A new image processing method for damaged bill recognition

A large volume of bills are circulated and exchanged for goods and services. These bills are vulnerable to tearing, fluid or dirt contamination, wrinkling, hand-writing, print peeling off, and other damage. Severely damaged bills may cause bill handling machines malfunction and are unfit for circulation. It is required that damaged bills be recognized and removed from the money supply chain. This paper presents a novel image processing method for damaged bill recognition. Different from the commonly used image recognition algorithms, the developed method takes the length of a variation curve of reflected light intensity from the billpsilas surface as an evaluation parameter, and compared it against a predetermined standard. The effectiveness of the proposed method is experimentally verified in this paper.