artificial neural network (ANN), convolutional (CNN). study was carried out over an area 900 km2 province Caserta, Italy. HDF5 file, pre-processed ASI at reflectance level (L2d), converted GeoTiff a custom Python script facilitate its management Qgis. Qgis plugin AVHYAS used tests. results show that CNN gives better terms overall accuracy (0.973), K coefficient (0.968), F1 score (0.842).