Abstract
An approach to multidimensional representation of underwater environments is presented with results of applications of 3-D sonar mapping. A nondeterministic model incorporates information from multiple knowledge sources and creates a framework for real-time processing. Probabilistic methods account for nonideal sensors; spatial decomposition and numerical techniques treat amorphous underwater features and facilitate an incremental approach. An emphasis on representational and modeling issues is maintained with examples drawn from computer simulations and field data from profiling and imaging sonars.

This publication has 0 references indexed in Scilit: