Investigation on satellite data fusion operation effects on accuracy of forest type classification (Case study: Mazandaran province)

The combination of panchromatic band with multispectral data is known as data fusion operation, which increases ground resolution of multispectral data. The results of the application of this method on the accuracy level of forest type classification were investigated. The multispectral data and PAN band from the ETM+ sensors belong to forest district located in "Azarrood" forestry plans in "Savadkooh" were used for the sake of this survey. The bands were controlled according to Radiometric and geometric error separately. Band 1, was omitted because of the existence of radiometric error and its less importance in vegetation cover study. Geometric correction was performed by 22 round control points with DEM, up to orthorectification level, with up precision of least than half pixel. Fusion operation was performed by PC method. The supervised classification was performed with using basic and synthetic bands (ratio, principle Analysis), with maximum likelihood, minimum distance, parallelepiped algorithm. The highest overall accuracy was 34.12% and kappa coefficient 23.15% for separability six vegetation type. The result shows the reduction of data correctness in compare with the result of produced forest type map without fusion. That as the reasons, we can indicated are, selected method type for fusion operation, heterogeneous area. Therefore, the repetition of this operation in another area with using other fusion methods is proposed.