Editing strategies for digitally correlated gestalt data

Abstract

Gestalt digital elevation and other digitally correlated data often have discontinuities that occur along boundaries between adjacent patches and the data usually have lower accuracy than manually digitized elevations. A production method for detecting such discontinuities, which involves the analysis of the mean height of the four corner points formed by neighbouring patches, is first discussed. A second method is then proposed utilizing the absolute second differences of triplets of points with an empirical test based on essentially the same principle as the statistical test for differences between sample means. Some considerations for filtering the edited data are briefly mentioned to indicate the type of data processing that can be done to meet various application requirements.