A Hierarchical Approach toward 3-d Geo-spatial Terrain Merging

Nowadays, DTM databases are vital for a variety of applications, such as visualization or terrain analysis. Therefore, in the fields of civil engineering, mapping and geo-information, or geophysics, this data is on the functional route leading from data acquisition to derived applications. One of the main issues when dealing with DTM database representation is data merging, which involves the integration of data from different sets. Various factors cause global-systematic errors as well as local-random ones, which reflect on geometric and radiometric differences. This paper describes a new approach to merging DTM datasets, which analyze the local inconsistencies of DTM geo-spatial databases in order to give an appropriate integration solution. The idea is to implement a hierarchical solution of pyramidal approach, in which local geometric discrepancies are monitored. The proposed solution suggests the implementation of two topographic zoning working levels – global and local. This solution offers control over the various levels of errors, and is implemented as follows: zonal division of the whole datasets area into patches, in which a local registration is extracted for each; sub-zonal division, in which an accurate 'local' ICP matching process is achieved. This process yields the extraction of a new database, which stores data representing local discrepancies (in the form of transformation parameters) of the integrated DTMs. This new approach yielded accurate results for DTM datasets merging, therefore achieving a singular, unified and spatial continuous surface representation of the terrain relief.