M2DSOMAP: Clustering And Classification Of Remotely Sensed Imagery By Combining Multiple Kolionen Self-organizing Maps And Associative Memory

摘要

This paper investigates a hybrid neural network framework by combining unsupervised an supervised neural learning paradigms on a unified representation platform of multiple Kohonen 2 dimensional self-organizing maps (M2dSOM) with the assistance of associative memory for clustering and classification of remotely sensed (RS) imagery. In contrast to the boundary (hyperplane) representational form of decision region by the Multi-layer Perceptron (MLP), the M2dSOM is a regional form of such for both cluster region and decision region. The formation of the clusters and the transformation from clusters to decision regions are implemented by unsupervised and supervised self-organizing learnings on M2dSOM, respectively. A new supervised learning algorithm is proposed that exploits the input portion of supervising samples to discover mismatches between. cluster and decision regions by a k-winner selection process and then correct the cluster boundaries based on a majority vote for a new cluster membership from the k winners. Finally, an associative memory is employed to form a mapping between clusters and classification labels by samples. Two association configurations are suggested. Analysis of this mapping SONN model (called M2dSOMAP) in relation to RS imagery analysis with comparison to other methods is briefly discussed. Preliminary experiments on imagery analysis have shown that the M2dSOMAP has at least as good a classification performance as MLP and MLC and provides a number of advantages over both.