SELECTION OF OPTIMUM BANDS FROM TM SCENES THROUGH MUTUAL INFORMATION ANALYSIS

The selection of optimum band subsets from remotely sensed scenes for visual interpretation or automatic processing is an interesting task w hich will assume growing importance with the availability of highly mu ltispectral data from future sensors. The usual methods for the mathem atical evaluation of the best combination of channels are based on par ametric statistical procedures such as eigenvector analysis and calcul ation of separability measurements. These procedures are not easy to b e interpreted or computationally expensive and are not suited for eval uating the probabilistic information which can be exploited by non-par ametric processes. For this kind of application, a method based on mut ual information analysis is put forward in the present paper. Mutual i nformation analysis is a statistical procedure which, using the concep t of system entropy, is capable of mathematically evaluating the proba bilistic information common to different variables. When applied to re motely sensed scenes superimposed on ground references related to some theme (for example vegetation types), information analysis can indica te which channels express more information about that theme. The metho d was applied to some Landsat TM scenes from three Italian areas about which ground references were available. Some mixed parametric-non-par ametric classifications were then performed to test the results of the analyses. From these tests the subsets identified were demonstrated t o be significantly more informative than standard subsets, which testi fies to the efficiency of the procedure.