Applications of simulated annealing minimization technique to unsupervised classification of remotely sensed data

The research in this paper is designed to develop algorithm and applications to apply the simulated annealing to unsupervised classification and compare the classification results with that of the K-means. It is known that the K-means can only produce local minimal solutions. We propose to use the simulated annealing to solve this problem and find a global solution in the clustering process. Details about how to choose parameters and cooling schedule are discussed based on experiment results. Finally, the future work to extend this research is proposed and presented.