QUALITY ASSESSMENT OF POLYGON LABELING

The high costs in time and resources of data validation makes the corr ect selection of a simple but robust adjustment method crucial. In thi s paper we describe an approach suitable for a polygon labeling assess ment. The approach uses the geometric distribution where the sample si ze is a function of the desired confidence level of the database. It a llows the detailed random verification of a selected map. All the area is equally exposed to testing and, if needed, to correction.