Scholarly record

ACCURACY ASSESSMENT OF COLLECTED IN THE TOPOGRAPHIC OBJECTS DATABASE BY THE RESISTANT TO OUTLIERS METHODS

Marek Ślusarski

Abstract

Spatial databases collect information about geographic objects that represent real world phenomena. Stored and shared data sets are not homogeneous, because they are obtained on the basis of diversified source materials. Distributed sources of spatial data are characterized by a different organization in terms of completeness, consistency and accuracy. In Poland, the topographic objects database (BDOT10k) is one of the basic databases covering the entire country. The positional accuracy is a basic element of the spatial data quality, which directly affects the value of geospatial analysis. The article presents methods resistant to outliers (modified by the author), which have been used to assess the accuracy of data collected in the topographic objects databases. During research three resistant techniques were analyzed: the method of the smallest absolute deviations, the Danish method and the method of selecting alternatives. In the studied methods, the modification consisted in using median and absolute deviation of the median to diagnose outliers, instead of the mean value and standard deviation. The comparative reference criterion for the applicability of modified resistant methods is classical statistics based on large, truncated random samples. The numerical values of the calculated estimators show that there is a satisfactory agreement in estimating the positional accuracy of the spatial objects.

Publication details

Back to publication list