Unsupervised classification of radar imagery of wetlands using the soft competition scheme

The authors apply a technique developed in the field of vector quantization to the problem of unsupervised classification of radar imagery of wetlands. The method can perform better than traditional unsupervised methods, such as k-means, because it performs soft classification at each step. The method is applied to Alaska data and shown to give results that are similar to previous results using supervised classification. Results are also shown for data from Belize.