PATTERN RECOGNITION FOR AUTOMATIC INSPECTION IN FACTORY FLOORS USING MODELLING TECHNIQUES BASED ON IMPLICIT FUNCTIONS AND DISTANCE MAPS

摘要

One of the most interesting stages in the field of computer vision is related to object modelling and its application to pattern recognition problems. The modelling stage starts from previous scene segmentation, aiming at supplying a set of shape describing primitives to be used in further stages as pattern recognition. This symbolic description should verify the property of uniqueness together with traaslation, rotation and scaling invariance. Moreover, object modelling could be applied to image analysis and synthesis, which converts it in a useful tool. The paper is focused on 2D modelling techniques easily extended to the 3D case, with application to the detection of serigraphy errors in glazing terminals of the Tile Manufacturing Industry, through parameter comparison of primitives.