The traditional hard classification techniques are parametric in nature and they expect data to follow a Gaussian distribution, they have been found to be performing poorly on high resolution satellite images, as classes in these images tend to exhibit extensive overlapping in spectral space. This produces spectral confusion among the classes and results in inaccurate classified images. A major drawback of such classifiers lies in the difficulty of integrating ancillary data, which follows a non Gaussian distribution nature. Ancillary data provides extra spectral and spatial knowledge, which improves the classification accuracy. Classification done using such knowledge is known as knowledge base classification. The present study explores a non-parametric decision tree classifier to extract knowledge from the spatial data in the form of classification rules. The classified image overall accuracy was found to be 86.66% using the Decision Tree method and with kappa values. 8133 respectively.