RANDOMNESS-BASED INTEGRATION OF MULTI-VIEWPOINT IMAGERY FOR AS-IS ROADWAY MODEL GENERATION

摘要

A new framework is presented for identifying roadway pattern in multi-viewpoint imagery. By applying fractal sampling to scene and satellite imagery, chromatic complexity of random texture is extracted and adapted for multi-viewpoint association. The feasibility of the framework is investigated through experimental studies.