Remote Sensing Based Two-Stage Sampling for Accuracy Assessment and Area Estimation of Land Cover Changes

摘要: Land cover change processes are accelerating at the regional to global level. The remote sensing community has developed reliable and robust methods for wall-to-wall mapping of land changes; however, changes often occur rates below errors. In current publication, we propose a cost-effective approach complement maps with sampling approach, which is used accuracy assessment accurate estimation areas undergoing changes, including provision confidence intervals. We two-stage in order keep accuracy, efficiency, effort estimations balance. Stratification applied both stages gain control over sample size allocated rare classes on one hand cost constraints very high resolution reference imagery other. Bootstrapping measures area estimates verification rely quality visual interpretation units based time series satellite imagery. To demonstrate operational applicability it deforestation an characterized by frequent cloud low rate Republic Congo, makes monitoring particularly challenging.