ABSTRACT

When natural resources are mapped using remotely sensed images, sampling design strategies are required to collect field observations of variables of interest at the selected locations and at the same time to create spatial explicit estimates of the variables (Brus and de Gruijter 1997). Moreover, the obtained ground data can also be used to assess the quality of the resulting maps (Campbell and Wynne 2011). Thus, selection of sampling design strategies is very critical for both mapping and accuracy assessment of natural resources.