ABSTRACT
Correct, accurate, and updated representation and modeling of our natural environment is recognized as an invaluable resource. One of the more widespread and applicable data type representations is achieved by using digital terrain model (DTM) data sets. A variety of applications, such as visualization or terrain analysis, in the fields of mapping and geoinformation (or geophysics) can benefit from using upto-date DTM data sets. A key issue still to be addressed when working with DTM data set representation is data merging: integrating data from different sets. Various factors cause global systematic errors as well as local random ones, which reflect on geometric and radiometric data representation. In this chapter a new approach for merging DTM data sets is presented, which analyzes local inconsistencies inherent in geospatial data sets prior to actual data integration. The concept of implementing a hierarchical approach is introduced, in which global geometric discrepancies are monitored, and then used locally for accurate spatial modeling. This approach
leads to a more qualitative and reliable representation of the natural environment, thus offering control over the various levels of errors. As part of the proposed merging solution, the extraction of a new DTM look-alike database is introduced, which stores data that represents local discrepancies in the form of affine transformation parameters of the whole integrated area. The hierarchical approach produces a singular, unified, and spatially continuous surface representation of the terrain relief, achieving a more accurate modeling result of the terrain than any of the original data sets.
