Remote Sensing Image Classification based on Combined Clustering – WAE Classification Model

Abstract

Classification of remote sensing images plays a major role in the research areas of Image processing. Classifying the images into different area reveals lot of new details to the society. Thus, considering the importance of the field, the paper proposes a new method of classifying the remote sensing image. The proposed model is constructed, which is based on the unsupervised classification method. It works based on Autoencoder (AE) network with the clustering logic and is named as clustering-wishart-auto-encoder (WAE) classification model. The clustering algorithms can explore the label information automatically. The clustering-WAE classification model, embeds the K-means clustering algorithm into the objective function of the WAE model, thus, producing the efficient classification accuracy. Experiments are carried out to discuss the outperforming results of the proposed method and they are found to be good.