Defect detection is an essential step in paper currency sorting. In this paper, an edge-based algorithm is proposed to detect the scratches and cracks appearing frequently on paper currency. An area-based image registration algorithm is used to overlay the sensed and referenced paper currency images. To ensure accurate correlation with the subjective feelings of human beings, an edge intensity differential of two images is then constructed from the edge information extracted by the Kirsch operator. The defect feature extracted from edge intensity differential is sensitive to the odd edge-information, and is robust to the gray (or edge ) intensity change. The paper currency image is divided into several overlapping subzones. Within each subzone, the defect feature is calculated to estimate the level of contamination. The proposed algorithm has already been applied to a practical sorting system, and the experimental results reveal that it is robust when applied to low quality paper currency.