| Popis: |
At present, the most technology of counting money is to use the money counter in financial fields. The paper presents a new method for automatic counting paper money which is based on image processing technology. Firstly, the paper money image is acquired by CCD. After analyzing the feature of image, we find that in Cr-space the edge of each paper money is enhanced. Then we use the north-west sobel operator for filtering and north sobel operator for detecting edge. Although the image-processed better highlight the edge of each paper money, the edge is rough and its variance is high. It is hardly to threshold the image for getting the single-pixel edge linked. After Different segmentation algorithm was been used for deriving the edge of paper money, we find the Two-dimensional Histogram θ-division algorithm is suitable for our purpose. The experimental result is proved satisfied. The detecting rate reached 100% in controlled environment for RMB. However, if we want to detect other kinds of paper money such as dollar, there also have several problems to be solved. |
