Advanced Semi-Supervised Possibilistic Fuzzy C-means Clustering Using Spatial-Spectral Distance for Land-Cover Classification

With the explosion of information, characteristics of increasingly complex data, the use of traditional methods in data processing has proved ineffective. Computer applications are increasingly becoming important and essential in many areas such as biology, medicine, psychology, economics, image processing and many other disciplines. A variety of multi-spectral satellite image classification, clustering algorithms have been developed and applied to analyze the surface of the earth. In this paper, we propose a novel semi-supervised possibilistic fuzzy c-means clustering on spatial-spectral distance (SPFCM-SS) for multi-spectral image land-cover classification by the extension of the possibilistic fuzzy C-means (PFCM) algorithm, in which spectral information and spatial information of the pixels are used coupled with labelled data to increase the accuracy of clustering results when the data structure of input patterns is non-spherical and complex. Experiments were performed for multi-spectral satellite image data and clustering efficiency indexes were used to compare the performance of the proposed algorithm with other similar algorithms. © 2018 IEEE.