CGH-ANN based system in interference pattern recognition

Interference figures are often subject of interest in subsurface sensing technologies. Proper further processing of them is essential for interpretation of data covered in such images. This interpretation is often possible after recognition of the interference patterns. The article presents pattern recognition system suitable for dealing with interference figures. The system consists of optimized computer-generated hologram used for feature extraction and artificial neural network used as classifrer of features. This pattern recognizer was tested with images of intermodal interference occurring in the optical fiber. If this fiber is embedded in the polymer composite material then such subsurface sensor together with mentioned pattern recognition system can be used for determining stress and distortion of that material. Since polymers are wide utilized for different constructions, including airplane wings, presented hybrid system can be used for real time, nondestructive monitoring of working stresses occurring in these constructions. The recognition of critical compressive stress can be therefore an early alarm signal of possible forthcoming danger.