Automatic Pixel Classification in Remote Sensing Satellite Imagery Using a New Multiobjective Simulated Annealing Based Clustering Technique

An important approach for unsupervised landcover classification in remote sensing images is the clustering of pixels in the spectral domain into several partitions. In this paper, a multiobjective optimization algorithm is utilized to tackle the problem of partitioning where a number of different cluster validity indices are simultaneously optimized. New multiobjective clustering algorithm uses the newly developed simulated annealing based multiobjective optimization technique (AMOSA) as the underlying optimization criterion. Here, center based encoding is used. Each cluster is divided into several small hyperspherical subclusters and the centers of all these small sub-clusters are encoded in a string to represent the whole cluster. For assigning points to different clusters these sub-clusters are considered individually. But for the purpose of objective function evaluation, these subclusters are merged appropriately to form some variable number of whole clusters. Two objective functions, one reflecting the total compactness of the partitionings based on the Euclidean distance, and another reflecting the total symmetrical compactness of the obtained partitioning are considered here. These are optimized simultaneously using AMOSA to detect the appropriate number of clusters and the appropriate partitioning from remote sensing image data sets. A new method is also developed to determine a single solution from the final Pareto optimal front provided by the newly developed multiobjective clustering technique (multicenter-AMOSA). Different landcover regions in remote sensing imagery have also been classified using the proposed technique to establish its efficiency. Results are compared with those obtained by fuzzy C-means (FCM) clustering technique and a recently developed symmetry based automatic clustering technique, VGAPS, both qualitatively and quantitatively. Seminar on Spatial Information Retrieval, Analysis, Reasoning and Modelling 18-20 th March 2009. ISI-DRTC, Bangalore, India 99