STATISTICAL GREY-LEVEL MODELS FOR OBJECT LOCATION AND IDENTIFICATION

This paper presents a new method for modelling and Locating objects in images for applications such as Printed Circuit Board (PCB) inspectio n. Objects of interest are assumed to exhibit little variation in size or shape from one example to the next, but may vary considerably in g rey-level appearance. Simple correlation based approaches perform poor ly on such examples. To deal with variation we build statistical model s of the grey levels across the structure in a set of training example s. A multi-resolution search technique is used to locate the best matc h to the model in an area of a new image to sub-pixel accuracy. A fit measure with predictable statistical properties can then be used to de termine the probability that best match is a valid example of the mode l. We describe a 'bootstrap' approach to training and a method of auto matically refining the final model to improve its performance. We demo nstrate the method on PCB inspection, showing the approach is robust e nough for use in a real production environment.