No abstract

We describe methods for using colour and texture to discriminate cloud and sky in images captured using a ground based colour camera. Neither method alone has proved sufficient to distinguish between different types of cloud, and between cloud and sky in general. Classification can be improved by combining the features using a Bayesian scheme.

The employment of Artificial Neural Networks to the classification of meteorological data has been considered in previous papers and found to offer promising results. We compare the performance of the Bayesian Classifier with two different Neural Network architectures. The classifiers were used to segment images of cloud into four different meteorological classes on the basis of spectral-textural measurements. The experimental design is based on a set of 60 hand-classified images, random sampling of which enables us to generate training and test sets. This design allows us to carry out a repeatable comparison of the classifiers with different training and test data.