REAL-TIME COLOR RECOGNITION IN SYMBOLIC PROGRAMMING FOR MACHINE VISION SYSTEMS

It is impossible to collect more than a tiny proportion of all of the possible examples of a given hue to form a training set for a machine that learns to discriminate colours. In view of this, it is argued tha t colour generelization is essential. Three mechanisms for learning co lours, as defined by a human being, are described. One of these is bas ed upon an idea developed by A.P. Plummer and is implemented in a comm ercial device known as the ''intelligent camera''. This implementation can learn the characteristics of coloured scenes presented to it and can segment a video image in real-time. This paper presents four proce dures that allow the range of colours learned by such a system to be b roadened so that recognition is made more reliable and less prone to g enerating noisy images that are difficult to analyse. Three of the pro cedures can be used to improve colour discrimination, while a fourth p rocedure is used when a single and general colour concept has to be le arned. Several experiments were devised to demonstrate the effectivene ss of colour generelization. These have shown that it is indeed possib le to achieve reliable colour discrimination / recognnition for such t asks as inspecting packaging and fruit. A practical system based upon the intelligent camera and controlled by software written in PROLOG ha s been developed by the authors and is being used in a study of method s for declarative programming of machine vision systems for industrial applications.