Towards the Development of a Multisensorial Vision System: An Adaptive Approach to the Low-Level Phase

摘要: The use of vision systems in order to accomplish the task recognizing and describing 3D outdoor scenes is an open problem faced by current research on Computer Vision. In past few years new methodologies have been developed which make it possible integrate different information sources about a given surrounding environment into single recognition framework. system proposed this paper has designed reach high degree reliability, adaptively integrating data coming from variable number physical virtual imaging sensors. addressed lies optimizing choice low-level methods for processing outputs set sensors, extracting that better describe, at symbolic level, scene under examination. system, integrates techniques typical Pattern Recognition, Artificial Intelligence, Image Processing characterized modular structure capability evaluating results obtained during whole process. To end attention focused correlating numerical aspects computational process; consequently, made assess progressive integration according accurate evaluation criteria.