Image Defect Recognition Based on Rough Set

This paper applies rough set theory to recognition system for image defect, and designs a decision algorithm on rough set suitable for image defect recognition. Firstly, the image is made regionalization and sequential discrete set is proposed, the continuous attributes of image is discretized. Then the decision table model on discrete condition attributes and decision attributes is constructed. Further the condition attributes significance function and reduction algorithm is given. A novel approach for decision rule analysis and rough set recognition is proposed. Finally, this paper takes the example for fabric defect recognition to validate these algorithms. The result shows the rough set algorithm is effective for image defect recognition with less calculation and fast speed.