Global Strategy for improving agricultural statistics, Implementation Plan for Africa, Research and development, Remote Sensing and area frame sampling

The paper was structured classifying the approaches for the use of remote sensing data for agricultural statistics into two groups: pure remote sensing approaches and integration of remote sensing and ground survey. Pure remote sensing approaches are generally risky, since they leave the analyst a large margin of subjectivity. The second approach needs a considerable amount of ground data collection; it allows to quantify the contribution of remote sensing to improve the accuracy and therefore a cost efficiency analysis. In most African landscapes only high resolution images are suitable. The cost of such images need to be taken into account. A major topic of research is the cost efficiency of the integration of remote sensing and ground survey in African countries. Other research topics were proposed including the use of AFRICOVER for stratification, the use of GPS for field measurement, small area estimation, assessment of the most appropriate area frame for specific landscape types and the integration of administrative data with the the other data sources. Some questions were risen, particularly concerning the possibility of combining households surveys with remote sensing data.