Development of a neighbourhood graph for trafficability analysis

Successful trafficability analysis requires usage of many data sources, e.g. elevation and land use data, in high resolution. However, this requires a large storage capacity and may lead to unduly long execution times when trying to find suitable paths. To allow for efficient representation and search in an area of interest (AOI), a searchable graph is created where the best paths between two arbitrary locations in the AOI can be found. Terrain features, identified from laser-radar data, and land use data are fused to create a compound map of the AOI. In order to create the graph, the map is partitioned into areas of homogeneous trafficability characteristics. These areas are represented by the nodes in the graph and the edges in the graph represent paths between neighbouring areas. Different methods to partition the map is described and analysed. Dijkstra´s algorithm are combined with the A*-algorithm to find the best paths between two locations in the AOI.