Geintra

Image transmission by incoherent optical fibers bundles requires a previous calibration to obtain the fibers spatial in-out correspondence in order to reconstruct at the output side an image presented on the input of the system. This information is recorded in a look-up table for later reordering of the fibers positions and reconstructs the original image. This paper shows how to minimize calibration time and improve the quality of the recovered image, by using a fibers detection process reducing the number of points to take into account in the calibration process. There is, also, an improvement in the quality of the recovered image compared with others techniques. In this paper two detection methods of circular patterns have been applied to a fiber bundle image captured by a CMOS sensor. The former utilizes the algorithm of circular Hough transform based on the gradient field of the image. A second one is introduced as a new hybrid algorithm that mixes some morphological transformations and distance transform. The results demonstrate that this technique offers a remarkable reduction of processing time and good detection.