There has been a lot of research in the field of definition of the representation, the data fusion, and the generation of probabilistic occupancy maps. Here the occupancy maps are processed as textured images to extract meaningful information which is useful for the robot navigation. modern segmentation techniques' application is considered over the probabilistic occupancy maps. The local binary pattern and contrast is used as a basis for enhancements. The enhanced LBP/C segmentation technique handles occupancy uncertainty. free, unknown, and occupied are the three deterministic occupancy states defined. The regions characterized by a given range of occupancy states by increasing the number of classification levels which makes the approach flexible. The ground-based probabilistic grids is used to evaluate the performance of the system. The method is used for mobile robot navigation with collision avoidance. The evaluation by aerial and satellite images also supports the claims made.
A method for compact representation of space occupancy is provided by the occupancy maps. efficient robotic platforms can come up when mobile robots are navigated from such maps. The autonomous perception capabilities suggest that regions are scanned without any gap between
the viewpoints. The uncertain range measurements collected from several viewpoints are merged using the Bayesian merge technique

get the report here:
http://goo.gl/P4I3