Image edge detection based on otsu and rough set theory

A new algorithm of image edge detecting was proposed, which was based on otsu method and the set´s approximation of rough set theory. Firstly, the global threshold of image segmenting was selected through otsu technique for this algorithm. Then there was a 3*3 pixels window, according to the attributes of the relations between threshold and window pixels´ grey levels to build decision table. It could deduce the upper approximation and the lower approximation of image edge sets. And at last, the edge image was obtained using the difference between the two approximation sets. The experimental result shows that the method can successfully preserve the image edge details, make the edge image meticulous and vivid, and have powerful ability in its resistance against noise pollution and better continuity and closure.