Extracció de característiques per a la detecció automàtica, òptica i digital de formes en color

In this project we intend to develop a system for pattern recognition applied to polychromatic or coloured objects. This fact adds the colour distribution information to the information usually used inimage recognition such as shape, scale, orientation. An optical system based in correlation techniques for the image recognition will be built. The filter will be positioned in the set up by computer control. The intensity distribution in the correlation plane will be digitalyzed and analyzed to obtain the features that will permit the recognition by discriminant functions. The features will be obtained by different types of computer generated filters: synthetic discriminant function, inverse filter, etc. We will also develop a digital approach of the polychromatic object recognition. It permits to obtain the correlations between the chromaticity distributions. These distributions are obtai