REMOTE IMAGE CLASSIFICATION USING IMPROVED DECISION TREE AND NEURAL NETWORK

Neural systems speak to a generally utilized option to manage remotely sensed picture information. The change of spatial and otherworldly determination in most recent era Earth perception instruments is required to present greatly high computational necessities in neural system based calculations for grouping of high dimensional information sets, for example, hyper ghostly pictures, with several phantom channels and fine spatial determination. A huge preference of neural systems versus different sorts of preparing calculations for hyper ghostly imaging is that they are inalienably amiable for parallel usage. Accordingly, they can advantage from advances in minimal effort parallel figuring architectures, for example, heterogeneous systems of PCs, which have soon turned into a standard apparatus of decision for managing the huge measure of picture information sets. This paper proposes another strategy to characterize remote sensing picture regions with blend of information mining procedures. Gotten aftereffects of proposed framework gives better results contrasted with past strategies.