Mapping woody vegetation clearing in Queensland, Australia from Landsat imagery using the Google Earth Engine

摘要: Monitoring of vegetation clearing in Australia is the province state governments. Only recently have data and services become available for generalised access to change detection tools suited this task. The objective research was examine if a globally cloud computing service, Google Earth Engine Beta, could be used predict decreases woody with accuracies approaching those obtained by government Queensland, Australia. This compared remote sensing results derived reported Queensland Government, using their standard methods. Four approaches were investigated Landsat-5 TM 7 ETM+ time-series algorithms through Application Programming Interface: (1) Classification Regression Tree (CART) (2) Random Forest classifiers; normalised (3) Normalised Difference Vegetation Index (NDVI) (4) Foliage Projective Cover (FPC) combined spectral index, between two image composites. CART classifiers produced highest user's (78–92%) producer's (55–77%) mapping against loss maps Government when detecting within epochs which training available. Extrapolation without reduced accuracies. FPC NDVI more robust calculating probability, as no required, can hence tuned provide automated alerts large events selecting suitable thresholds. provides foundation build further capacity use accessible, free, online datasets processing detect an manner.