A symbiotic organisms search algorithm for feature selection in satellite image classification

The image classification performance depends a lot on the best choice of the descriptors and the techniques used to extract them. With the exponential growth of data in the field of remote sensing, classifying these massive images still remains an open and challenging issue. The high dimensionality of the feature space, not only increases the time and space complexities, but also may reduce the image classification performance in terms of accuracy and time to build the classification model. To overcome this challenge, this paper presents a novel feature selection method based on a combinatorial optimization algorithm for training a feed-forward Artificial Neural Networks to select a small number of features while maintaining good classification rates. The performance of the proposed method is tested on real image dataset and compared with other state-of-the-art methods. The experimental results show that the proposed method has good performances.