Aerial image clustering analysis based on genetic fuzzy c-means algorithm and Gabor-Gist descriptor

In the study of identifying homogeneous regions in remote sensing images, fuzzy clustering is one of the most frequently used algorithms. Commonly used method of fuzzy cluster analysis is the fuzzy C-means algorithm (FCM), which easily traps into local optimal solution. An algorithm combining FCM with genetic algorithms is introduced for aerial remote sensing image fuzzy clustering analysis. The input image features are extracted based on a new descriptor which combines Gabor descriptor with Gist descriptor. The dimension reduction of the extracted feature vector is processed through principal component analysis. Then the extracted features from in-house aerial images dataset are clustered with proposed method. Experiment shows that this method can get a good clustering effect.