Performance Analysis of Biogeography Based Land Cover Feature Extractor for Building Hybrid Intelligent Models

This paper is an analytical study of the performance governing factors of the biogeography based land cover feature extraction technique which is characterized by its ability to perform differently on different natural terrain features contained in a satellite image. From the discussion, we establish the fact that the classification efficiency of BBO for a given land cover feature is inversely proportional to the degree of disorder and directly proportional to the similarity index of the digital number (DN) values of the pixels comprising that land cover feature when viewed in any of the bands of the multi-spectral satellite image. In order to verify our proposed hypotheses, we calculated the entropies and similarity indices for each of the land cover feature in two bands on two different datasets and found the same results in both the bands for each of the datasets we took, thus validating the theory. This finding is of prime importance since a prior analysis of the performance of biogeography based feature extraction technique on different types of terrain under consideration will improve the analyst’s decisive capabilities for the selection of the most appropriate technique for the feature extraction task in hand. This in turn can be applied for building efficient artificially intelligent hybrid classifiers by applying the BBO technique on the extraction of those features on which it shows maximum classification efficiency as demonstrated in the paper. Performance Analysis of Biogeography Based Land Cover Feature Extractor for Building Hybrid Intelligent Models