Ant Colony Optimization algorithm for remote sensing image classification using combined features

Applying ant colony optimization algorithm on the remote sensing image classification is a new research topic, and the preliminary experiments showed many promising characters, but there are also some shortcomings such as needing longer computing time and the classification accuracy is not high enough when using single feature of the image. In order to overcome these defects, we propose to combine gray feature and texture features to improve the classification rate in this paper. We also investigated the relationship between the number of ants and the classifications accuracy. The experimental results prove that the improvement achieved by using combined features vector.