High Resolution Remote Sensing Image Classification of Coastal Zone and Its Automatic Realization

Using the Quickbird remote sensing image of coastal zone of Yalu River mouth, this paper introduces the conception of separating layers, puts forward a high resolution remote sensing classification method of land use by integrating Maximum Likelihood, principal components analysis and band math. Its classification precision of water, mudflat and plough land are all higher than 80%, and the classification precision of reed is higher than 70%. The application results show that the method is feasible and can reduce the workload of artificial information retrieval. Finally, an automatic classification software system is developed in the IDL and ENVI environment. The system is robust and can be expanded, providing a model for object classification of river mouth coastal zone.