From a visual examination of the object-based CIROnOff image
classification, it seemed more accurate than the 82.0% stated by
the accuracy assessment. This contradiction between the
improved visual interpretation and the lack of improvement in the
accuracy assessment supports issues related to the use of standard
error matrix analysis as applied to object-based classifications
(Carleer and Wolff, 2006; Gao and Mas, 2008; Radoux et al.,
2008). This is an area for further research that is beyond the
scope of this paper.