Tropical land cover and cover characteristics mapping using Planet satellites images from Norway’s International Climate and Forest Initiative (NICFI)

ABSTRACT: The Planet satellite images provided by Norway’s International Climate and Forest Initiative (NICFI, https://www.nicfi.no/) are open data covering the Amazon forest and, more broadly, tropical forest regions between  30 degrees North and 30 degrees South. The Planet NICFI images are multispectral, including red, green, blue, and  near-infrared bands, with a spatial resolution of 4.78 meters for the Normalized Analytic Basemaps. The reflectance  measurements differ from classical satellite missions, such as Landsat, Sentinel, or CBERS, because absolute  radiometric accuracy is not guaranteed. Since September 2020, the temporal resolution of the NICFI images has been  one month, and each image is a mosaic composite of the best daily acquisitions within that month. Consequently,  Planet NICFI images are mostly cloud-free, providing the best freely available multispectral dataset for monitoring  Land Use and Land Cover (LULC) changes in tropical regions. In this session, we will explore recent developments in remote sensing using Planet NICFI data, including near real-time deforestation and degradation mapping, canopy  height mapping, object detection such as mining or forest blowdown, and urban land cover classification. The  presented work utilizes both deep learning and traditional classification methods.