During validation, a confidence-building procedure will be performed, which will consist in a systematic quality control of the CCI land cover products. This procedure is intended to meet two main objectives: the elimination of macroscopic errors and an increase in the overall acceptance of the land cover product by users. Systematic quality control is also a way of assessing if the remotely sensed data have been correctly classified, i.e. if the errors are due to limitations of data quality rather than to poor classification procedures.
The quality control is based on a systematic descriptive protocol in which the map is segmented into regular cells, each cell is visually examined and the cells accuracy is documented in terms of type of error, landscape pattern, reference material used, etc. Ascertaining the nature of the errors occurring in the cell is of primary importance. Indeed, statistical accuracy assessment merges in the category “error” many different cases that quality control can easily document. Such information can be profitably used for improving the map during the updating phase. Once all the cells have been visited, the results will be presented in a tabular manner or on a map. It will then be possible to investigate the influence of the parameters (heterogeneity, dominant class) on the quality of the land cover map.
This systematic quality control will be integrated into the classification procedure, with the results of the analysis employed for removing errors and improving the map.
