Proceedings of the 5th International Association of Science and Technology for Development (IASTED) International Conference on Robotics and Manufacturing, May 29-31, 1997., Cancun, Mexico
This article defines a method for computing a proximity compatibility function among fragmented 3-D surfaces for environment modelling. Fragmented surfaces are a common occurrence after the segmentation process has been applied to 3-D sensory data, in particular for data taken from large indoor environments. This proximity compatibility function among surfaces gives an indication on how close the surfaces are to each other based on a common gap defined between the boundaries of the surfaces. This particular approach performs most of the computations in the 2-D image plane and when required will use the 3-D information in the data. This is simpler than tackling the whole problem in Euclidean space and as an added bonus, sections of the algorithm pertaining to the 2-D image plane can be applied to classical 2-D intensity images.