Forest mapping accuracies are improved using a supervised nonparametric classifier with SPOT data

Abstract

A new supervised nonparametric classifier produces an image showing the empirical probability of correct classification for a pixel as well as a thematic image. This allows an analyst to visually locate those parts of the image where classification success can be improved. The algorithm was tested using SPOT XS data over a forest plantation in southeast Australia. The classifier produced thematic maps of higher accuracy than those from conventional supervised classifiers.