Land use land cover analysis aided by remote sensing provides the civic authorities a broader picture of the geographic area. This work aims to get the land cover map of the Bangalore Urban region using multispectral remotely sensed images. Three individual classifiers, namely random forest classifier, support vector machine classifier and the k-nearest neighbour, have been used. For the study area under consideration, the spectral signatures (which is the basis for classification) of water, vegetation and soil are not very distinct, which may cause misclassification. Also, the study area has imbalanced classes, because of which the individual classifiers may not give optimal results. The novelty of this work lies in the handling of these two shortcomings. A random unweighted hard MV (Overall accuracy = 82.75), priority-based unweighted hard MV (Overall accuracy = 86.06) and weighted hard MV (Overall accuracy = 86.92) classifiers are implemented here, and it is shown that they provide a better overall accuracy compared to the individual classifiers.