Modelado de sistemas de transmisión y reconstrucción de imágenes basados en mazos de fibra óptica no coherentes

Zusammenfassung

This thesis belongs to the scientific field of visual inspection systems using computer vision and artificial intelligence. Among the possible scenarios for implementing these systems, the work has been focused on the issue of remote monitoring of environments. This issue is of great interest in many industrial and research sectors. The purpose of this thesis is to propose a new global model calibration and image transmission using incoherent optical fiber bundles. The model presented includes the problems to be solved to calculate the transfer function needed in the transmission of images, and also some general aspects for designing. It has been developed a methodology for calibration and imaging transmission which has been tested by an experimental setup capable of characterizing different incoherent bundles and also evaluate different spatial calibration and imaging strategies. To carry out the experiments shown, has been developed an application that facilitates the empirical evaluation of the model referenced to a wide range of conditions. The results presented demonstrate the advantages that a system like this can bring out to the inspection of remote environments with difficult access, or where it is risky to use conventional electronic cameras.