Multispectral landsat image classification using fuzzy expert systems

A hierarchical fuzzy expert system is proposed for multispectral Landsat image classification to overcome difficulties with conventional maximum-likelihood classifier (MLC) based on normal distribution and easily incorporate other collateral data, such as vegetation index, digital elevation model, etc. The hierarchical structure is to reduce fuzzy rules to incorporate as many useful data sources as possible. Adaptive-Neural-Network Based Fuzzy Inference System (ANFIS) is used to build up fuzzy rule based systems to adapt training data. The expert system is tested for the classification on Landsat 7 ETM+ image and results are effective for multispectral image classification.