ABSTRACT
Noise suppression algorithms have been used in various tasks such as computer vision, industrial inspection, and video surveillance, among others. The robust image processing systems need to be fed with images closer to a real scene; however, sometimes, due to external factors, the data that represent the image captured are altered, which is translated into a loss of information. In this way, there are required procedures to recover data information closest to the real scene. This research project proposes a Denoising Vanilla Autoencoding (DVA) architecture by means of unsupervised neural networks for Gaussian denoising in color and grayscale images. The methodology improves other state-of-the-art architectures by means of objective numerical results. Additionally, a validation set and a high-resolution noisy image set are used, which reveal that our proposal outperforms other types of neural networks responsible for suppressing noise in images.
ABSTRACT
In this work the feasibility of employing two well-known techniques already used on designing optical fiber sensors is explored. The first technique employed involves monomode tapered fibers, which were fabricated using a taper machine designed, built, and implemented in our laboratory. This implementation greatly reduced the costs and fabrication time allowing us to produce the desired taper length and transmission conditions. The second technique used fiber Bragg gratings, which we decided to have mechanically induced and for that reason we devised and produced our own mechanical gratings with the help of a computer numerical control tool. This grating had to be fabricated with aluminum to withstand temperatures of up to 600°C. When light traveling through an optical fiber reaches a taper it couples into the cladding layer and comes back into the core when the taper ends. In the same manner, when the light encounters gratings in the fiber, it couples to the cladding modes, and when the gratings end, the light couples back into the core. For our experimentation, the tapering machine was programmed to fabricate single-mode tapers with 3 cm length, and the mechanically induced gratings characteristics were 5 cm length, and had a period of 500 µm and depth of the period of 300 µm. For the conducting tests, the tapered fiber is positioned in between two aluminum slabs, one grooved and the other plane. These two blocks accomplish the mechanically induced long period grating (LPG); the gratings on the grooved plaque are imprinted on the taper forming the period gratings. An optical spectrum analyzer is used to observe the changes on the transmission spectrum as the temperature varies from 20°C to 600°C. The resultant attenuation peak wavelength in the transmission spectrum shifts up to 8 nm, which is a higher shift compared to what has been reported using nontapered fibers. As the temperature increases there is no longer a shift, but there is significant power loss. Such a characteristic can be used as well for sensing applications.
