Surface blemish detection from passive imagery using learned fuzzy set concepts

An image analysis method for real-time surface blemish detection using passive imagery and fuzzy set concepts is described. The method develops an internal knowledge representation for surface blemish characteristics on the basis of "experience", thus facilitating autonomous learning based upon positive and negative exemplars. The method incorporates fuzzy set concepts in the learning subsystem and image segmentation algorithms, thereby mimicking human visual perception. This enables a generic solution for color image segmentation.