A new way of calculating the sub-pixel confusion matrix: a comparative evaluation using an artificial dataset

A new way of calculating the sub-pixel confusion matrix: a comparative evaluation using an artificial dataset

Stien Heremans and Jos Van Orshoven

KU Leuven, Department of Earth and Environmental Sciences Celestijnenlaan 200E, 3000 Leuven (Belgium) (stien.heremans@ees.kuleuven.be)

Abstract: This paper introduces a new, alternative method for calculating the sub-pixel confusion matrix. It calculates the off-diagonal matrix elements from the slope coefficients of the regression relations between the overestimations and the underestimations of the area fractions. The new method is set against existing approaches for calculating the sub-pixel confusion matrix through the use of an artificial dataset. The results show that it is able to compete with the existing approaches and in some cases even outperforms the latter.

Keywords: sub-pixel classification, confusion matrix, off-diagonal elements

AttachmentSize
HeremansAccuracy2012.pdf77.25 KB